I spoke about being an addict, and about being addicted to cheap dopamine. What I maybe didn’t quite realize at the time is that I am the type of person who basically HAS to be addicted to something. It has been this way my whole life, and every single time I have tried to find that so-called “balance” in life, I have failed.

I don’t know if this is something a lot of other people can relate to, but I suspect it might be. While there are plenty of people who don’t fit this mold, and who can find “balance”, the majority of people I know have a small number of things that they centre their personalities around. They might not be addictions in the full and true sense of the word, but they’re obsessions, or devout hobbies, or pillars of personality. It could be video games, or running marathons, or traveling the world, or making money, or gardening, or playing a sport, or politics, or whatever. But most people have something, or somethings, that define them — and that they let themselves be defined by.

There’s a YouTube clip that I watch roughly once a year. It changed my life and the way I viewed the world the first time I watched it, and whenever I re-watch it, it is a wonderful reminder of how the world works, how our brains work, and how we have the power to choose how to view the world and our own reality.

The clip is titled This is Water by the late, great, David Foster Wallace, and I encourage everyone reading to give it a watch/listen if you haven’t before:

The whole thing is gold, but the part I really want to focus on is the thing he mentions towards the end about choosing what to worship:

What I got out of this was not necessarily to start worshipping a religion. I am a non-religious person myself. But rather, the realization that I would always be worshipping, or focusing, on something. If that was the case, I might as well try and choose things that are going to result in a better life for me and my family.

Video games, instagram, youtube, netflix, etc, are not great things to worship (for me). After cutting all of these things and many other screen related things from my life, I had a lot of time to fill, and have replaced it primarily with two things: golf, and reading. I was playing a good amount of golf before, but now I am playing more. I was barely reading before, and now I am reading for 3+ hours a day.

Most of my down time, my non-work time, my pleasure time, is spent on these two activities. Am I spending “too much” time on them? Probably. Am I “addicted” to them? In many senses of the word, probably. Do I think my life would be better off if I could find a way to have balance and spend a little less time on these two things and a bit more time on a few other things? Probably.

But I don’t know how to do that. It’s not how my brain is wired. I have always had addictions, and obsessions, and things that defined me.

Rather than fighting it, I am now choosing to embrace it. I am choosing to choose my addictions. I don’t imagine that I will stick with these two forever — there will be a point where I get over reading, or golf, and start looking for something else. The challenge to myself is to replace one relatively healthy addiction with another one.

It’s about constantly choosing the lesser of evils. I’m not saying it’s GREAT to be addicted to anything; it’s not GREAT to be addicted to reading, or golf, or exercising, or gardening, or anything. There are real downsides to getting addicted and obsessed and hyperfixating on things. But it is a hell of a lot better to be addicted to these things rather than being addicted to drinking, drugs, gambling, video games, watching TV, and so on and so forth.

Again, I don’t know how much of this is relatable to others. On the one hand, I know many people who are able to have that elusive “balance” and seem to be able to not get uber obsessed and hyper fixated on anything in particular. Good for them, I am envious. On the other hand however, I also know many people with addictive personalities, and maybe if that’s you, reading this will give you some food for thought.

My life, like everyone’s, is always in a constant state of flux. I try to improve it, learn from my mistakes, learn from others, and strive to better myself year after year. I don’t always succeed, but it’s in running experiments like my dopamine fast that help me understand myself better and discover ways to improve my life.

By writing about my experiences, I help myself work through things, and, hopefully, I help others relate to me and help them come to similar conclusions on their own. I know I get a lot out of reading the words of others, and that’s all I can hope for for anyone reading this now.

Cheers, and we should be back with some more crypto-focused content next week. I don’t want to jinx it, but it feels like we are getting closer and closer to a bottom and to the start of something bullish, so it’s probably a good time for all involved to be paying a bit more attention out there.

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

Looking back across those fifteen years, the pattern is almost comically clear. The stretches where I journaled were the stretches where I played my best and won the most. The stretches where I didn’t were the ones with the slow leaks, the bad habits creeping back in, and the downswings dragging on longer than they needed to.

Some of it was that the very act of journaling was making me better. Some of it was the version of me disciplined enough to journal also being disciplined enough to do everything else properly. Honestly, it doesn’t matter which. The lesson was identical either way. When I wrote things down, things went well. When I stopped, they didn’t go as well. It took me an embarrassing number of on-off cycles to accept this.

Then I repeated the whole cycle in crypto, because apparently some lessons need to be learned twice (or five times, or, heck, twenty times). I’ve had periods where I tracked every single trade, and periods where I tracked nothing and yolo traded at high frequencies based on nothing but vibes. Once again, the tracking periods were better. Shocking, right?

So today I am gonna share a bit about what I tracked for poker, what I track for crypto and trading / investing, and then also share a template at the end for anyone to use to begin to do your own journaling.

What I tracked

During the stretches when I did journal for poker, after each session I logged the date, how long I played, and my results, and then a comments column for additional notes. Things like whether I was playing after a big meal, if I was low on sleep, if I had a fight with my partner, if I felt in the zone, etc. Whatever was true went in, no matter how unflattering.

All of it was useful. The comments column especially so, because when I lined it up against the results, the patterns fell out on their own. The low sleep sessions underperformed. Playing after a big meal underperformed. Nothing surprising in hindsight, but there’s a world of difference between vaguely suspecting something and seeing it in your own spreadsheet in your own words. The comments column was quietly building a map of my weaknesses.

And then there were two more columns, which turned out to be the special ones.

The two scores

After each session, alongside everything above, I scored two things. And while I did, I tried to deliberately ignore the results column:

It’s so simple, but it changed everything for me. Believe it or not, it changed everything beyond poker — it changed how I thought about the world.

If I lost money while playing well through a stretch of bad luck, I’d finish the session feeling genuinely good because I got to put a high number in the “how well I played” column. The spreadsheet sat there reminding me the only thing I controlled was the quality of my decisions, and if my decisions were sound, over a long enough timeline things would work out in my favour.

If I won money while playing badly and getting bailed out by the deck, I’d feel a little uneasy. Which was exactly the right feeling to have. A lucky win is the most dangerous thing in gambling and in markets, because it teaches you the wrong lesson at the moment you’re most receptive to learning it.

The two scores moved my emotions onto the things within my control, and away from the things outside it.

Long time readers might remember Letter 24, where I wrote about aligning your emotions with your actions instead of your outcomes. This spreadsheet was the mechanism behind the philosophy. White knuckling your way into caring about process over results doesn’t work, at least it never did for me. You need a ritual. Mine was a spreadsheet, a comments column, and two numbers (and then later on, this entire mindset was bolstered for me by reading Stoic philosophy, and realizing that bright minds had been saying the same things for thousands of years).

Bringing it to crypto

Everything above translates directly to trading and investing, with one adjustment. In poker, you’re constantly making decisions. Every minute or so at one table, and multiple times a minute if you’re playing more than one table at a time. A crypto position might run for months, and even “short term” trades are often at least hours from open to close. So the journal entry happens in two parts, at entry and at exit.

The thesis and invalidation lines do the heavy lifting. If you’re unable to write a two sentence thesis, you don’t have one and shouldn’t be making the trade, so you’ve learned something valuable before risking a single dollar. And a thesis without an invalidation isn’t a thesis at all, it’s just hopium.

The two scores are where the poker magic carries over.

Playing well in crypto looks like: I sized sensibly, I entered according to a plan, I honoured my invalidation, I exited for a reason rather than a feeling.

Getting lucky looks like: an unrelated narrative pumped my bags, a listing announcement I never saw coming, the whole market ripped and lifted everything including my mediocre pick.

After twenty or thirty closed trades, sort your journal by those scores and look at where your profits came from. If they cluster in the high-luck, low-quality trades, you’re not truly winning. You’re being paid to build bad habits, and the market collects on those eventually. It always does. I would much rather know this about myself from a spreadsheet than find out the expensive way, and I say this as someone who has found out the expensive way more than once (because, again, I am a dumb dumb).

You have to be honest

None of this works unless you are brutally truthful with yourself.

“I bought because X was euphoric and I didn’t want to miss out” is a perfectly valid entry. So is “entered this at 1am on four hours of sleep”. They’re also some of the most useful sentences you’ll ever write about yourself, because six months later you get to see, in your own words, exactly what euphoria and exhaustion cost you. Not as a vague feeling where you go “yeah yeah I know I shouldn’t have done those things”, but an explicit number where you can go “oh… oof. ouch, that hurts, maybe I reallllllllllllly shouldn’t FOMO trade”.

Nobody reads your journal except you. There’s no audience to perform for, no followers to impress, literally no reason not to write the whole truth. Lying in your own journal is like lying to your doctor. Technically possible, and entirely self defeating.

Reading it back

Writing the entries is half the habit. Reading them back is the other half, and it’s the part where you actually get to reap most of the benefits of the process.

When I am being disciplined and journaling, my cadence is simple. I’ll take a quick look at the journal every morning. Then once a month I’ll go through and read the whole month in one sitting and look for leaks or patterns or anything that will help me improve future decision making. In poker we called recurring mistakes leaks, and knowing your own leaks is most of the battle.

The comments column is often where they live. Mine, straight from my own entries: I size too small on my best ideas, I struggle to cut positions even when the evidence and my gut are telling me to, and my worst entries happen first thing in the morning (usually because I see some narrative that has run over night, get FOMO, and before I have properly woken up or had my coffee, I have entered a dumb trade). I know all three because they’re written down, in my own words, over and over and over.

I’m always working on improving these things. It’s a constant battle for me, and it will be for you. Once you plug these leaks, you’ll have different ones in the future. One of my biggest leaks a few years ago was struggling to take profits, and fortunately I have gotten a lot better at that in recent years.

The template

To make starting as easy as possible, I’ve built a free template you’re welcome to copy, share, and make your own:

Trading Journal Template

It’s all pretty self explanatory in the read me page. You don’t have to use this, it’s just an example. Some people like doing this with a physical book. I encourage you to make it your own — even if you do copy the template, add things that you want to track, and remove anything that you’re finding unhelpful.

Some journaling tips

• Write the entry before the trade, or as soon after as you can manage. Hindsight contaminates everything it touches and if you wait too long, it’s hard to trust what your genuine thoughts were at the time of the trade.

• Make sure to use the comments section. Low sleep, too many tabs open, after a fight, after a big meal. They might seem like boring or inconsequential details today, but patterns will emerge over time.

• Score every closed trade twice: how well you played it, and how lucky you got. Let those two numbers, not the PnL, tell you how to feel.

• Make sure you’re honest with yourself every step of the way. This is for your eyes only, you’re only hurting yourself if you’re not honest.

• Read it back monthly (or regularly). Your leaks are in there, written in your own words.

Final thoughts

The goal was never to be happy when you make money and upset when you lose it. Anyone can be like that — it’s our natural, default state. It’s what everyone is. You’re not going to outperform the masses by behaving the same as them.

The goal is to be happy when you decide well, and a little uneasy when you decide poorly, whatever the outcome. Get there and the violent swings of crypto trading lose most of their power over you, and somewhat magically, the results tend to follow.

A journal is one of the best ways to achieve this goal. Not because writing things down is some mystical magical voodoo wizardry, but for the simple reason that your memory is a biased narrator, while a written record isn’t. You can’t trust your memory, but you can trust your past words (again, as long as you were honest with yourself) — I’m hammering this point home because of how important it is.

And speaking as someone who has lived both sides of the coin — both journaling and not journaling — in two different careers now: the journaling periods are better. Start today. Your first entry takes five minutes, and future you, reading it all back a year from now, will be grateful that you did.

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

I was a professional poker player for 15 years, and if I had to pick the single most valuable thing the game gave me, it wouldn’t be the money, or the freedom, or the stories, or anything else. It would be one concept: expected value.

Poker players shorten it to EV. Once it gets into your head, you start seeing it everywhere. It becomes the operating system underneath every decision you make.

So today I want to explain what EV is, how I apply it to basically everything, and also (because this part often gets skipped) where the framework falls apart.

What EV is

Expected value is the average outcome of a decision if you made the same decision an infinite number of times.

Simple example. I offer you a coin flip. Heads you win $200, tails you lose $100. Half the time you win $200, half the time you lose $100, so on average each flip earns you $50. The flip is +EV. You should take it, and you should keep taking it as many times as I’m silly enough to offer it.

Note what EV does not say. It does not say you will win. You might lose the flip. You might lose five in a row. You might lose 20 in a row! EV says nothing about any individual outcome. It only tells you whether the decision was good.

This is the part most people never internalise, and it’s the part poker beats into you whether you like it or not. In poker you’ll make the “correct” play and lose the hand constantly. All day, every day, for years. Eventually your brain rewires itself and you stop judging decisions by their outcomes and start judging them by their quality (the same concept is behind my popular Infinite Regret post from a few years back).\

The failure to do this leads to results-oriented thinking, and it is considered one of the deadliest leaks in poker.

It’s one of the deadliest leaks in life too.

EV in the wild

Once the concept settles in, you notice almost every decision has an EV attached to it, whether or not anyone bothered to calculate it.

Insurance is negative EV. It has to be, otherwise insurance companies wouldn’t exist. You pay more in premiums, on average, than you receive in payouts. I still buy insurance, and I’ll explain why in the section on where the framework breaks.

Lottery tickets are horrifically negative EV. Roughly half your money evaporates the moment you buy the ticket.

Wearing a seatbelt is one of the most +EV actions available to a human. The cost is two seconds of mild inconvenience. The payoff, in the rare cases it matters, is your life.

Asking a question in a group setting when you don’t understand something: +EV. Small social cost, occasionally large informational payoff, and half the room had the same question anyway.

Going to the gym, backing up your files, getting the weird mole checked, spending ten minutes preparing before an important call. All +EV. None of them guarantee anything. A gym session doesn’t promise you health any more than pocket aces promise you the pot. You are simply buying better odds, over and over, and letting the long run do its thing.

And of course, investing. Every position in my portfolio is, in my head, a bet with a probability distribution attached. In last week’s letter I ranked five tokens on forward looking expected value, and this is genuinely how I think about all of them. PRISM sits in my portfolio as a lottery ticket where I think I’m paying less than the ticket is worth. PHA is a patience trade where the odds are decent but the timeline is long. I don’t know which ones will work. I don’t need to. I need the collection of bets, sized correctly, to be +EV in aggregate.

You don’t need the numbers to be exact

A common objection: “but I don’t know the exact probabilities in real life.”

Correct. Neither do I. Outside of casinos and coin flips, nobody does.

The good news is you don’t need exact numbers, because the alternative isn’t precision, the alternative is vibes. Even rough EV thinking beats no EV thinking. When I estimate a 30% chance of something, the true number might be 20% or 40%. Fine. I’m still miles ahead of the person operating on “it’ll be right” or “it feels risky”.

Over time this becomes automatic. Picking a queue at the shops is an EV calc. Trolley volume, cashier speed, etc. Same thinking for deciding whether or not to switch lanes in traffic because one seems to be moving faster. Or for my golf-sympathetic readers, you could imagine an EV calc for deciding whether to go for it and hit a shot over water vs laying up, or hit a shot through the trees vs punching out.

These are all relatively small decisions, but multiply them by the thousands of decisions you make each year and the compounding is enormous. A person who makes slightly +EV choices across their health, money, career and relationships will, over a decade or two, end up in a wildly different place to a person who doesn’t. Same as the slightly better poker player over a million hands.

Where the framework breaks

This is arguably the most important part to understand.

EV thinking has many failure modes, and I’ve fallen into every one of them.

The first is ruin. EV is an average across many repetitions, but some bets you only get to make once. A coin flip for your entire net worth at 2 to 1 odds is +EV and you should basically never, ever take it if your net worth is anything significant. Averages are meaningless to the player who busted out. This is why bankroll management exists in poker, and why position sizing exists in investing, and why I try to never put ruin on the table no matter how juicy the odds look. Survive first. Be +EV second.

The second is false precision. Assigning a number to something makes it feel rigorous, even when the number came from the same gut feeling as the vibes it replaced. “I estimate a 65% chance this token 5xs” sounds analytical, but often it’s just wishful thinking and hopium. I catch myself doing this most with positions I’m already attached to, where the EV maths mysteriously always comes out in favour of the thing I wanted to do anyway. EV frameworks are only as honest as the inputs you feed it.

The third is unquantifiable value. Back to insurance. Negative EV in dollar terms, yes. But EV in dollars ignores the sleep I get knowing a disaster won’t wipe us out, and it ignores the brutal maths of diminishing returns: losing everything hurts far more than the equivalent gain helps. Some things are worth paying a premium for, and peace of mind is one of them. The dollar EV and the life EV of a decision are different numbers, and the second one is the one you’re supposed to be maximising.

And the fourth, the big one: some areas of life are actively damaged by the framework. I said this in Letter 56 and I stand by it: nobody wants a partner who thinks there’s a 94% chance they love them. Friendship, marriage, etc. There are things that go beyond EV calculations. The whole point of loving something is you’ve stopped keeping score. Run the EV on all of your relationships and you’ll end up with a spreadsheet and no one to show it to.

Putting it together

So here is how I’d summarise 15 years of poker and another chunk of years applying it to everything else:

1. Judge decisions by their quality, not their outcomes. You will make good decisions and lose. This is normal and it means nothing.

2. Rough probabilities beat vibes. You don’t need precise numbers, you need to be directionally honest about odds and payoffs.

3. Never bet ruin. A +EV bet you only survive one side of is usually a bad bet. Size everything so you always get to keep playing.

4. Watch for wishful thinking. If your EV estimates keep endorsing what you already wanted, the estimates are the problem.

5. Know where the framework ends. Money and decisions, yes. People you love, no.

The meta lesson, if there is one, is a framework is a tool and not a religion. EV made me a living at poker, and it shapes nearly every financial decision I make today. It’s also, deliberately, switched off for the parts of my life I care about most.

Try and be an investor, not a gambler.

And with the people around you, try and be neither.

As always, thank you for reading, and leave any questions or comments below :)

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

All five have featured in recent letters. $VVV (and its sibling $DIEM), $PHA, $CARDS, $PRISM and $ETHFI.

I'll look at where they were when I wrote about them, where they are now, what happened in between, and what I think happens next.

$VVV and $DIEM (Venice AI)

When I covered it: twice. Letter 102 in February at $6.64, then a full update in Letter 112 in mid May at around $17.39.

In Letter 112 I told you I was taking some profits off the table while keeping a long-term bag of both VVV and DIEM. Part of the reasoning: when the biggest podcasts and traders all start talking about something at once, it tends to mark a local top rather than an early entry.

Now: VVV sits around $11, market cap near $550m. Down roughly 33% from Letter 112, and still up roughly 76% on Letter 102. It printed a high of about $21.32 on June 3 before a significant retracement to where it is today.

A large part of that is due to the news that Venice closed a $65m Series A led by Dragonfly at a $1 billion valuation, structured as equity plus token warrants, with no sale from the treasury’s 30m+ VVV. Erik Voorhees has been on X defending the terms after critics argued the deal underprices the token. The debate worth watching is whether future value accrues to equity holders or token holders. Every crypto project raising equity faces this question eventually, and Venice is now facing it in public.

The supply side keeps tightening regardless. Emissions drop to 3m VVV per year this month, down from 10m at launch. The revenue-funded buy and burn continues, with burn rates doubled in April, and Venice V2 with text-to-video is on the roadmap.

DIEM sits around $1,310 with a $49m market cap, off its May 20 high of about $1,927 (near where it traded at the time of Letter 112). The mechanism is unchanged. One DIEM equals one dollar of Venice API credit per day, forever, minted only by locking staked VVV. At current prices the market pays roughly 3.7 years of compute upfront for perpetual access. New DeFi rails have appeared around it since the letter, including liquid staked DIEM (wstDIEM) and collateral markets on Morpho.

Looking forward: of everything in this letter, Venice has the most complete flywheel. Real revenue, shrinking supply, a second demand sink in DIEM, and now $65m of runway. The profit-taking call from Letter 112 aged well, with the price a third lower since. The pullback from $21 to $11.70 looks like froth coming out after a spectacular run from December. Still bullish, and still holding a solid position in both VVV and DIEM.

$PHA (Phala Network)

When I covered it (Letter 113, late May): around $0.03 to $0.04, market cap in the $30m to $40m range.

I did not pitch it as a high conviction buy. I framed it as a small, asymmetric bet on the confidential AI narrative, with a long time horizon and no expectation of a quick rerate. I told you to watch the metrics rather than the price. Token throughput on Phala Cloud, enterprise customer count, GPU TEE capacity.

Now: the token sits at around $0.026, market cap near $22m. Down roughly 25 to 30% since the letter, and it shed close to 14% in the past week alone.

The frustrating part is the business keeps doing business things. Since I wrote about them they added DeepSeek V4 Flash and updated Qwen models to the confidential compute catalog, and continued repositioning from crypto privacy project to enterprise AI infrastructure company. None of it helped the token.

Which brings me back to the core tension from the original letter. Phala’s revenue is mostly off-chain. Enterprises pay for confidential compute with credit cards through Stripe, and almost none of it routes to PHA holders through any on-chain mechanism. A real business sits next to the token rather than underneath it. Until something changes there (a buyback, a fee switch, anything connecting cloud revenue to the token), PHA trades on narrative alone. And right now there is no narrative bid (especially with Venice struggling and attention flowing to other sectors). I flagged the history of compute tokens like Render, Akash and Filecoin in the letter, all of which built real products while their tokens bled. So far PHA is following the script.

Looking forward: the thesis is intact and the token is cheaper. If the AI privacy narrative catches a bid, or the team announces a revenue-to-token mechanism, a sub-$25m market cap against a ~$3m ARR business rerates quickly. Without either of those, I expect more drift. This one is a patience trade, and patience trades test you.

$CARDS (Collector Crypt)

When I covered it (Letter 115, early June): around $0.21, market cap around $55m.

My take at the time: owning a tokenized card is a bet on the card. Owning CARDS is a bet on the business of selling packs. These are two different bets, and people are often confusing them. The bull case was the near 1x revenue multiple. The bear case was supply, with roughly three quarters of tokens still locked. For the airdrop hunters I pointed at the pre-token platforms instead, Phygitals, Beezie and Courtyard.

Now: the token is at around the same price, though it did hit an ATH of around $0.38 a few weeks ago — making for a pretty clean almost 2x if you took profits near there.

It’s been an eventful month. On June 23, Maelstrom (Arthur Hayes’ family office) published a bullish report on CARDS with a $4 price target (absolutely ludicrous target) but which still sent the token up 16% in a day and lit up the timeline. On the product side, a partnership with Loopscale will let people borrow USDC against vaulted card NFTs, and the team has flagged an on-chain tokenized cards index plus expansion into One Piece and sports cards.

The bear case from the original letter has not gone anywhere though. Roughly a quarter of supply circulates. There were token unlocks on June 29, and the next unlock is on July 29 and releases about 28.8m tokens (roughly $6m at current prices, 1.4% of total supply), with plenty more behind it. The gacha mechanic also lives in a regulatory grey zone between gaming and gambling, and the whole thesis leans on pack-ripping volume holding up.

That said, volumes are regularly hitting new highs and the TCG sector is still on fire so there’s still a lot of momentum behind it.

Looking forward: revenue remains real and the P/E style maths remains compelling relative to most of crypto. My read is the same as it was. You are buying a genuinely good business through a token with heavy dilution ahead of it. A good token to keep on a watchlist and buy when there are significant retracements and sell when the timeline gets overzealous about it.

$PRISM

When I covered it (Letter 116, mid June): $242.91, with the market cap in the low single-digit millions.

I called it watchlist material, or one to take a longshot punt on. A small, eyes-open bet for people who enjoy cool onchain mechanics and funky new defi innovations and experiments.

Now: the token sits around $294, up roughly 20% since the letter.

The token had already been on a wild ride before the letter went out. It peaked above $1,100, then retraced hard, and I published with the price at ~$242. It’s been on a wild ride since too, yo-yo’ing all over the place. Fortunately, the market seems to be appreciating something about it and supporting the price for now at least.

Looking forward: Spectrum V2 ought to be launching any day now and a lot hinges on how things go there. It looks promising, but this is still a typical assymetrical bet for me — probably zero, but a good enough chance for it to be a hero.

$ETHFI (EtherFi)

When I covered it (last week’s neobank letter): around the $0.30 mark.

I said if you want token exposure to the crypto neobank sector, consider both ETHFI and AVICI. The historical buybacks were the reason for ETHFI, plus the team flagging more token value to come in some other form.

Now: around $0.42 to $0.44. Up roughly 35% in the week since the letter.

I got lucky on timing here, plain and simple. Any time you are up 35% in a week there’s bound to be some luck involved, but hey, we’ll take it. The honest version is I wrote about a token near its lows with improving fundamentals, and the market happened to agree seven days later. Sometimes it’s 7 months later, sometimes it’s never.

What the market is waking up to is the value accrual machinery I covered. The business behind it holds up. DeFiLlama puts annualised protocol revenue around $34m, out of total fees above $150m. Messari’s February valuation report had a stat worth repeating for a neobank letter follow-up: Cash is now the largest revenue line at roughly half of protocol revenue, so the card already out-earns the staking business.

TVL sits in the billions, second among liquid staking protocols behind Lido, and the Cash migration to OP Mainnet moved $220m in TVL and 70,000 live cards without a single declined payment. As spending products go, this is the one with the most protocol behind it.

Looking forward: With around 93% of tokens in circulation, the unlock overhang is mostly nonexistent — the main reason I was enticed by the token last week. Of the crypto neobank plays with a live token, this remains my pick for cleanest alignment between business and holder. A 35% week does not change the thesis, it just makes it slightly less appealing to buy cause of the higher price. I bought more just yesterday though, so I think the price is still reasonable.

Final thoughts

Writing these recaps is always going to be a mixed bag. Anyone with a newsletter will happily show you their winners. The useful information lives in the full ledger, including what I said at the time, so we can look at our losses alongside our wins.

If I had to rank them on forward looking expected value today, knowing everything above: Venice and ETHFI share the top spot, CARDS comes next with the dilution asterisk in big bold font, and PHA and PRISM bring up the tail: PHA as the patience trade, and PRISM last as the lottery ticket.

As always, none of this is financial advice. I hold positions in some of these and my bags shape my views no matter how hard I try to stay objective. Do your own research, size for the possibility of zero on the small ones (or on all of them tbh), and be real with yourself about which of these is an investment and which is a punt.

In other words, try and be an investor, not a gambler.

As always, thank you for reading, and leave any questions or comments below :)

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

I first wrote about neobanks back in November of last year, so I won’t rehash the old other than to say, I was bullish then, and I am bullish now.

I think that anyone in crypto not using a crypto card is almost certainly leaving money on the table. Aside from offering extremely competitive cashback and other rewards, many of them are also running points programs that will add token rewards on top of everything else.

Most cards seem to market around their cashback %, but to me, that’s disingenuous. The reality is that it doesn’t really matter whether a card has 1%, 2%, 4%, 6%, 8%, or 10% cashback if there is a cap on spending per month and you’re going to go far over the cap. You also need to take into account any subscription fees for the card.

So, with the power of AI, I built a tool to help find the best card based on your expected monthly spend, taking caps and fees into account. It’s never going to be perfect because of the many other intangibles you need to consider: points programs, travel perks, free or discounted subscriptions, insurances, etc. You also need to look at the other fees: FX fees, ATM withdrawals, crypto-to-fiat conversions, and so on.

Still, I think it’s helpful to evaluate the cards based on the expected monthly cashback as the starting point, and work from there.

You can try it out here: Crypto Card Finder.

That’s one way I have been thinking about the neobanking space lately, but there are three main ways I am looking at things:

1. Investing in existing tokens

2. Farming projects with upcoming tokens

3. Finding the best card(s) to use in terms of real cashback %

Investing in existing tokens

The neobanking sector is still small. A look at the category on coingecko shows a total marketcap of ~2.3 billion, with over half of that coming from Mantle alone.

There aren’t too many tokens here that I am overly excited to be buying into. The majority of them have massive token overhangs, with an FDV of 2-4x their marketcap.

The two that stand out are EtherFi (almost fully diluted) and Avici (fully diluted). While I am bullish on Tria, I am also an investor, and I know my tokens are going to start vesting in a couple of months along with many others, so it is hard to recommend others buy into the token at this point in time. Plasma falls into a similar category, with roughly 75% of their tokens unvested and yet to hit the open market.

I recommend people use those apps and cards because they both offer good products with solid cashback and other rewards, but I would be careful about ploughing money into either token, and consider both ETHFI and AVICI to be slightly more appealing to hold from a pure tokenomics perspective.

EtherFi

EtherFi is probably the most defi-native protocol and card out there. It is an excellent place to park funds that you can earn yield on, and even borrow against, as well as get a card you can use for everyday spending.

The ETHFI token sits at a marketcap of ~$300m and is almost fully diluted. The team has spent over $15m on token buybacks, and has proven to be mindful of making decisions that help support the token price (despite the trajectory of the token, which I think can be largely attributed to the broader market).

They have paused buybacks for the time being but have said they’re working on a new plan that will bring even more value for the ETHFI token, TBD what that looks like and if it involves a different form of buybacks. Something to keep an eye on.

This is a great article if you want to dig deeper into the numbers behind EtherFi and their ETHFI token.

Avici

Avici is a Solana based neobank that shares a lot of similarities with EtherFi in that it’s also very defi-native and crypto-forward. It allows you to retain full ownership of your funds and even borrow against them while spending money via one of their Visa cards, just like you would with a bank account.

AVICI (the token) was launched via MetaDAO in October 2025, and thus, the ownership is actually held by a DAO. The company itself had been around for about 2 years prior to that, with the founding team determined to build around the token in a way that aligns incentives in a way that benefits all parties.

This article by co-founder RamXBT is a must read for anyone considering buying into the token, and I wager it would be hard not to want to be a part of what they’re building after you read it.

Sitting at a marketcap of “only” $13m (fully diluted), this seems like the value buy token to me. It has a lot going for it, and their monthly volumes are up and to the right.

Farming projects with upcoming tokens

My general philosophy when farming airdrops is to use products fairly organically, and to not go too far out of my way doing mental and financial gymnastics to try and get an airdrop.

That can sometimes work if you pick the right protocol and the timing is right, but the majority of the time, you end up wasting time and capital.

When it comes to crypto cards, I think it makes sense to try a variety of them and see what suits you best — and it’s probably a good idea to try and put a little bit of volume through on the ones that have likely airdrops coming up.

If you like any of them, and the fees and cashback seem reasonable, then making one of them your primary card could lead to a very lucrative airdrop.

There are well over 50 crypto cards at this point and it seems like there’s a new one popping up every other day. It would be crazy to try and catch them all, so instead, here is a shortlist of three that I believe to be some of the best ones to use if a potential airdrop is something you’re after:

1. XPlace — They have announced and planned a token and are running an XP program, but there is no TGE date set as yet.

2. Bleap — A token is confirmed but there isn’t a points program, this might mean you’re very early (or that they’re not planning on running points). TGE is expected in early 2027.

3. KAST — A $KAST token is confirmed on the roadmap with a points program running, but the token has been delayed from an initial launch date of Q2-Q3 2026 with no new launch date given.

Many of the largest card projects already have tokens (EtherFi, Tria, Plasma). Many others are unlikely to ever launch a token. It’s a fine line trying to find projects that have a chance at becoming large enough for their token to be valuable, and that haven’t yet launched a token.

There will likely be more that pop up in this sector in the coming months, and I think it’s very much worth keeping an eye on.

Finding the best card(s) to use in terms of real cashback %

Cashback is not the be-all-and-end-all for cards, but I do find it to be a great starting point for evaluating rewards. Some of the other rewards are great, like airline lounge access, AI tool access or discounts, travel insurance, hotel and flight discounts, and so on. But they’re all so intangible — some people might use all of them, others might use none of them. It’s hard to compare apples to apples.

The same goes for fees. For some people, FX fees are a major consideration. For others (aka those living in the US or spending in USD), it is not. For some people, yield on the money you have in the protocol matters. For others, it is insignificant.

So, as mentioned at the top, I built a tool to help rank cards based on their real effective cashback percentages. The site has all of the data for everything else so you can compare the things that mean the most to you, but the starting point is cashback.

Notably a lot of companies will advertise high percentages but then have a $999/month fee. Or a cap on how much you can earn in cashback (most have this). Or they’ll pay out in a non-stablecoin.

Crypto Card Finder aims to take all of this into account. Note it’s still in the early stages, so if you have any feedback, please let me know and I will be regularly trying to tweak and update the site to make it the absolute best place to go to find out what card(s) offer the best cashback.

That’s a wrap for this week, hope you found this newsletter as insightful and helpful as ever.

p.s. thanks once again to everyone who reached out after my last two letters, I still owe some of you a response, and will be getting back to everyone soon. I appreciate you all 🙏.

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

My brain was totally burned out. Despite not having done anything overly productive or “work-heavy” for a while, my brain was still sending me all sorts of alarm bells and warning signs that it needed a break, that it was overloaded. Overloaded on dopamine.

You see, for the last however many years, the majority of the time I have been going through life trying to fill every waking minute with some source of dopamine. I couldn’t go for a walk or drive a car without a podcast in, I couldn’t eat a meal without a TV show on, I couldn’t do anything without some secondary or even tertiary form of stimulation. Even when I was doing “good things” like rock climbing or playing golf or working out at the gym, I was never doing those things alone. I was talking with others, or I was listening to a podcast. I was never alone with my thoughts.

Something had to give, and so for the last three days I:

• Listened to zero podcasts

• Spent no time behind a computer screen

• Went for walks without my airpods

• Kept my phone on grayscale mode

• Deleted YouTube and Instagram from my phone

• Deleted the apps on my phone that let me check crypto prices

• Pulled zero gacha machines

• Opened zero new TCG packs

• Played zero video games

I’ll admit I did eat some junk food (had a few Tim Tams and two ice creams, sue me). I did watch a movie with my wife. It wasn’t a “pure” 100% true dopamine fast in the absolutely-zero-screentime and absolutely-no-dopamine-sources sense of things, but it was good enough for me.

So what did I do instead? I read a real, physical book (Tress of the Emerald Sea by Brandon Sanderson, great book!)

I spent time with friends and family and was fully present.

I cleaned the house, caught up on chores, and got all sorts of little things in order that had been piling up.

I took my dog for more walks.

I went for drives and didn’t listen to music or a podcast.

I sat with my thoughts.

I was BORED, while not actually being bored.

That last point is the thing I want to highlight and focus on.

Subscribe now

The importance of being bored

I didn’t quite realize how important it is to be “bored”, or rather, to spend time alone with one’s thoughts. Not only are we finding our lives jam packed with more and more content these days, we’re ALSO having less and less time to ourselves to process the content.

In ye olden days, I would be reading a book and then when I was away from the book, I would spend some amount of time thinking about the book and what I read. I did that again this weekend for the first time in god knows how long, and it made me realize how crazy it is that this part of life had all-but disappeared.

Instead of thinking about the book I was reading, or even the TV show I was watching, the video game I was playing, the whatever I was whatevering, I was immediately jumping into something new. No time to process.

It’s no wonder I can barely remember the characters from books or TV shows any more. I consume a piece of content and then immediately jump into a new piece of content, with approximately zero time to think about anything.

This constant jumping from one thing to another has left my memory in tatters and has had me questioning my sanity.

The antidote, I believe, was and is so simple: just do nothing for a while. Sit still, be idle, and be alone with your own thoughts.

Understanding addiction

This part might not apply to as many people, but it applies to me, and it will apply to some of you, and that makes me feel like it’s worth writing about.

Addicts are addicts. That might seem like an obvious statement, but it’s not always so obvious. Lots of people have “addictive personalities”, but there’s a difference between that, and being an actual addict. Being someone who is unable to prevent themself from causing themself significant harm by overdoing something.

Typical addictions are alcohol, drugs, gambling. Oftentimes an addiction can start and end with a singular vice, but oftentimes it’s not so simple. If someone is addicted to one thing, they are infinitely more susceptible to being addicted to other things.

Dopamine addiction is sort of a catch-all term for being addicted to all the types of things that provide cheap dopamine. As I mentioned in my previous post, it is particularly nefarious because there are so many things that fall into this category in today’s world, and it’s virtually impossible to completely eliminate them all from your life.

So, as addicts, we need to figure out how to navigate this brave new world, and live with a type of balance that we could never find from our other vices. I know I can never drink alcohol, because I am simply unable to do so in moderation. The “solution” is simple (not easy, but simple): never drink alcohol. It’s not so simple when it comes to dopamine, a naturally occurring chemical and one that nobody can (or should/would want to) eliminate entirely from their lives.

So I have to learn how to moderate my intake of dopamine from content, the internet, food, games, and all the other enjoyable things life has to offer.

I fear it will be a lifelong battle for me, but that’s okay. Everyone has their things they have to deal with. In the grand scheme of things, I am still extremely fortunate.

My strategy, for now, is to entirely eliminate the most dastardly sources of cheap dopamine from my life: short form video content (youtube shorts and instagram reels), as well as video games.I think that without these two categories of dopamine in my life, I will be able to manage the rest.

In an ideal world, I find a way to at least bring back video games, in moderation. They do bring me a lot of joy. We shall see though.

The other important thing for me, and I think this one applies to everyone, is to find ways to be bored again. To not fill every waking moment with external stimuli. To learn to sit with myself, to listen to my own thoughts, to appreciate silence, and to use my brain the way it’s meant to be used: to think.

I want to end by thanking everyone who reached out after my last post. There were a bunch of public comments, and I received a whole lot of private messages too. It seems I struck a nerve, and there are a lot of people out there going through similar things. Every now and then I write a piece that resonates like that one did, and it feels good on multiple levels. It helps me feel like I’m not alone, and it feels good to know that other people feel less alone by reading my experiences.

As you can surmise from how I am writing about all of this, I am figuring things out in real time. I feel like I made some solid progress over the last few days, but I’m also coming to terms with the fact that there’s a lifetime of constant vigilance ahead for me.

If you’re struggling with anything similar, please feel free to reach out to me if you want to chat further. I think everyone could take a few small steps that will go a long way to finding some more peace and balance in life.

And it all starts with allowing yourself to be bored once in a while.

Leave a comment

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

This post is sponsored by OndoPerps

Last week I got early access to OndoPerps and have been playing around a bit on it. It’s a clean product, super easy to use, with a great depth of options for trading rwa stocks and commodities onchain.

On their platform you can trade equity perps 24/7 with up to 20x leverage, unmatched liquidity, and tokenized stocks as collateral.

In line with the theme of this post, I decided to short Netflix because I think more and more people are waking up to the idea of wanting to escape the matrix (the infinite dopamine machine that tech company algorithms are holding over our lives), and get back to “the real world”.

And no, the irony is not lost on me.

Sign up here to check it out (using my link/ref code will get you a discount on trading fees): https://app.ondoperps.xyz/?ref=DDGEP7

My Name is Zeneca, and I am a Dopamineholic

My brain is fried. Burnt. Toasted. Frizzled. Frazzled. Cactus.

I have been feeling the degradation for a while but in recent months and especially the last few weeks, it has gotten so much worse. I can’t hold attention on anything. I can’t focus on tasks. I can’t complete anything. I can’t follow through with commitments. I am watching myself turn into a potato in real time, and I can see it happening, I know it’s happening, and it is still nigh impossible for me to stop it.

I think I am addicted to dopamine. I have always had an addictive personality. It has manifested throughout my entire life, from reading, to video games, to trading cards, to poker, to alcohol, to golf, to trading, to writing, everything. When I get infatuated with something, I go all-in, get ultra obsessed, and find it difficult to focus on something else.

It’s both a superpower and a curse. It has benefited me greatly at times, and it has brought me to the brink of ruin. Whatever “normal” people have in their brains that helps them regulate and find “balance”, I do not have. I find it so much more difficult than most people to say no to that next drink, or to say no to that next hit of dopamine, wherever it’s coming from.

I don’t know how to fix it either. I have tried many things lately: taking a break from all screens, no phone in the morning, greyscale mode on my phone, reading physical books, doing puzzles, walks without my phone, driving without listening to a podcast, eating without watching a TV show, deleting all video games from my computer, etc etc. They are all temporary fixes, and they do help, but I always find myself succumbing to the lure of the infinite dopamine machine and falling back into my old habits again.

The world is working against you

Dopamine is insidious. We talk about it like it is the pleasure chemical, but it is closer to the wanting chemical. Dopamine fires before the reward, not after, and so it leaves us in a constant state of anticipation, of being pulled towards the next thing, the next hit.

Your phone knows this, and the people who built every app you open understand it better than you or I ever will. The infinite scroll, the pull to refresh, the little red notification dot. None of it is an accident. It works the same way a slot machine works, and the payout is a reward you never fully predict (or even understand).

Variable rewards are the strongest hook we know of. A pigeon will peck a lever far longer when the food comes out at random than when it comes out every time. You and I? Sorry to break it to you, but we are pigeons.

When you flood your brain with cheap stimulation all day, your baseline for pleasure climbs. Quiet starts to feel like boredom, and a walk without a podcast feels like wasted time, or worse, it feels scary because you might have to be alone with your own thoughts.

A book becomes hard work. I noticed this with my own attention over the last few years. I used to read for hours. Now I feel a pull to check something after a few pages. My brain got trained, and not by me.

Your brain needs to be bored

Here is what I have come to believe. All of my best thoughts come from when I am doing nothing in particular. Walking, sitting in the shower, lying in bed before sleep.

There is a network in your brain called the default mode network. It switches on when you are not focused on a task. It is where a lot of your reflection and creative connection happens. Fill every idle moment with a screen and you never let it run.

You are trading the long term for a hit of nothing in the moment. It is a negative EV play you make a hundred times a day, except in the moment, it feels like you’re making a lot of little micro wins cause of the instant pleasure you’re getting.

Insidious.

What’s the fix?

Honestly, I kinda don’t know what to do at this point.

This is not a new problem. People have been talking about dopamine issues for years. There are infinite resources and ideas out there for how to fix things: leave your phone outside the bedroom, use apps that limit your screen time, block out time to be offline, no phone for the first hour of the day, yadda yadda yadda.

Those are all well and good but it’s one thing to know what to do, it’s another to be able to implement the solutions and then stick to them.

When you’re addicted to dopamine and your brain is fried, you’re in the worst possible state to be able to fight back. The tools you need to fight with (your brain, your willpower) have been weakened by the very thing you need them to fight against.

I honestly feel exactly the same way I felt when I was dealing with my alcohol addiction. It took me 15 years to kick that addiction. I’m not going to say that this is more difficult overall, but in some ways it definitely is.

You see, with alcohol, you can quit it in its entirety. I haven’t had a drink in 6.5 years, and I don’t ever have to have another drink for the rest of my life. But dopamine? I can’t “quit” that. I can’t eliminate all the sources of dopamine (nor do I want to). Perhaps I can eliminate all sources of “cheap” dopamine, but even that is tough, because we live in a world that all-but-requires the use of social media if your work is in the online realm.

Honestly, I am at a loss.

I’m going to try a 3 day break from screens and cheap dopamine and see if that helps heal me enough to be able to put in place some guardrails and then stick to them.

For the next 72hrs I won’t be using a computer. I will only use my phone for the most basic chat and actual telephone purposes, I won’t watch TV, I won’t play video games, I won’t listen to podcasts, I won’t eat junk food.

I did a similar thing a week or two ago for about 12hrs and it helped, but clearly not enough. So this is trying the same thing but taking it to the next level.

I think this is not an uncommon feeling out there. Maybe I am in a bit of a worse place than most right now, but speaking with friends and strangers-on-the-internet lately, I get the sense that almost everyone is feeling some level of this in their souls.

Anyway. Wish me luck. I’ll see y’all in 3 days, back with another Letter to kick next week off. I won’t be in Discord, Telegram, or on X until then. I won’t be playing video games, I won’t be trading, I won’t be behind a screen.

Time to try and enjoy the real world again, and heal my brain.

Hopefully 🤞.

Get $20 in free BTC by signing up to Swyftx here. Australian & NZ residents only.

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

Ultimately I decided to write this, because Prism is a cool concept, doing novel and interesting things onchain. Whether or not there’s value in buying the token is up for debate, a debate that we’ll evaluate from multiple sides during today’s Letter.

But simply being able to share a cool project with y’all is enough for me. I had a lot of fun researching this Letter, I hope you enjoy reading it.

Prism is the work of an anon who goes by 0xsolazy, aided heavily by another anon who goes by Colby. It went live on Ethereum in late May, and as I write this the token sits at a price of ~$240 with a market cap of ~$1.2m, having previously hit an ATH of ~$6m. PRISM is the base token of an entire, growing ecosystem, with the idea being that the products and services being built will flow back to support the token price in various ways.

How the token works

PRISM is a Uniswap v4 hook. If hooks are new to you, the tldr is that v4 lets a developer attach custom code to a pool so something fires every time a swap goes through. Most projects use a hook as a small helper sitting beside their token, whereas PRISM does the stranger thing of making the hook itself the token.

They’ve also added an NFT layer to it. Every whole PRISM you hold mints you a matching NFT (there is a cap of 5000 NFTs), and each of those NFTs is a 1/5000 slice of the same single liquidity position. Fees drip down to the NFTs in proportion to how many you hold.

Claiming the fees is open to anyone. They pile up inside the pool, and a single call from any wallet pushes them out to whoever holds the NFTs at the time, split by share, and nobody needs permission to do it.

Where the idea comes from

If this is ringing an ERC-404 bell, good, because the bones are the same. ERC-404 was an early 2024 fad where people fused a fungible token and an NFT collection so the two mirrored each other. PRISM runs on the same mirror trick in its tidier modern form, DN404. It is laid out across a handful of small contracts handling the hook logic, the token-to-NFT syncing, and the art drawn onchain.

The difference from the old 404 projects is the part I like. Most of those were NFTs cosplaying as a mechanism, where the picture was the whole point and nothing sat underneath it. Basically just memecoins with extra steps. PRISM turns the relationship around where the fee-earning slice of the pool is the point, and the NFT is used to assist in keeping track of the ownership and distribution of fees.

The scarcity only moves one way

5000 is the most NFTs there will ever be. Whenever someone sells a fractional amount of PRISM, the NFT underneath it burns for good. Fewer NFTs alive means a bigger fee slice for everyone still holding a whole one.

In theory it’s all well and good, but in practice it only matters if there are actually fees to be split. This is the biggest X-factor, and this project will live or die by whether or not the ecosystem is able to generate fees. So let’s look at the various ways that the protocol can currently make fees, and also look at how much has been generated to date.

Spectrum, aka index funds for onchain tokens

This is where things start to get very interesting for me. The base token PRISM is the value capture layer for everything. On top of it sits Spectrum, an index launchpad. Underneath both runs dStable, a yield-bearing settlement rail. Put them together and the pitch grows past hold = earn into something I find more interesting. It is a way to turn a crypto narrative into one thing you buy in a single click.

Let me set up the problem first, because it is one I have lived. Crypto is brilliant at minting narratives. AI agents one week, RWAs the next, then memecoins, then DePIN, then gaming, then Base szn (again), then socialFi, then TCGs, then whatever Twitter decides to care about next. The trouble is being dead right on the theme does not stop you picking the wrong token.

You can be right about the Base AI rotation, then you have to choose between Venice, Bankr, Virtuals, Nock, Rei and a handful of others, and one of them does a 20x while the rest flounder (or simply don’t perform nearly as well). So you end up holding eight or ten tokens to express one view, attention rotates between them, liquidity splits across all of them, and the whole exercise becomes a knife fight between people who already agree with each other.

Traditional markets sorted this out a long time ago. You want a sector, you buy the basket. Onchain has tried the same trick plenty of times and the attempts kept dying, and it is worth knowing why before you assume this one is any different. The old index tokens were two separate things stitched together, a vault holding the coins and a thin little market trying to price the wrapper, and the seam between the two is where they came apart.

You paid a management fee for the privilege, with the category's flagship charging close to one percent a year, roughly 20-30 times a plain S&P fund. The liquidity was rented with token rewards, so the day the rewards stopped the depth walked out and the spreads blew open. And because the wrapper traded in its own thin market, the price drifted above or below the coins it was meant to track, so being right on the call still left you buying rich or selling cheap.

Spectrum's fix is structural. The index trades natively on Uniswap with no separate wrapper market, and its price is bound to the weighted value of the basket by construction, so it does not drift to a premium or a discount. The depth does not walk away when incentives dry up, because the depth is the token itself. And rather than charging you a yearly fee to hold it, the index pays you a cut of its own trading fees for as long as you do. Same one-click, whole-sector exposure an ETF gives you, with the carry running in your direction for once.

Here’s how it works under the hood. Each Spectrum index is its own ERC-20 token, live on Ethereum and Base, a weighted basket with its own v4 hook and a pool paired against dStable, and a redeemable claim on the assets it holds. You buy in with ETH, USDC or dStable (more on dStable soon), and the hook routes your money into the components at their set weights. You get one token back, and it represents your proportional slice of the whole basket. When you want out, you burn the index token and take your share of the real assets at NAV. Own one percent of the supply, redeem one percent of the basket.

The hook handles the buying, the minting, the redemption and the routing, so you skip the spreadsheet and the eight separate trades.

One caution worth keeping in view. A basket takes the pain out of picking the one right token from ten. It does not take out market risk, since a bag of volatile tokens is still a bag of volatile tokens, and an index holding ten low marketcap coins in a new meta will swing wildly.

Now the fee model. Every mint and every redemption charges a one percent fee, and it splits three ways.

Read the creator line again. If you put together a useful basket, you earn thirty percent of its fee flow for as long as people trade it. Your top picks become a product with a revenue attached. Become an expert in any narrative or sector and launch your own index and earn fees for as long as the index is traded. This turns curation into a real business model, and it is the cleverest social hook in the whole design. We haven’t seen any indexes really take off yet, but all it takes is one for the flywheel to pop off and every KOL under the sun to jump on the bandwagon.

So how do you launch an index?

The process is a little complex right now, which might be why we haven’t seen too many launched. Another reason is the fee. It basically costs 0.1 ETH to launch a token (technically there’s a dutch auction every 10 blocks from 1ETH → 0.1 ETH, but I can’t imagine so many tokens launching that anyone would do anything but wait and pay the 0.1 ETH).

The fee is also yet another way that the protocol supports their token, with 90% of the revenue from index launches going to buying and burning PRISM.

Anyway, to launch a token, basically you go to the launch page on the Spectrum site and follow the instructions in their deploy guide. The gist of it is that you select your tokens and the weights you want for them all and that should be it, but it’s a little clunky and can require some troubleshooting.

The good news is that there’s a V2 of Spectrum coming out soon which should make this entire process a lot more user friendly, so it should look something like this soon:

As they say in this tweet, their mission with V2 is to make it as easy as possible to launch an index with a new and improved launch flow that strips away all complexity.

Okay let’s talk about dStable

The third piece of the ecosystem is dStable, and it is the least glamorous of the three, which is usually a healthy sign. It is a yield-bearing stable asset, and its job is to be the clean unit everything settles against. Without it, every index has to quote and route through messy ETH paths and multi-hop swaps. With it, Spectrum prices against one steady rail, which makes the system cheaper to use and easier to build on.

The flow is simple. You deposit ETH, the system converts it into stablecoins like USDC and USDT, parks around 95 percent into Aave to earn yield, and keeps a small buffer idle for redemptions. When the yield is harvested it splits three ways. Seventy percent flows to dStable holders, twenty percent buys and burns PRISM, and the last ten percent goes to a dev treasury.

So dStable pulls double duty. It hands you a yield-bearing dollar on capital you parked anyway, and it quietly becomes a second burn engine for PRISM, one that runs off Aave yield rather than trading activity.

There are a lot of impressive elements to how this all works under the hood, and I strongly recommend reading this post by Colby to learn more:

If you can’t be bothered to click and read the whole post, these last few points are at least worth knowing:

Giiifting, and what else might come

I mentioned at the top that PRISM was the base token of a growing ecosystem, with more products and services that will launch and be built on top of it. Giiifts are another one of these, and will be launching soon.

While I think the idea is cool, I’m a little uncertain about how much usage it will actually see. I think most likely once it launches a bunch of people will play around with it and have some fun, but then move on to the next thing, and it won’t gain escape velocity and become a mainstay service. I hope I’m wrong though, and there is a moonshot chance that it really takes off and becomes a narrative in its own right.

Either way, the idea is that using it will benefit PRISM:

Whether or not this idea works, it’s more the fact that because of the interoperability and composability that we all know and love about blockchains, anyone can build anything they want on top of the Prism ecosystem.

If more builders get excited by the infrastructure, things can really take off.

How’s the revenue actually looking?

A great site for tracking the Prism ecosystem is PrismBeat.

The lifetime revenue is not bad, but most of that came when the token was trading higher and the consensus was a lot more bullish. Volumes have slowed way down, and there was only $3,500 in revenue over the last 24hrs.

I say “only”, but the math actually works out pretty decently even at these low numbers, with each PRISM token earning its owner ~$0.70 in revenue for the last 24hr period.

70 cents doesn’t seem like much, but when you multiply it by 365 you find that it’s actually bringing in $255.5/yr for every PRISM held. Considering the price of one token (currently $242.91 as I write this), that’s a pretty damn good return, even at low volumes.

If you could be confident volumes would stay at these levels or go higher, then it seems like a pretty obvious buy. We can’t be confident about that though. Things can certainly stay here or go much higher, but as with all things in crypto, they can go much lower too.

One last thing to note is how little revenue has come from the Spectrum / index side of things so far, only around $5k. If that product takes off, then the flywheel is really going to start going crazy. If not, then maybe/hopefully one of the other products does. If not, well, then things start to look a little more bleak.

The bull case for PRISM

PRISM is built to soak up value from the whole system through buy-and-burn pressure as well as fee sharing, and right now it arrives from three live taps, with room for more.

The three live taps are the start. The whole reason to build Spectrum and dStable as separate products is that each new thing bolted onto the system arrives with its own fee, and each fee routes a slice into the same burn. Gifting, once it ships, is a fourth tap. The product after it is a fifth. The more the ecosystem grows, the more lines feed the same sink, and that compounding is the part of the design I find most interesting from a value-capture angle.

It’s sorta like how PunkStrategy works, where every Strategy launched afterwards had a way of tunneling value back to the initial PNKSTR token.

Stack all of the taps on top of the scarcity ratchet from earlier and you get one hell of a potential flywheel. More indexes launch, more people mint and redeem, more fees flow, more PRISM gets bought and burned, fewer NFT claims survive, and the holders left standing own a bigger share of everything.

PRISM is the asset that all captured value flows into.

The bear case for PRISM

While in theory everything sounds pretty freaking incredible for this project, there are many very real, very valid concerns:

1. The biggest and most recent one is that the main person behind the project has launched it and already moved on to other things. In this post Solazy says he’s working on a cool concept that will be live soon, and it’s unrelated to Prism, which is now primarily developed by Colby and the community.

2. The Spectrum Indexes are imperfect. A Spectrum basket is immutable once it launches, with no way for the creator to rebalance, so if one of the components dies or gets rugged you are stuck holding it inside the index until you redeem.

3. Revenue is fairly low across the board. All the fees and buybacks are great in theory, but if volume dies right down, it’ll be tougher and tougher to get people excited.

4. Complex onchain ecosystems are ripe for being exploited with the recent advances in AI. I continue to be wary of all of DeFi at this particularly vulnerable time, and am mindful of how much of my money I put into any one project.

5. The v4 hook design is still relatively new, so data tools struggle with it, meaning you will see explorers and charting sites report broken or misleading liquidity for these pools. DexScreener for example will see liquidity = 0 and mark the indexes as suspected scams; not great.

So, plenty to be wary of.

Where I land

This is still early and experimental. What pulls me in is not just the PRISM token on its own, it is the shape of the whole system. The idea of turning a narrative into one basket you buy in a single click, with the people who curate it getting paid and value flowing back to one asset underneath, is a fresh answer to a problem we’ve had for years.

I’m not overly concerned about Solazy moving on to other projects. It’s not great, but it’s bolstered a lot by the fact that Colby has effectively taken on a leadership role. For most projects like these: the techy, cool, interesting ones, you usually need other builders and devs to get excited and get involved for the project to succeed. And while that would obviously help a lot here, I do think the Spectrum product alone is interesting enough to support the ecosystem.

I want to see what the V2 launch is like, and I especially want to see some sort of solution for rebalancing indexes, and solving the problem of creating an index that ends up having a token that rugs in it.

Over my many years in crypto I have seen sooooooo many projects come and go where the tech was genuinely innovative and interesting and the price just never went anywhere. It’s one of the reasons so many people eventually get jaded in this space; the things that go up in price aren’t always the things that “deserve” to go up in price.

I think there’s a very high chance this project, unfortunately, fizzles out. Like, 90-95% chance. If it doesn’t gain momentum, doesn’t generate revenue, and if new builders don’t join and/or Colby decides to move on, then it’s going to be extremely hard to see the protocol thrive.

In this post he mentions multiple teams being added to the ecosystem, and many core products being in flight at the same time:

Which is cool and all, if it’s true (I have no reason to believe it’s not true, but I also don’t have any additional proof that it is).

Anyway, as I said, I consider this a long shot. IF it works out though and if there is renewed, sustained, and consistent attention / building / revenue / etc etc etc, then I think this is the type of token that can do many, many multiples from there. It’s another asymmetrical bet, but one I am pretty comfy and happy making personally.

I don't recommend anyone else make it, but do your own research and thinking, look at your own portfolio and allocations, and if it makes sense for you, dip your toes in.

The Prism ecosystem is one of the more original things I have seen ship this year, heck, probably in multiple years. The design is genuinely clever, and the narrative-as-basket idea is one I keep coming back to.

I find it genuinely cool, exciting, refreshing, and awesome to see out in the wild.

For that, if nothing else, I am happy.

Resources

PRISM contract address: 0xbd3AB5859f244CC9F51Ee0Ca755c5cf663D80040

Prism Website

Prism Twitter

PrismBeat

Spectrum

dStable

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

Why is the market crashing? Well aside from it being a general bear market for crypto, the pending implosion of Saylor’s MicroStrategy and the fears around a future pending implosion from quantum computing have got the crypto markets well and spooked.

It’s been a while since I’ve done a portfolio review post, and while I don’t need a market crash as a reason to do one, it’s generally a good time to take stock of your portfolio situation and re-assess whether or not you’re still happy with how you’re allocated. So, that’s what we’re doing today.

Last portfolio update on February 26th, Bitcoin was sitting at 66k. We’ve had a bit of a rollercoaster ride since then, getting back up over 80k, then crashing below 60k, and now sitting at around 62.6k. In other words, BTC is down 5.2% since last update.

My portfolio, on the other hand, is up about 25%. Great success! While not everything has done well for me (REKT and some of the BANKR ecosystem tokens are down pretty bad since then), I significantly outperformed by being heavy in AI/Privacy coins. ZEC, TAO, HYPE, and especially VVV/DIEM have carried my portfolio.

Those tokens unsurprisingly now make up the bulk of my allocations, but I have been slowly allocating some more of the speculative portion of my portfolio over to TCG tokens.

Here’s the full breakdown of every token I hold:

Read more

I’ve spent the last month watching my life slowly descend into TCG madness. My wife’s brother is staying with us at the moment and he has binder after binder of Pokemon cards and has been collecting them for years. My wife got intrigued, bought a couple of packs when we were at the shops one day, opened them, and was immediately hit with a wave of nostalgia and affection for “the good ole days”.

I felt something similar to that but my vice was always Magic the Gathering, and I quickly found myself pulling out my stash of cards from 20-30 years ago and sorting through them.

Almost every day at home we are now ripping open packs of physical cards and having an absolute blast doing it.

It seems we’re not alone as well. The entire world has seemingly been going crazy for Pokemon, and we find ourselves at the top of an insane bull run in the world of cardboard cards.

I have so many thoughts and things to share on this topic. It reminds me a lot of NFTs. Surely it’s a bubble, surely it will burst spectacularly at some point. Surely it’s entirely unsustainable. But, also, maybe there’s still time to capitalize on the upside?

It’s a risky game to be playing, but my thesis is that there just might be a few more months of mania left — and we all know that the last few months of a speculative bull run are always the craziest.

There are a lot of ways to get exposure to this sector but since I am a crypto person, I will be specifically talking about the onchain avenues for exposure. What platforms you can be using, how you can use them, what airdrops you may or may not be farming, what tokens already exist out there from the higher marketcap to the truly speculative bottom-of-the-barrel meme-stuff, to everything in between.

Settle in cause this is gonna be a doozy, but by the end you should hopefully have a solid understanding of what a potential onchain TCG meta might look like, and what tokens I am currently allocated in.

Remember that the mania is physical first

Before we touch a single token, you have to understand the physical side of things, because none of the onchain stuff works without it.

Trading cards stopped being “just a kid’s hobby” a while ago. The global TCG market is projected to hit roughly $15.1 billion in 2026, with Pokemon making up the bulk of that. 2026 is also the year of Pokemon’s 30th anniversary, and The Pokemon Company (TPC) and community are treating the release of the 30th anniversary set as a massive occasion.

TPC says it has produced 85 billion trading cards since inception, and roughly 10 billion of those (about 12% of every Pokemon card ever made) were printed in the last financial year alone. Demand is (currently) so far ahead of supply that they’re printing at a pace that would have been unthinkable a few years ago, and it’s virtually impossible for people to get their hands on the things they want. Stores sell out to preorders, and there are lines dozens or hundreds of people long waiting to get their hands on the latest drops whenever supply does hit the shelves.

This, to me, once again screams “top” and “bubble”, and I am sure that we will eventually hit an inflection point where the music stops. But for now, let’s allow ourselves to succumb to the delusion for a while longer. I personally think the market won’t cool off until at least the 30th anniversary drop (in mid September), but I am also treating that as a date when risk is at its maximum level (rather than at like an 8/10 level it might be at now).

What “onchain TCG” actually means

TCG means Trading Card Game. Pokemon is a trading card game. MtG (Magic the Gathering) is a Trading Card Game. One Piece is a trading card game. There are many more.

In the context of onchain TCGs, and of collectibles, you can basically eliminate the G part of it all. These are trading cards, treated as collectibles and financial assets — not many people are actually using them to play the game (MtG is the big exception, and it’s probably why not many people invest in MtG cards the same way they do other cards).

When it comes to trading these physical cards, there are real limitations. There are liquidity issues, shipping costs, high marketplace fees, fears of frauds or fake cards, and so on. Being in Australia, it’s not that easy or simple for me to buy the cards I want without having to deal with expensive shipping costs (and customs fees).

Bringing the cards onchain solves most if not all of these issues.

The way it works is a platform buys real, graded cards (PSA, CGC, BGS) and stores them in an insured vault. Each physical card gets a matching NFT. You can hold the NFT, trade it instantly 24/7, or redeem it to have the real card shipped to your door. The card never has to move for ownership to change hands*, which is the entire point.

*the big asterisk obviously being that these are not fully self-custodied NFTs in the sense that there’s still trust involved — you trust in the company (or companies) that are storing and securing these assets and that everything will work as-intended when it comes to transferring ownership via the NFT and when it comes to redeeming the assets.

Nonetheless, these online platforms are doing a spectacularly good job of driving liquidity onchain. They are buying massive amounts of cards to store in their vaults in order to improve the liquidity of their marketplaces, and they have also introduced gacha (gambling) mechanics into the whole system too.

The engine that drives almost all the volume are these gacha machines. You pay a fixed price (anywhere from $1 to $1,000 depending on the machine) and pull a random card from a pool. It's a digital pack rip. All top platforms publish the odds and the expected value, and most offer an instant buyback, usually around 85% to 90% of the card's market value, so you can cash out the second you pull if you want.

A word of warning: most sites will advertise that the expected value of a gacha opening is higher than the cost of the pack itself, ie, pay $50 and get $53 in value back. While that might be strictly true if you’re considering the fair market value of what you get, most of the time what you get is highly illiquid, and you’re better off selling it back to the platform at the 85-90% discount, which, after their own marketplace fees, means you’re definitely losing money, on average, every time you play the gacha game. *exception again is that you might make up the value with a future airdrop, or if you hold the cards and they go up, but both are just more levels of speculation baked on top.

A look at the numbers

Volume is really starting to take off on these platforms:

From about $10.4m in January 2025 to $230.1m in May 2026. That's a 22x run in seventeen months. It hasn't gone straight up (notice the cool off late last year), but the trend since March has been relentless, with each of the last three months setting a new record.

We’re also seeing more and more new platforms pop up and it’s worth keeping an eye on them and their volume, especially for potential airdrops if you can get in early.

As you can see, most of the platforms have yet to drop a token. I don’t know whether they will all be dropping one, but some of them surely will, and if this meta continues and the volume continues to go up up and away, there are surely some excellent airdrop farming opportunities to be found in this sector.

The Larger and Established Platforms

There are 4 platforms that stand head and shoulders above the rest in terms of volume:

1. Collector Crypt

2. Courtyard

3. Phygitals

4. Beezie

They all operate in largely the same way. Deep marketplace liquidity that is growing at a rapid pace, gacha machine style pack openings that you can do from $25 to $1000, automatic buy-back offers of 85-90% of the value of your pulls.

Beezie is the smallest of the set but they have been around for years and have a strong reputation within crypto, so it very much deserves a spot amongst the top dogs.

Collector Crypt has their $CARDS token that has been on a tear lately, and we’ll dive into the token a bit more in the next section. Courtyard, Phygitals and Beezie are all running points/XP programs which should, hopefully, lead to a nice airdrop in the future, but we have all been burned by protocols farming users for points ad naseum so I always take these things with a giant grain of salt now.

Still, probably worth chunking through a bit of volume, and great platforms to explore if you’re genuinely looking to collect some cards.

The Up-and-Comers

These are the newer platforms that are growing fast but are still only churning through a fraction of the volume of the giants. I think these represent some of the best potential airdrop opportunities since you’re getting in at a much earlier stage and competing with far fewer people at this point. The risk with these though is that they never reach escape velocity and either don’t launch a token, or they launch one but it doesn’t get valued particularly highly.

1. Renaiss

2. Mnstr

3. Gacha Game

4. Pull.fun

5. Dyli

6. Slabz

These all operate in basically the same way and take the same approach as the bigger platforms, although some also have their own spins on things. Also, as I was researching this post, I found more and more and more of these platforms. I had to stop adding them at a certain point. I think we’re going to see many more pop up and try and ride the hype train, and I think a good general strategy is to try and use the ones that have been around a bit longer.

I’m not gonna do a deep deep dive on every platform individually (though if there’s sustained interests and people want that for next week, maybe I can do a follow up).

One thing to be aware of with ALL of these platforms is that.. the odds are stacked against you. Take a look at this chart of the top gacha machines, ranked first by EV and then by Gross EV:

Generally speaking, if you’re playing the gacha games, you’ll want to pick the ones with the best EV (overall if you’re planning to hold the cards, gross if you’re planning to sell most of them back).

EV is what they will advertise on their websites. “Oh look, our $50 gacha machine has an expected return of $53.06!!!” — sounds great, right?!

There’s some truth to that, in the sense that the actual fair market value of the items you win might be worth slightly more than the cost to play the game. But the vast, vast majority of things you’ll win are non-grails, non-highly-sought-after items, and if you actually tried to sell them on your own then you’d most likely lose money.

The Gross EV looks at the 85-90% buyback offer the platforms offer to instantly buy the cards back off you for, and is mostly what you should be looking at.

A gross EV of -2% or -3% means that for every $100 worth of spins you’re doing, you’re losing $2-3 on average.

On the plus side, having real odds known up front is a bit of a welcome change. It’s not such a bad price to pay for some fun, for the chance at winning something big, and especially for the chance of winning an asset that you think might appreciate in price and that you can now hold a secure digital version of.

On the down side, it’s still gambling and “the house” is still gonna win. So, gamble responsibly. I think everyone can win together for some amount of time — in a bull market (even a micro bull / isolated bull, within a macro bear) then things can get silly for longer than anyone might expect.

But eventually the music has to stop and when that happens, you don’t want to be left holding the bag.

Tokens that give you exposure to this sector

Everything up til now has been talking about things at the platform and airdrop farming level. Truth be told, there aren’t too many fantastic ways to get exposure to this movement aside from buying individual cards yourself as investments. That’s probably a decent strategy if you know what you’re doing / can take the time to research, but it’s not the approach I am really taking myself.

For starters, I think that outside of a very very very small percentage of cards, almost everything is WAY overpriced right now. Buying in at these prices feels like a recipe for disaster, especially given the general illiquidity of cards.

So, instead, I am choosing to find second order plays to get my exposure. It’s still risky, just risky in a different way, and in a way I personally understand a bit better and have more experience with. Namely: onchain tokens.

The conversation has to start with $CARDS here, the token behind Collector Crypt.

It sits at a $55m marketcap with an FDV of a whopping $430m. That initially scared me off and I had little interest in buying into a token that had such a massive supply overhang, especially since there was no current mechanic for using revenue to buy back the token or support the token price.

This looks kinda bad, but as I looked into it more however, my fears started to subside.

The Foundation tokens are earmarked for longer term holding and growth and the team has indicated no intention to sell those any time soon. The pre-seed unlocks have already started, vesting 14m tokens per month, and will run until August this year. This is the most significant sell pressure, but the market has proven an ability to shrug them off and continue to go up in price as the unlocks occur.

I believe the team unlocks only begin next year, and that’s a bridge to cross as we approach it. I am really looking at this TCG meta as a trade between now and September or so, and I don’t think the tokens entering the market between now and then are significant enough relative to the revenue the protocol is making + the marketcap to deter it from being a reasonable investment.

They are also positioning themselves to be the “Hyperliquid” of TCGs, wanting to be more protocol than platform. It’s some good marketing, but how much truth is there to it? Hard to say at these early days but there’s some signs of life with other projects building on top of them and tapping into their liquidity to build out their own projects.

Thus, $CARDS is the best way to get exposure in size to the sector. I own a healthy position now and am comfortable holding through the next few months. It’s worth noting that I am not “early” to this trade. It has already done a 5x from the lows, and some people are choosing this as their time to exit the trade.

My logic is that this becomes a strong trade IF the overall thesis of the whole onchain TCG meta continues to evolve and play out as I am thinking/hoping it will: with continued significant growth and mania, going hand in hand with the IRL mania in the trading card world.

Alright with CARDS being the main player in this space, let’s now take a quick look at some of the smaller cap plays — even higher risk, but potentially higher reward.

Read more

Lots of tokens have pumped a lot: VVV, DIEM, ZEC, NEAR, SERV, the list goes on. I caught some of them and missed others. The question at times like these is “what’s next?”

There are a couple of ways to think about answering that:

1. Looking for more beta plays within the existing narratives

2. Looking for the next narrative

While I think looking for the next narrative is overall the more fruitful endeavor and where the largest gains will be made, today we’re going to take a look at one token that I have my eye on that sits nicely in the AI x Privacy nexus and hasn’t run much (…yet?)

Perhaps next week we’ll look at what the next narrative(s) could be.

Today, however, we take a deep dive into a token that is down 97% from its peak, has been around for years, is building a real product that sits at the heart of the AI x Privacy sector, and recently partnered with Venice AI (yes, that Venice) to help with their privacy.

So without further ado…

Read more

This post is sponsored by Euphoria

Euphoria has fully gamified the trading experience, abstracting away any reliance on order books and making it a "click the box" style game.

It’s honestly an incredibly fun way to trade, can see this going viral or becoming a mainstream app what with the gamblification-of-everything we're seeing in society lately.

Is it going to replace orderbook trading for sophisticated traders? No. Is it a fun way to spend a few spare minutes you might have here and there? Hell yes.

Sign up and try it out here.

zkTLS

One thing I am always interested in is when and how crypto connects to “the real world”. Stablecoins are an obvious and great example, but there’s something a bit more technical that works behind the scenes / under the hood that is pretty damn cool and worth knowing about: zkTLS.

The name stands for zero knowledge transport layer security. TLS is the protocol behind that little padlock symbol you might sometimes see in your browser, essentially helping secure some websits/data. zkTLS wraps that connection in a zero knowledge proof, so you can prove things about what the server told you without showing anyone the actual response.

As a quick example, let’s say you have a bank acocunt with a balance of $50,000. A DeFi lender wants proof you have at least $10,000 to your name before giving you an undercollateralised loan. With zkTLS, you can generate a proof that says yes, my bank balance is above $10,000, and provide it to the lender. The lender sees the proof but they don't see your real balance, your name, or anything other identifying information. Privacy secured.

This all works using the underlying ZK-Proof technology, which I wrote about a couple of years back.

zkTLS gets even more interesting when you consider AI agents and getting them to do things for you in a secure manner. When websites start catching up and implementing the ability to check for proofs, you can provide proof to your agent that you have X, Y or Z and they can then go out into the real world and do things on your behalf and be able to prove that you do have X, Y and Z, without actually being a security risk of leaking any of the underlying information.

I think in a number of years, this tech is going to be uniquitous and powering a lot of things in our digital world, and I think it’s cool and interesting to know about this kinda stuff and how things work, at least at a very high level. Hopefully you found it cool and interesting too!

Thanks as always for reading, and see you next week with another little learning!

Subscribe now

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

You see the numbers quoted everywhere. 100k token context window, 1 million tokens, 10 million tokens. The numbers keep going up, the marketing keeps getting louder, and most people have no clear sense of what any of it actually means.

So today, we fix that.

By the end of this Letter you’ll know what a context window is, why a bigger number doesn’t always mean a better model, and how to think about all of it next time you’re choosing which AI to use. I’ll also share some tips and tricks for optimizing your context window regardless of what model you’re using.

The simple version

A context window is the model’s working memory.

Everything the model needs to do its job in a single request has to fit inside it. That includes your prompt, the system instructions, any documents or images you’ve uploaded, the conversation history from earlier in the chat, the tools the model has access to, and the response it’s about to generate.

All of it shares one budget, and the budget is measured in tokens. I covered tokens in Letter 108, but as a refresher, a token is roughly 3/4 of a word. So a 200,000 token context window holds around 150,000 words of stuff at once. A million tokens is around 750,000 words, which, for reference, is the length of a fairly long novel.

How we got from 4,000 tokens to 10 million

Let’s appreciate just how fast things have accelerated over the last few years.

When ChatGPT launched in late 2022 it was with the GPT-3.5 model and that had a 4,000 token window. You couldn’t paste a long article in without hitting the wall.

By early 2023 we had 8k and 16k models and by late 2023, OpenAI shipped GPT-4 Turbo with 128k. Soon after, Anthropic pushed Claude to 200k and then Google blew the doors off with Gemini 1.5 Pro and 1 million tokens in early 2024.

In March this year, Anthropic made Claude Opus 4.6 generally available at 1 million tokens with no pricing surcharge and one month later they shipped Opus 4.7, which kept the 1M context, and that’s basically become the standard across the board.

Meta’s Llama 4 Scout claims 10 million which seems a bit excessive to me since the quality of output as the context window expands degrades significantly (more on this soon).

Overall though coming from when ChatGPT launched to today, it’s a 250x to 2500x increase.

Larger context windows are more expensive to build. Transformer models (the architecture LLMs are built on) use something called self-attention, where every token pays attention to every other token in the sequence. Doubling the context window roughly quadruples the computation, so a 1 million token model isn’t five times more computation than a 200k model, it’s actually 25 times more.

Most of the recent gains have come from clever engineering around how attention works, prompt caching, and efficient memory management.

Is bigger always better?

I’m no expert on this topic as a general rule (though word on the street says no), but when it comes to AI context windows, bigger is not always better.

The context window size is the maximum the model can technically accept, but it’s not necessarily the amount it can use well / optimally.

Some researchers at Chroma published a study last year called Context Rot. They tested 18 different state of the art models, including Claude, GPT and Gemini, on simple retrieval tasks at different input lengths. The task itself stayed the same, only the length of the input changed.

Every single model got worse as the input got longer. Some models that were scoring near 100% on short inputs were under 50% on the same task with a longer input.

The community adopted the name of the study and the phrase “context rot” quickly became part of the AI vernacular. The gist is simple: as you stuff more into a prompt, the model becomes less reliable (even when none of the extra content should matter and genuinely add helpful context).

A few things are going on under the hood to cause this context rot:

• Models tend to pay more attention to the start and end of a prompt than the middle. Important information buried in the middle of a long context can get ignored. The researchers call this lost in the middle (scientists aren’t known for their originality in naming conventions, see: the Very Large Telescope in Chile).
  

• When there’s content in the prompt that looks similar to what you’re asking about but isn’t quite right, the model gets confused. Even one of these distractors can throw it off.
  

• Where information sits in the prompt also affects how well the model uses it. The same fact at position 1,000 performs differently to the same fact at position 50,000.
  

• When the question and the answer use different words for the same idea, the model struggles more as the context grows.

The takeaway is that the advertised context window figure is a ceiling, not a target, and you should generally be clearing your context window and starting fresh well before you get close to the ceiling for practically every task you use AI for.

Another reason bigger isn’t always better is the cost / usage. The larger a context window you’re utilizing, the more tokens you’re sending back and forth and you’re either paying for that via the API or you’re using your monthly subscription usage limits up. Keeping your context windows small and manageable is not only better from the perspective of ensuring better quality outputs, but it’s fiscally prudent too.

What this means for what you’re actually doing

If you’re using AI for work that involves a lot of information at once, the context window matters; most people are not actually doing this though.

Most documents are smaller than people assume. For example, a 50 page PDF is around 25,000 tokens. You can paste a document like that in and have a long back and forth conversation with the AI and be well under 100k tokens.

If you’re building and using your own agents however, this becomes really important. Agentic workflows burn through tokens fast. Like, really, really fast. The agent reads a page, decides what to do, calls a tool, reads the result, repeats. Every step adds to the context. When you chat with your agent, it’s not only reading your entire chat history every time for context, it’s also loading other skills and tools and bits of memory in order to work as well as it does.

I can hit a 200k context window within an hour or two of chatting back and forth.

Coding just adds to this. If you’re coding you can burn through tokens insanely fast once subagents and parallel work gets involved.

This is one of the reasons local models can be a very appealing option for certain work. There are no per-token charges or usage limits when you have a local model. I covered local models in Letter 107 if you want to learn more there.

How to actually manage your context window

This is the part of the Letter I want you to take away most. Below are the practical tactics I use, alongside guidance from some of the experts who actually build these models.

The bigger lens here is something Anthropic’s engineering team calls context engineering. They published a deep dive on it in September 2025 and it’s worth a read if you want to go further than this Letter takes you. Their summary line is the best I’ve seen on the topic:

That framing has actually changed how I use AI day to day. Instead of asking what should I put in, I now ask what can I leave out (once again, “less is more” wins the day).

Here are the tactics that flow from that.

1. One task per chat, and start fresh often

This is the single most useful habit I’ve picked up. If you’re researching one topic and then pivot to something else, start a new chat. If you finish drafting something and want to refine it, start a new chat. If you spend an hour debugging one issue and the next issue is unrelated, start a new chat.

Teresa Torres, who writes about AI product work at Product Talk, put it really well: every long chat is accumulating noise, and that noise is actively making the next response worse.

2. Use the handoff technique

When you do need to carry context forward, get the model to write a dense summary of what matters before you start over. Copy it and paste it into the new chat as the opening message.

This is essentially what Claude Code does automatically when it hits the context limit. Anthropic calls it compaction, and they describe it in their context engineering post as the first lever you should pull. They preserve architectural decisions, unresolved bugs and key files while throwing away redundant tool outputs and old messages. You can do the same thing manually with any AI tool.

3. Watch the 50% line

Arthur Clune, who works on AI products, made a really sharp observation in a recent post. LLM accuracy starts declining noticeably once the context window passes about 50% full. Anthropic’s own compression triggers automatically at 95%, which means by the time you see the message, you’ve been operating in degraded mode for a while.

That’s why you should not just wait for Claude Code to automatically compact, since that only happens when you get close to the limit. You can manually run it yourself by typing /compact whenever you want.

4. Curate before you work with AI

Pulling in 5 relevant documents is going to get you a better result than stuffing the context with 50 documents and hoping the AI can “just figure it out”. The model doesn’t always know what’s important (and it certainly doesn’t know what’s important to you specifically). You know, so use whatever edge you have as a human while you can.

5. Put the important stuff at the start or end

The lost-in-the-middle research has held up across model generations. Information at the start and end of a prompt gets more attention than information buried in the middle. If you have a long document and one paragraph really matters, restate that paragraph at the end of your prompt, even if it’s already in the document a couple of times.

6. Use Projects or Custom GPTs for persistent context

Claude has Projects. ChatGPT has Custom GPTs. Both let you set up a persistent system prompt and upload reference files that don’t count against each individual chat’s context. If you have a recurring use case, put the instructions and reference material in a Project or Custom GPT instead of repeating them in every chat.

7. Use RAG for big knowledge bases

RAG stands for Retrieval Augmented Generation. Instead of dumping a thousand documents into the prompt, you index them in a vector database, and only pull the most relevant chunks for each question. This is what tools like Perplexity, NotebookLM and ChatGPT’s deep research feature are doing under the hood. I’ll cover RAG properly in a future Letter, it deserves a dedicated piece.

8. Pick the right model for the job

Gemini 3.1 Pro is cheap for huge document processing. Claude Opus 4.7 is the strongest at reasoning over the context once it’s in there, with Sonnet 4.6 a great middle option if cost matters. GPT-5.5 is OpenAI’s current frontier and sits between them on price. If you’re processing a 500 page PDF, Gemini is probably your tool. If you’re working through a complex multi-step problem, Opus or GPT is probably your tool.

9. For agents, separate read tasks from think tasks

If you’re building agents (or using Claude Code, Cowork, Cursor and the like), Anthropic’s research on long-running agents is gold. The key insight is that exploration burns context fast. So they recommend subagents, which are spawned for specific read heavy tasks (search across files, review code for security issues) and return only a summary. The main agent’s context stays clean.

You can apply this manually too. If you want Claude to research something deeply and then act on the findings, do the research in one chat, get a tight summary, and use that summary as the starting point in a new chat for the action work.

10. Don’t paste images you don’t need

Images are expensive in token terms. A single high-res screenshot can use 1,000 to 2,000 tokens. If you've uploaded 20 screenshots into a long chat, that’s 20,000 to 40,000 tokens just on the images alone. I’m a huge fan of screenshotting and adding things to chat, but it can get a bit out of hand at times, and this is a good reminder to not go overboard with them.

Closing thoughts

Hopefully this has helped you understand a bit more about what context windows are and how they work. Most people just see headlines that brag about bigger context windows and think that bigger is always just straight up better, but the reality is that I am rarely going over 200k tokens for any of my tasks these days. Occasionally I’ll get to 400-500k before compacting, but not that often.

Maybe / hopefully one day this problem gets solved and we can reliably use 1 million or even 10 million token context windows without things getting lost in the rot of it all.

Until then though, keep an eye on your context. Regularly start new chats when you have new tasks, handoff from one session to another, and repeat the important things at the start and end of your prompts if you want better outputs.

Thanks as always for reading, and see you next week!

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

This post is sponsored by Moonberg

Moonberg is at the forefront of bringing AI into trading strategies, allowing you to test trading strategies with real data and agents before going live.

Their latest QuantFi product is worth checking out. It has:

• 686 unified datasets (onchain metrics, sentiment, momentum, etc)
• Realtime stream on 5,000 tokens
• A no-code formula builder and backtest engine
• Much more

The whole idea is to allow everyone to access AI trading capabilities regardless of experience level, and to test strategies in a safe and reliable way before deploying any actual funds.

It works by utilizing their own agentic infrastructure on the platform. The agent reasons about strategy in steps. so if your spec is ambiguous, it asks. If your idea has an obvious bug ("you'll get stopped out every time"), it tests before deploying.

It's a full agentic pipeline that owns the strategy from idea to execution, with the data and infra plumbed in.

Discover more here.

What Is DePIN, and Why Should You Care?

A lot of crypto narratives sound good in theory but fall apart when you ask the simple question of who is actually using this.

DePIN is one of the few corners of crypto where the answer is real customers paying real money for real services. Not huge by crypto standards, but the trend line is up and to the right, and the revenue is real.

What it means

DePIN stands for Decentralised Physical Infrastructure Networks. The idea is simple. Instead of one company owning all the GPUs, hard drives, or wireless towers, the network is built by thousands of individual contributors. Anyone with the right hardware can plug in and earn tokens for the work they do.

Traditional model: AWS owns the data centres and rents them out. Profit goes to Amazon.

DePIN model: people run GPUs at home, the protocol pays them in tokens, customers pay the network to use those GPUs. Profit goes to contributors.

The main categories

• GPU compute & AI infrastructure. Networks that aggregate GPUs from independent operators and rent them out for AI workloads and rendering. The hottest category right now because of AI, obviously.
  

• Wireless & telecom. Mobile networks built on community deployed hotspots, often partnered with traditional carriers for fallback coverage.
  

• Storage. Decentralised storage networks that pay people to store data on their hard drives, increasingly used for AI training data and Web3 applications.
  

• Data networks. Networks that crowdsource real world data collection. Things like mapping via dashcams, sensor networks, or monetising idle internet bandwidth for AI training.
  

• AI networks. Decentralised networks for AI model training and inference, where models compete for rewards or where users access compute through token staking.

Why I find it interesting personally

DePIN is one of the rare places in crypto where the technology touches the physical world in a way you can actually see and use.

We have spent decades building infrastructure by companies raising infinite capital to build huge data centres, and DePIN is an alternative option that once again returns power and rewards back into the hands of “the people”. The hardware is already out there in billions of homes and pockets. The missing piece was a way to coordinate contributors and pay them fairly for useful work, which is exactly what blockchains can be good at.

The implications get more interesting the further you zoom out. If this works, infrastructure stops being something owned by a handful of giant companies and starts being something owned by anyone willing to participate. A mobile network without a telco, a cloud without AWS, or a mapping service without Google. None of these have been possible before because coordination at that scale used to require a central authority.

Whether DePIN ends up being the model that wins or just a stepping stone to something else, the experiment is one of the more interesting things happening in tech right now.

One of those things in crypto that I think is worth paying attention to even if you never touch a token.

For now, thanks for reading, see you next week with another little learning!

Subscribe now

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

That was 70 days ago. Since then, the ecosystem has soared with both VVV and DIEM hitting new highs after new highs.

So today I’m doing an update on what’s changed, what’s driven the move, and where I think things go from here.

If you’re interested in leveling up your AI learning journey even more, then check out the new company I have launched alongside a couple of friends: The Stoa of AI.

We create video courses and have weekly live workshops and calls that show you practical ways to implement AI into your daily workflows.

We’re in early access mode with discounted pricing, check us out here: https://www.skool.com/thestoaofai

A quick look at the numbers

Things have boomed across the board, so here’s a quick snapshot of where some of the key numbers are standing today:

What’s happened over the last two months?

There are five key things that happened between Letter 102 and now:

1. Verifiable end-to-end encryption (March 18)

This was a technical milestone that a lot of people had been waiting for. One of the main criticisms of Venice up until this point was that their privacy was on a “trust me bro” basis. On March 18, Venice shipped four privacy modes, the most important being Trusted Execution Environment (TEE) and full End-to-End Encryption (E2EE), through partners at NEAR AI Cloud and the Phala Network.

This made things shift from “trust us, we don’t log your prompts” to “you can cryptographically verify that we cannot see your prompts.” Each response includes attestation evidence you can independently verify. It’s a real, hardware level guarantee.

For regulated industries, this is a big deal. A US federal court recently ruled that using standard AI tools could waive attorney-client privilege precisely because they lack confidentiality guarantees. If you’re a law firm, hospital, or financial institution, you literally cannot use ChatGPT for client work in a defensible way.

While most will probably still choose to host their own local LLMs, Venice at least gives anyone with confidentiality concerns an alternative option and way to use models without leaking private information to Big AI.

VVV jumped about 10% the day of the announcement and that was the start of the real run.

2. Programmatic buy and burn launched (April 15)

Up until April 15, Venice’s buybacks were discretionary. The team would use a portion of monthly revenue to buy and burn VVV on the open market, but the timing was at their discretion.

On April 15, they switched to a programmatic engine: every new Pro subscription triggers an automatic, on-chain $1 market buy of VVV that gets immediately burned. The buybacks became verifiable, predictable, and directly tied to revenue.

3. Burn sizes increased (April 27)

There was some criticism around that first announcement that the buyback amounts “weren’t enough”, and so twelve days later, they cranked the burns up, with a new tiered structure based on the level of subscription someone signs up for:

• Pro subscription: $2 in VVV burned (up from $1)

• Pro+ subscription: $5 in VVV burned

• Max subscription: $10 in VVV burned

The market, unsurprisingly, liked this decision too.

4. Staged emission reduction (effective May 1st)

The Venice team committed to reducing the VVV emissions by 50%. From their X post:

• May 1: 6M → 5M annual emissions

• June 1: 5M → 4M

• July 1: 4M → 3M

For context, emissions started at 10M/year at launch. By July 1 they’ll be at 3M/year. A 70% cut from the original rate in under 18 months.

Combined with the programmatic buy and burn, this is the path to VVV becoming net deflationary. They’re still a long way off this becoming a reality but with emissions reducing and buybacks increasing and overall adoption and usage of the platform skyrocketing, everything is moving in the right direction.

5. StrikeRobot partnership (May 7-11)

This one is really cool. As you know, I am bullish on Robotics too (both in general and onchain robotics), having written about the sector earlier this year.

Venice is now powering humanoid robots with private AI vision and decision making, through a partnership with StrikeRobot because, quote, “Your AI shouldn’t spy on you.”

I bring this up not because robots are necessarily going to move the needle on revenue this year, but because of what it signals. Venice is being chosen for high stakes physical applications where privacy actually matters. A robot in your home, watching you, making decisions about you. Do you really want OpenAI, Google, or Anthropic to have access to that data?

A deeper look at DIEM

I covered the basics of DIEM in Letter 102, but it deserves a much bigger spotlight here because I think it’s still one of the most underappreciated parts of the Venice ecosystem. Most VVV analysis I read fixates on the buy and burn, which is of course important, but DIEM is the part that really excites me (partly because I am actively using it myself on a daily basis).

So let me actually explain what DIEM does and why it matters.

A quick refresher

Each DIEM token equals $1 per day of Venice API credit, forever. If you hold 100 DIEM, you have $100/day of Venice inference, every day, indefinitely.

DIEM can only be minted by locking staked VVV (sVVV). The mint rate is dynamic and rises as DIEM supply approaches Venice’s target. While your VVV is locked, you continue to earn 80% of normal staking yield. To unlock your VVV, you burn the equivalent amount of DIEM.

So the cycle is:

VVV staked → VVV locked → DIEM minted → DIEM burned → VVV unlocked.

Why this is structurally clever

DIEM creates a futures market for AI compute.

If you’re a developer or AI agent using the Venice API daily, it can be hard to budget because regular API pricing fluctuates and costs are unpredictable.

DIEM solves this by letting you buy your compute upfront. 100 DIEM costs you whatever the market price is today. From that day forward, you get $100/day of inference with no future bills or costs. It’s a fixed-cost compute futures contract.

Not to mention the potential benefit of being able to sell your DIEM at any point too, meaning if you sell in 2 years at the price you got it for, your compute was essentially free.

The DIEM market right now

What I’m watching is the ratio of DIEM market price to its underlying fair value. Each DIEM produces $1/day of compute, so over 365 days that’s $365 of value (less if you discount for time, more if you assume AI compute costs rise over time). If DIEM trades at $200, payback is around 200 days. If it trades at $1,500, the payback period is over four years. The market is essentially pricing in expectations about how much Venice API access will be worth in the future, and that’s a real economic signal.

How DIEM creates VVV demand

Every time someone mints DIEM, they have to lock VVV. That VVV is taken out of the circulating, sellable supply. It’s effectively a second supply sink on top of the buy and burn.

Think of it this way. If you’re an enterprise developer who needs predictable AI compute costs, you have two paths to access Venice:

• Path A: Subscribe to Pro/Pro+/Max and pay Venice in fiat. Venice uses that revenue to buy and burn VVV.

• Path B: Buy DIEM on the open market and stake it for daily credits. The DIEM you bought required someone to lock VVV to mint it.

Both paths create VVV demand. Path A through direct burn pressure. Path B through removing tradeable VVV from supply. As Venice’s user base scales, both paths grow simultaneously.

Some valid criticism to the model, and some counterpoints to it

There’s been ongoing pushback on the DIEM model with the main argument from skeptics being that when developers buy DIEM and use it for compute, no new cash hits Venice’s balance sheet (unlike Pro subscriptions which do generate cash revenue). The concern is that DIEM could cannibalize Pro subscriptions and drain future cash flow.

Venice founder Erik Voorhees has pushed back on this directly by saying that DIEM was designed to solve pricing instability and create programmable compute, not to be a primary revenue mechanism. The Pro subscription model funds Venice as a company and DIEM creates a secondary economic layer.

I think one important thing to keep in mind is that DIEM holders aren’t necessarily the same population as Pro subscribers. Pro subscribers are mostly individuals and small teams who want a chat interface and convenience, and often are not crypto native at all.

DIEM holders are developers, AI agents, and enterprises that are crypto native and that want guaranteed, programmable access for production workloads. They’re different markets, and Venice serving both expands the total addressable market rather than cannibalizing one with the other.

Where I think things go from here

Here’s how I’m actually thinking about my holdings and positions in both VVV and DIEM personally.

The fundamental thesis has gotten objectively better since Letter 102 with programmatic burns, faster emission cuts, a real enterprise privacy product, and 4x user growth. Every dial moved in the right direction.

I was bullish then, and I am obviously still bullish now, from a fundamentals and company / ecosystem perspective.

From a price perspective? I am taking some profits off the table. I think we’re at the point where the market is frothy for all things Venice, and while there’s always still room to run, we’re at that point where the Banklesses and Ansems of the world are talking about the fascinating tokenomics of DIEM.

If this was 5 years ago, when Bankless and Ansem started talking about something, it was considered early. Today? It’s more likely to be a local top than to be early.

In other words, I’m not frothing at the mouth to buy VVV at $18 or DIEM at $1500. I’ll gladly take a few chips off the table here, and if we get a significant dip (40-50% or so) then I would be looking to add more to my stack.

I’ve got a bag of both that I am holding longterm — especially my DIEM which I am using daily to power my Hermes agent(s). But in this market and climate, I think it would be silly to not be thinking of taking profits too.

Final thoughts

I think Venice is one of a very small number of AI x crypto projects where the fundamentals genuinely justify a real position over a multi year timeframe.

That doesn’t mean it goes straight up from here, it won’t. It doesn’t even mean it’ll be up from here in a few years. Nobody has any idea what is going to happen on a multi year timeframe in this space, and anyone telling you otherwise is lying.

The best thing to do, in my experience, is find good projects, with good fundamentals, at good prices, and buy sensibly, and take profits sensibly. Much easier said than done, but if you’re thinking about positioning in the AI x crypto space, this remains one of the best projects out there. Just perhaps not right this minute at these all-time-high-ey prices.

I’ll write another update in a few months (either as a standalone or as part of a portfolio review).

In the meantime, I’d love to hear what you think. Reply directly or leave a comment below.

Thanks as always for reading!

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

This post is sponsored by Umia Finance

Umia Finance is doing some cool things, building the onchain venture creation layer for agentic ventures.

I plugged in my Yoshizenco github repo into their model and it gave us a pretty solid valuation!

Maybe it knows we’re cooking up yoshi v2 to come out soon.

Go check them out here, you can plug in your own github repo or even just point it at another repo that you’re interesting in seeing the valuation of.

What Are AI Agents, Actually?

A few weeks ago I wrote a Little Learning on what AI agents are and how they work. The short version: an agent is AI that does things, not just says things. You give it a goal, it figures out the steps.

Up until now, they’ve had a pretty big limitation. They can manage your wallet, trade tokens, swap assets, and move value around onchain. But they couldn’t easily spend at a real merchant. If your agent wanted to pay for an Uber, buy a domain, or grab a SaaS subscription, it had to stop and wait for you to step in.

That changed last week.

MoonPay launched the MoonAgents Card on May 1st. It’s a virtual Mastercard debit card that lets AI agents spend stablecoins directly from a self custodial onchain wallet, at any of the 150 million or so merchants in the world that accepts Mastercard.

It still requires some human intervention to setup, but once that’s done, the agent can have full autonomy in a way that can’t really exist with a regular mastercard connected to a bank account.

Here’s how it works:

1. You connect your onchain wallet to the card.

2. You complete identity verification (once).

3. You set the spending rules and permissions for your agent.

4. From that point on, the agent can pay merchants whenever the conditions you set are met.

5. The smart contract authorizes funds at the moment of purchase, the stablecoin gets converted to fiat, and the merchant receives regular money.

The key part to me is being able to set spending rules and permissions.

Once an agent has real world spending power the possibilities of what it can do expand so much. It can run a small business workflow from start to finish if you let it.

CZ said recently that AI agents will make a million times more payments than humans, and they’ll use crypto to do it. Brian Armstrong made a similar point. Agents can’t open bank accounts, but they can own a crypto wallet. So the path of least resistance for machine-to-machine commerce ends up running through stablecoins. This is exactly why I’ve been so bullish on the x402 payments standard and ERC-8004 as the rails for all of this.

This card is one of the first concrete products that proves the thesis, acting as a real bridge between the agentic world and the real economy.

The card is currently live in the UK and Latin America through MoonPay’s CLI, with the US and EU rollout coming in the next few months.

Exciting times ahead.

Want to dig deeper? A few good reads:

• The Block’s coverage of the launch

• MoonPay Agents documentation if you want to see what the developer side looks like

• Decrypt on Pay.sh, the new Solana x Google Cloud agent payment rail (very related, worth a read)

Subscribe now

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

There was roughly $651 million stolen across roughly 40 separate exploits; more than one incident every day.

If you have funds anywhere onchain, this is an important letter to read.

I’m not doing this to scare you (actually, maybe I am, but with good intentions). I’m doing it primarily because I think the patterns behind these hacks are going to keep getting worse before they get better, and it’s the time to be proactive about protecting your funds.

What happened in April

While there were over 40 attacks in April, two attacks did almost all the damage, and both were linked to North Korea’s Lazarus Group. Let’s take a look at both of them, and then a brief look at the rest.

Drift Protocol ($285 million)

Drift was the largest perpetuals exchange on Solana with around $550 million in TVL. On April 1, attackers drained $285 million in 12 minutes, wiping out more than half the protocol’s TVL.

The setup started in October 2025 at a major crypto conference. The attackers posed as a quantitative trading firm and spent six months building relationships with Drift contributors. They had legit, professional looking backgrounds. They even deposited over $1 million of their own funds to make everything seem normal and above board.

Once they had gained their trust through social engineering, they used a Solana feature called durable nonces to get Drift Security Council members to unknowingly pre-sign transactions. Durable nonces let you sign a transaction now and execute it later, you can think of it kinda like signing a blank check.

On April 1, the attackers executed those pre signed transactions. Two transactions, one second apart, transferred admin control. They whitelisted a fake token, deposited 500 million units of it as collateral, and withdrew $285 million in real assets. The DRIFT token dropped 42% within hours and SOL fell 5.5% on the day.

KelpDAO ($292 million)

KelpDAO is a liquid restaking protocol on Ethereum. They issue rsETH, which represents staked ETH and circulates across more than 20 chains via LayerZero’s bridge.

On April 18, attackers minted 116,500 unbacked rsETH worth around $292 million. Roughly 18% of the entire rsETH supply, created out of thin air.

The attack exploited KelpDAO’s configuration of LayerZero. When you bridge a token across chains using LayerZero, the protocol uses something called a Decentralized Verifier Network, or DVN. The DVN’s job is to watch the source chain, see that you burned tokens on one side, and tell the destination chain to release tokens on the other side. LayerZero’s documentation says you should configure your bridge with at least two independent DVNs so neither one alone can authorize a release. Two sets of eyes, basically.

KelpDAO didn’t do that. They configured their bridge with a 1-of-1 DVN setup. A single verifier with full authority over a $392 million escrow. LayerZero had explicitly recommended multi-DVN in their integration checklist. KelpDAO went with the default and never changed it for some reason (hubris, probably).

The attackers were able to manipulate this DVN by attacking the RPC nodes, essentially allowing them to generate 116,500 rsETH out of thin air. They then did what every modern hack playbook describes and they deposited it as collateral on Aave and borrowed real ETH against it.

The fallout was arguably worse than the loss itself. Aave saw $8.4 billion in deposit outflows in 48 hours. Total DeFi TVL dropped by more than $13 billion across the board. Lending platforms like Morpho, Spark, Lido and Beefy froze certain operations. The AAVE token fell 17% and ZRO fell 12%.

Trust in DeFi was and is at an all time low.

The full April hack list

Drift and KelpDAO got the major headlines but they weren’t alone. Around 40 (yes FORTY, wtf) separate incidents hit the industry across the month:

What’s causing all these hacks?

If we rewind a few months, all the hack postmortems on X had people saying the same thing: that the main cause of hacks wasn’t bug exploits but rather social engineering. People problems. And that’s certainly still an enormous part of things, especially for the larger hacks, and even the majority of hacks so far this year.

Most protocols have gotten pretty good at auditing their contracts, of double and triple and quadruple checking that they have no exploits, and in feeling relatively safe in the comfort of their code.

But I think the narrative is starting to shift once again back to one that smart contracts are absolutely still a problem. And they’re about to be a much, much bigger problem.

Enter Mythos

On April 7, Anthropic announced a new model called Claude Mythos Preview. They didn’t release it to the public, and didn’t even discuss releasing it to the public anytime soon.

They didn’t release it because of how powerful it is.

To quote Anthropic directly, Mythos can find and exploit software vulnerabilities at a level that “can surpass all but the most skilled humans.” Over a few weeks of testing, they used it to find thousands of zero-day vulnerabilities in every major operating system and every major web browser. Some of the bugs it found were 27 years old. They had been sitting there since the late 1990s, missed by every human security researcher who ever looked at the code, and Mythos found them in days.

It also did things that supposedly freaked out the Anthropic team. In one test, it chained four separate vulnerabilities together and broke out of its own secure sandbox, gained internet access, and emailed the researcher running the experiment (who, incidentally, was sitting on a park bench eating a sandwich at the time lol).

So instead of releasing it publicly, Anthropic launched something called Project Glasswing. They gave Mythos to a small group of partners, basically the biggest names in tech: AWS, Apple, Microsoft, Google, Nvidia, JPMorgan, Cisco, Palo Alto Networks, CrowdStrike, Broadcom, the Linux Foundation, and more. About 50 organizations total.

The goal is to use Mythos to find and patch vulnerabilities in critical software before equivalent capabilities show up in the hands of attackers, with Anthropic committing $100 million in usage credits to make this happen.

Why this matters for crypto

Mythos is the canary in the coal mine.

Anthropic has been pretty open about why they’re doing Project Glasswing instead of releasing the model. Their thesis is that this kind of capability is going to exist in the wild within six to eighteen months whether they release Mythos or not. OpenAI is reportedly working on something similar, and the UK AI Safety Institute already evaluated GPT-5.5 and concluded it has reached similar offensive cyber capabilities on their narrow tasks. Open weight / local models are getting better fast too.

Defenders need a head start. That’s the whole point of Glasswing.

AWS and Microsoft and Apple are in Project Glasswing. The Linux Foundation is in it. You know who’s not in Project Glasswing?

Every DeFi protocol you use.

So when Mythos-class capabilities leak out into the wild (and they will, whether through open source models catching up, model weight theft, or jailbroken closed source models), the first wave of targets is going to be exactly the systems with the most value sitting on them and the least defensive resourcing. Aka, all of crypto.

Smart contracts are open source by design, meaning anyone can read them, anyone can fork them, and anyone can run an AI model over them looking for bugs. The same model that found a 27 year old vulnerability is going to find plenty of five year old DeFi contracts just waiting to be exploited.

What happens when finding a critical vulnerability in a DeFi protocol drops from a six month operation to a six hour operation? It’s not rocket science: you get a lot more hacks.

The subtle silver lining is that the same AI tools that find bugs can fix them. Auditing firms are already using current frontier models to assist their work, and if you’re a protocol team, you can (and should) run these tools over your own code right now too.

But the transition is going to take time and be bumpy. The next 12 to 18 months are going to be the most dangerous window for crypto security we’ve ever seen.

What this means for you

Every protocol you use has an attack surface that’s about to get scanned more aggressively than ever before. Which means your job as a user is to be more deliberate about which protocols you trust with your money, and how much.

Honestly, I am taking the nuclear option. Taking everything off every DeFi protocol for the time being, until I feel safe. As much as I hate to say it, cash in a savings account in a bank is seriously feeling like the safest option right now.

Crypto is still early, and I very strongly still believe in the future of DeFi. But this current period is overly risky in my view, and the rewards don’t outweigh the risks.

One hack for any sort of meaningful amount of funds can wipe out a decade+ worth of yield gains (if not much more).

The protocols that survive this period will be all the stronger for it in a couple of years, and will have built up a lot more trust.

The users who survive this period will be the ones that take their own security seriously.

Final thoughts

I started writing this letter because I wanted to talk about the April hacks and warn people about the dangers of AI. But the more I dug into it, the more I realized the April numbers are almost the lagging indicator. They’re still dominated by the old playbook, social engineering hacks dominating the bulk of the exploited money, with a growing number of smaller smart contract hacks acting as the canary in the coalmine for what’s to come.

The next wave is going to be different. AI driven bug discovery at a scale we’ve never seen, finding code level vulnerabilities faster than protocols can patch them. Mythos showed us what’s coming, and Anthropic trying to do the responsible thing by gating it isn’t necessarily doing us any favours. It might be good for the biggest tech companies in the world, but it’s not doing much for most of DeFi.

My strong recommendation is that you take an hour this week (ideally TODAY) and audit all your onchain holdings and whatever funds you have in any protocols, with a strong view to withdrawing the funds and parking them in cold storage.

I wish I had something more uplifting to end on, but the honest take is that crypto security is about to get harder before it gets easier.

Goodluck, godspeed, and as always, thank you for reading.

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

I might have missed the mark on some exact tokens, and not every call was correct, but I firmly believe that everything has been at least directionally accurate. And I keep coming back to this stuff because I genuinely believe it’s where the puck is going. But until you actually look at the numbers side by side, it’s hard to appreciate just how dramatic the rotation has been over the past year and a bit.

So today I want to zoom out and walk through the data. The collapse of the memecoin sector, the quiet but steady growth of AI tokens, and what I think it means for how you should be thinking about positioning in 2026.

Let’s jump in.

The memecoin collapse

It’ll be no surprise to anyone that the memecoin sector has collapsed, but let’s take a look at the numbers to see just how much it has.

The memecoin sector peaked at $150.6 billion in December 2024, fueled by Trump’s re-election in November and the broader political frenzy. The TRUMP token launch in January 2025 felt like it marked the pico-top, but actually the sector was well into decline already by then. By January 2026 the sector had fallen to $36.5bn, roughly where it still sits today.

That’s a peak-to-trough drop of over 75% in 15 months. Bitcoin held above $93k for most of that period and even hit new ATHs, so memecoins didn’t collapse because crypto collapsed, they collapsed because the engine that powered them ran out of fuel.

The volume side is even uglier. Memecoin trading volume peaked near $20B per day in early to mid 2025 and by the end of the year it had dropped to under $3B.

Pump.fun, the launchpad that powered most of the cycle, tells the same story from a different angle. Platform revenue collapsed from peaks of $7M+ per day in early 2025 to $1-1.5M per day by early 2026 (a number which still blows my mind tbh, it’s pretty insane that they’re still making that sort of money). But, that’s a 75-80% drop in fees, which aligns with the drops in overall marketcap and volume.

The other thing worth knowing about Pump.fun activity is that a lot of what was reported as “user activity” was never really human users to begin with. Dune Analytics dashboards from on-chain analyst @adam_tehc found that 93 out of the top 100 wallets by volume on Pump.fun are tagged as automated bots. The top wallets show 18+ hours of daily activity (which obviously isn’t human). Pump.fun’s founder Alon admitted that around 30% of all wallets on the platform have only ever made a single transaction, which he conceded are likely bots and AI agents.

So when you hear about the platform having “millions of users”, take it with a heavy grain of salt (a very good rule to follow for pretty much anything in crypto). A meaningful chunk of that activity is and was bots gaming the system to farm an airdrop that ended up never really materializing for them anyway (RIP).

The retail story is just as bad. According to a Dune dashboard tracking 1.4 million wallets that traded Pump.fun tokens, 96% either lost money or made less than $500 in profit. Out of 13.55 million wallet addresses across the platform’s lifetime, only 0.412% ever realized profits above $10,000.

Most people are probably better off buying a lottery ticket.

The token graveyard

I covered a lot of this in Letter 106 a few weeks ago, but it’s worth repeating because it’s genuinely staggering and it backs up the broader point. Over 13 million memecoins were launched in 2025. According to federal lawsuits filed against Pump.fun, 98% of those tokens collapsed within 24 hours of launch.

Less than 0.63% of Pump.fun tokens ever graduate to a real DEX. The other 99.37% die before they reach a $90k market cap.

Another thing to note is just how concentrated the sector is. DOGE and SHIB now make up 84% of the entire memecoin market cap, and honestly, when was the last time anyone got excited by either of those coins? Strip those two out and the rest of the sector — the long tail of every TRUMP, FARTCOIN, PIPPIN, and 13 million others — accounts for the remaining 16%.

So when people talk about the “memecoin sector,” what they’re really talking about is two coins from 2013 and 2020 plus a graveyard filled with Murad’s hopes and dreams.

A look at the AI side

While the memecoin sector was bleeding out, AI tokens did something different. They didn’t exactly skyrocket, and the story isn’t that AI tokens went to the moon while memes died. The story is that AI tokens grew steadily while memes died, and the gap between them has compressed dramatically.

In early 2025, the AI crypto sector was around $15B in market cap. Memecoins were peaking near $150B. The ratio was roughly 10:1 in favor of memes.

As of April 2026, the AI crypto sector sits at $22.6B, while memecoins sit at $36-38B. The ratio is now closer to 1.6:1.

That’s a 6x change in the relative sizing of two of crypto’s biggest narratives.

And it tracks with what I’ve been seeing on the ground. When I wrote Letter 99 (AI Season, Take Two?) back in February, the thesis was that the first AI wave was full of glorified chatbots and the second wave would have actual revenue and utility behind it. The data backs this up, not with artificial pumps of AI Agents masquerading as memecoins, but with real products and companies slowly building up over time.

Why the institutions care about one and not the other

Institutional money has been telling us this story for months, with a huge amount of attention being paid to the Bittensor network and TAO token.

Some data points from Q1 2026:

• Nvidia put $420M into TAO with 77% of the position staked

• Polychain Capital added $200M in TAO exposure

• Grayscale filed an S-1 to convert its Bittensor Trust into a spot ETF (ticker GTAO), with a decision expected by August 2026

• Grayscale raised TAO’s weighting in its AI fund from 31.35% to 43.06%, the largest single-asset reallocation they’ve ever made

• Bitwise also filed a TAO ETF in April

• BitGo and Yuma launched institutional custody and staking for Bittensor subnet tokens

Now compare that to the institutional news for memecoins in 2026 «insert tumbleweed gif here».

The biggest story is that Pump.fun is reworking its fee structure to try to bring traders back, and that they’re facing federal lawsuits.

The institutions vote with money, and the money is telling you which direction this is going. Again, it’s not everything, and institutions get it wrong all the time, and god knows there are plenty of sectors within crypto that got billions of dollars of funding that are now gasping for air. But there’s also plenty of evidence that institutional money does support genuine growth (most notably with the Bitcoin ETF(s).

I think it’s important to consider the difference between crypto-native VCs shoveling money into things like gamefi, socialfi, the metaverse, NFTs, and other sectors that struggle, and non-crypto-native companies like Nvidia voting with their wallets / company treasury.

The infrastructure layer

This is the thing I keep coming back to and one of the reasons I’ve been so excited about AI tokens. I called this out in Letter 93 (Ten Predictions for 2026) talking about the x402 protocol, and in Little Learnings #1 where I broke down ERC-8004, the onchain agent identity standard.

But the agentic payments space has changed a lot since I first wrote about x402 a few months ago, and the picture is more nuanced now.

x402 is one protocol in a much bigger landscape

When I first wrote about x402, it felt like THE answer for how AI agents would pay each other. It was the first mover, it was elegant, and Coinbase had a head start. The lazy version of this letter would be to keep banging that drum.

The reality in April 2026 is that x402 is one of several competing protocols, and its early adoption numbers are softer than the hype suggested. Recent on-chain data from Artemis shows daily volume of only around $50k and daily transaction volume actually dropped over 92% from its December 2025 peak as bot-driven activity normalized.

That doesn’t mean the thesis is wrong or the protocol is bad, but it does mean the landscape has changed a bit and the winners are less clear.

Enter Stripe’s Tempo

The biggest development you should know about, that I haven’t covered yet, is Stripe’s blockchain Tempo and its Machine Payments Protocol (MPP).

Tempo is a payments-focused Layer 1 blockchain that Stripe and Paradigm incubated together. It launched mainnet on March 18, 2026, after raising $500 million at a $5 billion valuation in October 2025. The same day they went live on mainnet, they released MPP, an open standard for AI agents to pay for services using either stablecoins or fiat.

The partner list for Tempo is genuinely insane. Design partners include OpenAI, Anthropic, Shopify, Visa, Mastercard, DoorDash, Klarna, Revolut, Nubank, Standard Chartered, Deutsche Bank, and UBS (phew). DoorDash announced last week (April 21) that it’s using Tempo to power stablecoin payouts to merchants and dashers across 40+ countries, which seems like a pretty darn big deal?

For context, Stripe processes nearly $2 trillion (TRILLION) in annual payments. They acquired the stablecoin platform Bridge for $1.1B in 2024. They acquired the wallet provider Privy. They are not messing around.

How MPP differs from x402

The two protocols are similar at the surface level. Both are designed for machines to pay machines, both can settle in stablecoin, and both use the dormant HTTP 402 “Payment Required” status code under the hood. The design philosophies are where they differ:

• x402 is permissionless and minimal. Pay per request. Each transaction settles individually. Anyone can run a facilitator. USDC on Base or Solana.

• MPP adds “sessions”. An agent pre-authorizes a spending limit, then streams thousands of micropayments that batch-settle in a single on-chain transaction. Stripe describes it as “OAuth for money”.

• MPP is rail-agnostic. It runs on Tempo today, but Visa has extended it for cards, Lightspark has extended it for Bitcoin Lightning, and Stripe has extended it for traditional payment methods. x402 is crypto-only.

• MPP launched with 100+ integrated services already including Browserbase, Parallel Web Systems, Ramp, and more. It has Stripe’s entire merchant network behind it.

If x402 is the open-source crypto-native version of agent payments, MPP is the enterprise-ready corporate-backed version. They’re not the same product even though they look similar on paper.

And it’s not just two protocols

Once you start looking at this space, you realize there are at least five major agentic payment standards in play right now, plus card network extensions:

The good news is these protocols are mostly complementary, not competitive. AP2 handles authorization. ACP handles checkout. x402 and MPP handle settlement. A real-world agent workflow might use AP2 for spend governance, ACP for vendor discovery, and x402 or MPP for actual machine-to-machine payment execution. Google’s AP2 has already explicitly integrated with x402 for crypto settlement.

The bad news is this is messy and nobody knows yet which protocol will dominate which layer, so it’s a bit tough to make speculative investment decisions in this sector — other than a spray and pray approach (which tbh is not a terrible idea, betting on the overall sector eventually taking off).

How this changes the investment picture

Three things you should take from this.

1. The bullish case for the agentic payments narrative is even stronger than I previously framed it. When Stripe raises $500M for Tempo and signs DoorDash, Klarna, Mastercard, Visa, OpenAI, Anthropic, and Shopify as partners, that is the largest fintech in the world betting heavily on this. Combine that with what Coinbase is building, and what Google is building, and what Visa and Mastercard are building, and you have every major payments player aligned on agent payments being a big freaking deal.
   

2. The thesis is broader than the original idea of “buy x402-related tokens”. The bigger picture is that stablecoin volume hit $5.7 trillion in 2025 (double the previous year), B2B stablecoin payments surged 733% year over year, and Mastercard just paid $1.8B for the stablecoin firm BVNK. The “crypto rails for AI agents” narrative is part of a much broader “stablecoins are eating payments” story that has nothing to do with memecoins or even most altcoins. This is why I keep saying the AI/payments intersection is where the real opportunity is, not in chasing the next AIXBT or Fartcoin or whatever.
   

3. Tempo specifically is a project to keep an eye on, even though it doesn’t have a token yet. Stripe and Paradigm have explicitly said Tempo will transition from a permissioned validator set to a permissionless one over time. There’s a strong chance that means a token at some point. That would be one of the more interesting launches in the space if it happens.

This is a pretty telling chart. In February 2026, stablecoin monthly transaction volume crossed the US ACH network for the first time ever, hitting $7.2T versus ACH’s $6.8T. And these are not the bot inflated raw numbers (which would put stablecoins at over $25T per year). This is Artemis’s adjusted volume that explicitly strips out MEV activity and intra exchange transfers.

If you don’t know what ACH is, it’s the rail behind essentially every US payroll, mortgage payment, and direct deposit. It’s been around since 1974. It processes roughly 29 billion transactions per year and roughly 93% of all American salary payments. It is the unsexy plumbing of the entire US economy. And it just got flipped by an asset class that didn’t exist 12 years ago… pretty mindblowing. To me it’s another great example of how fast things are accelerating in general in our world.

Visa and Mastercard understand this. That is why both are partnering with Tempo, integrating MPP, and signing up to the x402 Foundation. They would rather upgrade their infrastructure than be replaced by it.

So when I say I’m bullish on the AI infrastructure layer, I mean the whole stack of standards and rails being built right now to let agents transact autonomously. Some of it will be capturable through crypto tokens. Unfortunately (for us), a lot of it will be captured by Stripe, Visa, Mastercard, and corporate-backed L1s like Tempo. The protocol war is just getting started, and it’s going to be a very interesting space to watch.

The capital flow story

One more data point that puts this into context. In 2025, total crypto venture capital was $7.9 billion, up 44% year over year and almost half of that went into AI integrated blockchain projects.

BlackRock projects $5 to $8 trillion in AI capital expenditure between 2025 and 2030. Even a tiny fraction of that finding its way into decentralized AI infrastructure is more than enough to lift a $22bn sector materially.

By comparison, memecoin VC funding is essentially zero. There are no funds raising to invest in the next FARTCOIN (at least that I could find: if any funds are reading this with a flatulatory investment thesis, feel free to correct me).

Practical takeaways

The institutions are voting with their money

When Nvidia, Polychain, Grayscale, and the Linux Foundation all park serious money in the same sector, that’s a signal. When the other sector’s biggest news is federal lawsuits, that’s also a signal.

Memecoins still have a larger marketcap than AI tokens, but it’s concentrated, and the outlook isn’t looking so hot right now

DOGE and SHIB still make up 84% of the sector. The launchpad meta is down bad, and the odds are stacking against the millions of tokens trying to be the next FARTCOIN.

The AI sector is mostly noise too

There are ~900 projects with non-negligble market caps at the moment, and almost all of them will fail. Don’t mistake the broader narrative for an excuse to YOLO into AI memecoins. Letter 106 covered the brutal failure rates across crypto and AI tokens are no exception. Concentrate around the few names with real utility.

The agentic payment layer is the most interesting bet

This is bigger than just x402. Stripe’s Tempo and MPP, Google’s AP2, OpenAI’s ACP, plus card network extensions from Visa and Mastercard, are all converging on agentic payments. Watch the whole stack of standards and rails being built, not just one protocol.

Don’t assume the rotation is done

AI is at $22.6bn. Memes are at $36-38bn. If the trend continues at its current pace then the AI sector will overtake the Memecoin sector by the end of the year — hopefully as they both go up!

Closing thoughts

I’ve been saying for a while that AI is where some of the greatest opportunities in crypto are. Today we looked at the data to back that up in a really clear way. Capital is rotating from sentiment-driven assets to utility-driven ones, and the rotation has been quietly underway for over a year now.

Memecoins are a bet on attention. Attention is finite and rotates fast. AI tokens with real infrastructure are a bet on usage. Usage compounds. Over a long enough timeframe, the second one wins (attention will win any race in the short term, and ultimately attention is the thing that matters but the bet is that utility + usage will drive sustainable attention, whereas memetics based attention is shorter-lived).

That doesn’t mean every AI token is a buy. Most of them are still going to fail, just like nearly everything in crypto fails. But the sector itself has tailwinds that the memecoin sector simply does not have, and those tailwinds are real, sustainable, and getting stronger.

If you’re newer to AI agents and want to actually start playing with this stuff yourself, Letter 107 walks through how to set one up. Letter 108 covers how LLMs actually work under the hood. Both should help you build a better foundation for evaluating the projects in this space.

If there’s interest, I can do a deeper standalone piece on the agentic payments stack, breaking down x402 vs MPP vs ACP vs AP2 in more detail and which ones I think have the strongest setup. If that’s the kind of thing you want to read, lmk in the comments.

As always, my recommendation for most people remains the same: DCA into Bitcoin and a very small handful of other tokens, and wait. If you’re going to take some shots outside of that, I’d rather be taking those shots in AI than in the memecoin trenches in 2026.

Thanks for reading, hope you enjoyed this letter, and see you next week!

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

Whenever I talk about AI in this Newsletter, it’s usually with reference to the big cloud AI tools like Claude, ChatGPT, Gemini, etc. The way these models work is that you type a prompt, it gets sent to a server somewhere, processed, and the response comes back. Simple. This is the same whether you’re using the website interface or doing deep coding in Claude Code using your Claude Max subscription.

But there’s a whole other world out there of opensource AI that runs entirely on your own computer. These are local LLMs, and in 2026, they’ve gotten really good.

The space is, unsurprisingly, moving fast. In the past two weeks alone, GLM-5.1 became the first open-source model to top Claude Opus 4.6 on a major coding benchmark. Kimi K2.6 then dropped earlier today and took the crown over from GLM. The tooling and models keep getting better, and the gap between cloud and local keeps getting smaller.

I’ve been learning about and experimenting with local models on my Mac Studio for the last week and have been (pleasantly) surprised with how capable they. Obviously not quite as good a Claude Opus 4.7 and other frontier models for the super complex stuff, but for a lot of what I do day to day, local models are genuinely useful. And free. And private. And always available.

Even if you keep your cloud subscriptions (I do), having a local model as a backup or for specific tasks is one of the best moves you can make.

It’s also just genuinely fascinating and interesting, and learning how to own and run your own models is a really good skill to learn in this day and age.

Here’s what we’ll be covering in today’s post:

1. Why run a local model?

2. Hardware: what do you need?

3. The software tools

4. Which model for which task?

5. Getting started

6. Connecting local models to AI agents

7. Closing thoughts

If you’re interested in leveling up your AI learning journey even more, then check out the new company I have launched alongside a couple of friends: The Stoa of AI.

We create video courses and have weekly live workshops and calls that show you practical ways to implement AI into your daily workflows.

We’re in early access mode with discounted pricing, check us out here: https://www.skool.com/thestoaofai

Why run a local model?

Five main reasons.

1. Privacy. Your prompts, files, and conversations stay on your machine. No third-party servers. For anyone working with sensitive data, proprietary code, or confidential documents, this is a huge deal. Not to mention those who just care about their personal privacy and don’t want Big AI spying on them (or worse, leaking their data to nefarious actors).

2. Cost. Once you have the hardware, inference is free. If you use AI heavily, local models will often pay for themselves given enough time. You can also repurpose old devices you might have at home to run local models.

3. No rate limits. Frontier models burn through credits fast. Having a local fallback is a godsend, as is having models running tasks that will never hit a rate limit (and don’t count against your existing rate limits). Most people use a one-size-fits-all approach to AI and use models like Opus and Sonnet for simple tasks where they’re totally overkill, and a much simpler, local model can do just as well.

4. Offline access. This is a cool one. Once you have a model downloaded locally, it will work without the internet. You can interact with your model on flights, in remote areas, and just have a backup way to access the entire knowledge of humankind on your own computer.

5. Control. You get to choose the models and can tweak their configurations to your heart’s content. You won’t be surprised by a change in Terms of Service and you won’t get randomly blocked for violating terms (or cause of an error on their end). You can have full control over your entire AI stack when you run a local model.

That last one hit home a few weeks ago when Anthropic blocked OpenClaw and other third party agent frameworks from using Claude Pro/Max subscriptions. People relying on that setup were suddenly stuck having to switch to another provider or paying API costs that could easily be in the $50/day range.

Local models don’t have this problem.

As I said at the top, local models won’t match frontier models for the hardest multi-step reasoning. But for simple and everyday coding, summarization, drafting, web scraping, research, and Q&A, they handle 70-80% of what I throw at them.

The right setup is both. Cloud for the hard stuff, local for everything else.

Hardware: what do you need?

Before we get into the hardware itself a quick detour on quantization. You'll see this term everywhere in the local LLM world and it affects every hardware decision you make, so worth understanding upfront.

Full precision models store each parameter as a 16-bit number. Quantization compresses that down to 8-bit, 4-bit, or lower. The model gets smaller and faster, at the cost of a tiny bit of accuracy. A music analogy that clicked for me: FLAC is technically better than a 320kbps MP3 file, but most people can't hear the difference (I certainly wouldn’t be able to).

4-bit quantization produces output nearly indistinguishable from full precision for most tasks. If you come across models with names like Q4_K_M or Q3_K_M just know that these are referring to the same model just with 4-bit or 3-bit quantization.

The rule of thumb: a Q4 quantized model requires roughly 0.6-0.7 GB of memory per billion parameters (I explained parameters in last week’s post).

I would recommened that you stick with Q4_K_M models unless you have a specific reason not to.

Alright back to talking about hardware. The single most important number when it comes to running LLMs on hardware is available memory. This is VRAM on a PC or unified memory on a Mac. Everything else hardware related is secondary.

Here’s a handy chart for looking at the types of models you can run based on different hardware specs:

Macs have a unique advantage thanks to unified memory. The CPU, GPU, and Neural Engine share one memory pool. A Mac Studio with 512 GB of unified memory can actually run DeepSeek R1 at 671 billion parameters locally.

I’m personally running GLM5.1 at 744 billion parameters on my own Mac Studio (the Q3 version, which requires ~308GB of memory).

Mac vs PC: which should you buy?

This is a common question and the answer, like with most things, is “it depends”. Neither is universally better, they’re good at different things, depending on your situation / requirements.

• For small-to-medium models (under 24 GB), a PC with an NVIDIA GPU is faster (a lot faster). An RTX 4090 runs an 8B model at 100-140 tokens per second. An M3 Max runs the same model at 40-60 tokens per second. If you care about snappy responses and you’re only running 7B-14B models, PC wins.
  

• For large models (30B+), Macs win. Here’s why: NVIDIA consumer GPUs max out at 24 GB of VRAM (4090) or 32 GB (5090). Once your model exceeds that, the GPU has to shuttle data back and forth with system RAM over a slow connection, and performance suffers as a result. The Mac isn’t faster because Apple chips are faster, the Mac is faster because the whole model fits in memory at once.

Some other factors to take into account:

• Power and noise. A Mac Studio pulls about 60W under full load. An RTX 4090 pulls 450W plus whatever the rest of the PC uses. If you’re running inference all day, the electricity costs will add up over time. Macs are also silent. PC workstations with 4090s are LOUD. My Mac Studio sits on my desk and I rarely hear a thing from it.
  

• Price. A used RTX 3090 runs $700-900, but that’s just the GPU. You need a full PC around it which brings a realistic build to $1,500-2,000. On the Mac side, a Mac Mini M4 Pro with 24 GB unified memory starts at $1,399 as a complete machine. Macs also costs less to run daily thanks to the power difference mentioned above.
  
  At the higher level, a Mac Studio M3 Ultra with 256 GB unified memory runs about $5,999. Max it out to 512 GB and you’re at $9,500-10,000. But at that high end, there’s no comparable PC option either. A PC build that could run 671B parameter models needs multiple professional GPUs and costs $20,000+.

My recommendation, based on your situation:

If you’re on a tight budget and already have a PC: Drop in a used RTX 3090. Best value per GB of VRAM in 2026.

If you want a complete machine under $1,500 and mostly run 7B-14B models: Mac Mini M4 Pro with 24 GB ($1,399). Quiet, efficient, no assembly required.

If you want the fastest possible responses on small-to-medium models: Build a PC with an RTX 4090 or 5090. Around $2,500-3,500 total.

If you want to run 30B+ models or you want a quiet always-on machine: Mac Mini M4 Pro with 48-64 GB ($1,999-2,199) or Mac Studio with 64-128 GB ($2,400-4,500).

If you want to run the biggest open-source models (GLM-5.1, Kimi K2.6, DeepSeek R1 at full 671B) without a rack of professional GPUs: Mac Studio with 256 GB or 512 GB is the only consumer option that makes sense. Around $6,000-10,000.

What about the laptop you already own?

One thing to know before you go out and spend money is that any M1 MacBook or newer with at least 8 GB of memory can run a small local model. An M1 MacBook Air with 16 GB runs 7B models at 15-25 tokens per second, and if you have a MacBook with even more memory, you can run even more models.

These aren’t going to be anything fancy, but they can still be genuinely good for simple/basic tasks, and more importantly, you can at least get a feel for how local models work before doling out extra cash.

The Software tools

Hardware is the first step, but once you have the hardware, you’ll need some tools to manage and run the models on your own devices. Here are the main options.

LM Studio is the right starting point if you’re new to this. It’s a full desktop app with a clean and easy to use interfact. You download the installer, browse the built-in HuggingFace model library, click the one you want, and start chatting. Zero terminal commands required.

It has a live RAM monitor that tells you whether your machine can run a model before you commit to the download and will recommend the best models for you to download based on your hardware.

It also exposes an OpenAI-compatible API so you can connect it to scripts and agents if you want to (ie you can run Openclaw or Hermes agents on your local models).

Ollama is an overall better choice if you want to build things with local models, but it requires being comfortable with the terminal / command line interface (CLI). A few of the advantages of Ollama over LM Studio:

• It runs as an always-on background daemon. Install it once and it’s just there, listening on port 11434. LM Studio is a desktop app. You have to open it and flip the “Start Server” toggle every time you want API access. For agents, cron jobs, or anything that needs local AI available 24/7, Ollama is better/cleaner.

• It’s fully open source (MIT license). LM Studio is closed source and their free tier doesn’t cover commercial use. If you’re building a product or want transparency over what’s running on your machine, Ollama is the safer pick.

• Lighter memory footprint. Ollama is minimal. LM Studio is an Electron app and uses 300MB-1GB of RAM just for the GUI layer, before you’ve even loaded a model.

Ollama has the same API compatibility as LM Studio.

Ollama does have a native desktop app and it’s functional, but it’s minimal compared to LM Studio. There’s no live RAM monitor, no visual parameter controls, no side-by-side model comparison, no HuggingFace browser. It’s fine for quick chats, but not where Ollama shines. If you want a polished GUI, stick with LM Studio. If you want headless/scripting/agents, use Ollama. Or, an even better option…

You can install both! They don’t conflict, and that’s my recommendation. I’ve got LM Studio for quickly testing new models, and Ollama for anything I want to integrate into a workflow. If I had to pick one: LM Studio for a non-developer just starting out, Ollama for anyone planning to connect local models to OpenClaw, Hermes, or their own scripts.

Some more tools to know about:

Unsloth is for fine tuning models on your own data, which is a whole other very cool possibility for local models. The new Unsloth Studio released in March lets you train a model on your docs or writing style. I want to fine tune a model on all of my newsletters (or X posts) at some point and see how the model does compared to how frontier models are at writing in my own tone of voice.

HuggingFace is the repository where models live. Think of it as GitHub for AI, you don’t really need to interact directly with it but when you’re on Local LM or Ollama and you’re “downlading a model”, just know that you’re probably downlading it from HuggingFace.

llama.cpp and MLX are the engines underneath. Both Ollama and LM Studio use one or the other for inference. Most people never need to think about them.

Which model for which task?

This section literally went out of date twice while I was writing this letter. What follows is my take as of April 21, 2026. Half of this will probably be superseded in three months, if not sooner. Kimi K2.6 literally came out hours ago and I haven’t had a chance to try it myself yet, but I have used GLM-5.1 and it was probably the best choice before Kimi K2.6.

A few things to keep in mind before I share a comparison chart. The frontier open weight models (Kimi K2.6, GLM-5.1) are better than the smaller ones at almost everything. That’s the nature of bigger models with more parameters. But they need serious hardware to run locally, so for tasks that don’t need deep reasoning, a smaller model does the job at a fraction of the cost and latency. The practical question you should ask yourself is not “what’s the best model for this task” but “what’s the smallest model that handles this task well enough.”

Quick aside on benchmarks. I'll reference SWE-Bench Pro a few times in this post. It's the benchmark that matters most for coding. Instead of testing whether a model can write an isolated function, SWE-Bench Pro gives the model a real GitHub issue from an actual open-source project and asks it to fix it. The model has to read the codebase, understand the bug, write the fix, and submit code that passes the existing tests. A score of 50% means the model solved half the bugs thrown at it.

For context, Claude Opus 4.6 scores 53.4%. The newly released Opus 4.7 scores a whopping 64.3%. Anything in the 55-60% range is generally considered frontier, but that number is obviously constantly changing as frontier models get better.

On the top shelf for coding, two open-weight models stand out as of today.

Kimi K2.6 from Moonshot AI is the new king of the open-source coding world. It came out today. It is purpose built for long and complex coding tasks. Where other models start losing coherence after an hour or two, K2.6 has demonstrated 5 day continuous execution runs on real engineering tasks.

It can also orchestrate 300 sub-agents in parallel (wtf), which means you can throw something like “refactor this entire monorepo” at it and it’ll decompose the job across hundreds of specialized workers. It beats Claude Opus 4.6 on SWE-Bench Pro (58.6% vs 53.4%). If you’re building anything agentic or doing heavy codebase work, this is the best local model at the moment (but again.. this could all change by literally tomorrow lol).

GLM-5.1 from Z.ai is the older option (April 7, which is crazy to consider it as the “older option”), but still close on coding quality. It scored 58.4% on SWE-Bench Pro, so only sliiiightly worse than K2.6. Another great pick if you want frontier coding but don’t have the hardware to run the Kimi model.

On the practical side, Qwen3.6-35B-A3B (released April 16) will hit the sweet spot for most people. The MoE architecture means only 3B parameters are active per token even though the model is 35B total, which means it runs fast on a 24 GB machine. It handles images and video, not just text, and it has a context window that goes up to 1M tokens so you can feed it entire codebases or long documents.

It’s good at everyday coding, writing drafts, summarization, and agent workflows.

This is random but someone tested it on their laptop against Claude Opus 4.7 the day both were released, and the local model drew a better pelican riding a bicycle (very random and silly example but what’s life without a little whimy):

For smaller hardware, Qwen 3.5 9B is the practical option and runs fine on an 8 GB MacBook. It’s not going to handle complex multi file reasoning, but for some daily tasks (rewriting emails, summarizing articles, quick Q&A), it’s remarkably capable.

Getting started

If you want to try running your own local models, here are some instructions to get started for both LM Studioo and Ollama.

LM Studio:

1. Download LM Studio from lmstudio.ai.

2. Install it.

3. Open the app.

4. Click “Discover” and search for a model. The live RAM monitor tells you whether it will run on your machine.

5. Click download.

6. Click “Load model” when it’s done, and you’re off to the races. You can chat with the model directly in LM studio, or connect it to an agent like openclaw/hermes (i’ll explain how in the next section).

Ollama:

1. Install Ollama from ollama.com (one-line installer for Mac and Linux).

2. Then head to ollama.com/library or huggingface.co to browse models.

3. Every model listing should give you the exact command to run it. HuggingFace has a wider selection and shows you the file size so you can check it against your RAM before downloading.

4. Once you’ve found your model, run it in the terminal, it should look like this:

ollama run qwen3.5:9b

The first time you run a command like this it’ll download the model, then after that it’ll load the model from your hard drive. Once it’s downloaded/loaded, you can start talking to it immediately from the terminal.

It’s surpsingly simple to get up and running with local models. The whole setup from start to finish doesn’t take long, usually the longest part is downloading the model itself (a few GB to tens/hundreds of GB depending on the model).

This is literally all you have to do to have local LLMs running entirely on your own devices, and I recommend everyone with the hardware to at least give this a shot with some of the smallest models.

Connecting local models to AI agents

This is where things get interesting. Running a local chatbot is useful and cool and all that, but connecting a local model to an agent framework (Openclaw or Hermes) is the real unlock.

OpenClaw: Install OpenClaw, then in Settings > Config (or openclaw.json) add a custom provider pointing at http://localhost:11434/v1 for Ollama or http://localhost:1234/v1 for LM Studio. Set the API type to “openai-completions” and give your model a name that matches what’s loaded.

Hermes Agent: Install Hermes, then run hermes model to open the setup wizard. Choose “Custom endpoint”, enter your local URL (same as above: Ollama is http://localhost:11434/v1, LM Studio is http://localhost:1234/v1), and pick the model you’ve loaded. Switch models later with /model in chat.

Both Ollama and LM Studio expose OpenAI-compatible APIs, and both OpenClaw and Hermes speak that format, so it’s all pretty simple at the end of the day. Once you get it figured out once, you’ll find it very easy to try new models.

Closing thoughts

A lot of content about local LLMs out there tends to overhype things. While I don’t think everyone has to be using local models, and I very much understand the limitations of these models, I do think everyone passionate about AI would benefit themselves greatly by taking a day or two to tinker around here.

A local model is not going to replace Claude Opus 4.7 for complex multi-step reasoning. It’s not going to write content as well as frontier cloud models. It’s not going to debug a gnarly multi-file codebase as reliably.

What it will do is give you a private, free, always-available AI assistant that handles the majority of basic tasks that you throw at it, and, apparently, do a better job of creating an image of a pelican on a bicycle, sometimes?

And for a lot of people, that’s more than enough.

The quality curve is obviously real, and not all local models are created equal. Going from 8B to 14B is a noticeable jump. 14B to 32B is another. If you have the hardware for Kimi K2.6 or GLM-5.1 on a 512 GB Mac Studio, you’re running a model that beats Claude Opus 4.6 on SWE-Bench Pro. For normal hardware, Qwen3.6-35B-A3B on a 24-32 GB setup is the sweet spot in April 2026. You get near frontier quality on a standard machine.

The best overall approach I recommend for everyone is cloud for your hardest tasks, local for everything else (or for things that must be private). You don’t have to pick one or the other.

The local LLM ecosystem in April 2026 is mature. The last couple of months has seem pretty incredible leaps in quality, and if this trend continues, it’s gonna be absolutely mindblowing the AI power that us mere mortals can wield at home.

Honestly, these models are probably already better than you’d think. And having an AI that runs on your own machine, answering your questions with (near) zero running costs and zero data leakage, is yet one more of those things these days that makes me think of these fine words from one of sci-fi’s greatest writers:

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.

But for today, we’re taking a foundational look at LLMs. I’ve noticed that most people who use ChatGPT or Claude every day have no idea how they actually work.

Which is fine of course. You don’t need to know how an engine works to drive a car. But I think having a basic understanding of what’s going on under the hood makes you a better user. It helps you understand why the AI is good at some things and bad at others, it helps you ask better questions, and it makes you less likely to either over-trust or under-trust the outputs.

This turned into quite a long piece, here’s what we’re going to cover:

1. What is an LLM?

2. How does an LLM “learn”?

3. Wait, so it’s autocomplete?

4. What are tokens?

5. What about parameters?

6. How do LLMs actually generate their responses?

7. And what do these models really “know”?

8. What is training vs fine tuning?

9. Why are some models better than others?

10. Model sizes: why some run on your laptop, and others need data centres

11. How does knowing all of this help you?

If you’re interested in leveling up your AI learning journey even more, then check out the new company I have launched alongside a couple of friends: The Stoa of AI.

We create video courses and have weekly live workshops and calls that show you practical ways to implement AI into your daily workflows.

We’re in early access mode with discounted pricing, check us out here: https://www.skool.com/thestoaofai

What is an LLM?

LLM stands for Large Language Model. That’s what ChatGPT, Claude, Gemini, and all the other AI chatbots are built on.

• Language. These models work with language. Text in, text out. You type words, they generate words back. (Yes, they do images and audio and code now too, but at their core, they are language machines, and the word “language” can be used for whatever input/output is being generated by these LLMs).

• Model. In AI, a model is a program that has been trained on data to recognize patterns. If you showed someone who had never seen a cat, a million photos of different breeds of cats, eventually they’d get great at telling the difference between them. An LLM is the same concept.

• Large. These models are LARGE. They are trained on enormous amounts of data. We’re talking about a significant chunk of the entire internet. Books, articles, Wikipedia, forums, code repositories, academic papers. Billions, maybe trillions, of words.

Put it together and you get: a program that has read a huge portion of human text and learned the patterns of language from it.

How does an LLM “learn”?

The core learning of “training” process is surprisingly simple in concept. You take a sentence, hide the last word, and ask the model to predict what comes next.

“The cat sat on the ___”

The model guesses. If it gets it wrong, you adjust the model slightly so it does better next time. Then you do this billions and billions of times, with billions and billions of sentences.

Over time, the model gets good at predicting the next word. And then the next word after that. And the next. Until it produces entire paragraphs and pages that sound like a human wrote them.

This is a simplified version of the process (the technical term is “next token prediction”), but it captures the core idea. LLMs are, at their foundation, prediction machines. They predict what text should come next based on everything they’ve seen before.

Wait, so it’s autocomplete?

Kind of. This is a comparison that gets thrown around a lot, and it’s partially accurate.

Your phone’s autocomplete predicts the next word based on simple patterns. LLMs do the same thing, but with astronomically more data, vastly more computing power, and a much deeper understanding of context.

The difference in scale creates a difference in kind. Your phone’s autocomplete might suggest “the” after “in.” An LLM will write you a coherent essay about quantum physics, maintain a consistent argument across 2,000 words, and format it properly. Both are predicting the next word. One is doing it with such depth and sophistication that it produces something that looks and feels like understanding.

Whether or not it is understanding is one of the great debates in AI right now. I don’t think we need to settle it here. What matters for practical purposes is that the output is useful, and often impressively so.

What are tokens?

Tokens are the units that LLMs work with, and they’re also sorta considered the currency of LLMs. When you use a frontier model from Anthropic or OpenAI, you’ll generally be paying per token used.

Here’s something most people don’t realize: the model never actually sees your words. It only sees numbers.

When you type a message, the first thing that happens is your text gets encoded into tokens, where each token is assigned a number. The word “hello” might become token 15339. The word “the” might be token 1820. The word “cryptocurrency” might get split into two tokens: “crypto” (54219) and “currency” (26072).

These numbers are what the model works with. Every single computation that happens inside the model, all the pattern matching, all the predictions, happens as math on numbers. The model processes these numbers through its neural network, and outputs... more numbers. Those output numbers then get decoded back into words that you read on your screen.

Encode → Math → Decode. That’s the whole loop.

The process of converting text into numbers is called encoding. The process of converting the output numbers back into text is called decoding. You never see the numbers, and the model never sees the words. There’s a translation layer (called a tokenizer) sitting between you and the model, encoding and decoding back and forth.

So what happens during the “math” part? Each token number gets converted into a vector, which is a long list of numbers (hundreds or thousands of them) that represents the meaning and context of that token. The word “bank” in “river bank” gets a different vector than “bank” in “bank account” because the surrounding tokens influence the representation.

The model then runs these vectors through layers of calculations, adjusting and combining them, comparing every token to every other token to figure out relationships and context (this is the “attention” mechanism you might have heard about). After dozens of these layers, the final output is a probability distribution over every possible next token. The model picks one, decodes it back to text, and voila! you see a word appear on your screen.

This is also why LLMs are sometimes weird about things like counting letters in a word or doing arithmetic. The model doesn’t see the word “strawberry” as s-t-r-a-w-b-e-r-r-y. It sees it as one or two token numbers. It has no concept of individual letters because those letters got encoded away before the model ever touched them.

A token is roughly 3/4 of a word, or about 4 characters. Common short words like “the” or “and” are one token. Longer or less common words get split into multiple tokens.

This matters to you because LLMs have a limit on how many tokens they process at once. This is called the context window. Think of it as the model’s working memory.

If a model has a 200,000 token context window, that’s roughly 150,000 words it is able to hold in mind at one time. Some models now go even higher. Claude Opus 4.6, Claude Sonnet 4.6, and Gemini all support 1 million token context windows. That’s about 750,000 words, or roughly 10 to 15 full novels. Llama 4 Scout from Meta supports a whopping 10 million token context. These are staggering numbers compared to where things were a few years ago.

But something to keep in mind is that larger context windows aren’t necessarily or inherently better.

As you stuff more and more tokens into the context window, the quality of the model’s responses tends to degrade. Researchers call this “context rot.” The model doesn’t attend equally to everything in its context. It tends to pay the most attention to stuff near the beginning and the end, and less attention to stuff in the middle. A 2023 research paper found that when relevant information was buried in the middle of a long context, models performed significantly worse at finding and using it.

This means that giving the model more context isn’t always better. If you dump 500,000 tokens of loosely related documents into the context window and your actual question relates to a detail somewhere in the middle, you might get a worse answer than if you had only provided the 10,000 most relevant tokens. Quality context beats quantity of context. It’s a little counterintuitive, but it’s how it works.

Just like everything AI related, the models are getting better at this too. Claude scores at the top of long-context benchmarks, and the gap between small-context and large-context performance is shrinking with each generation.

What about parameters?

This is another big number you hear about. Lots of models tout billions or hundreds of billions of parameters; some have trillions. But what even are parameters?

Parameters are the model’s internal settings. Think of them as tiny dials, and during training, each of these dials gets adjusted slightly every time the model makes a prediction and gets feedback on whether it was right or wrong.

To put it more concretely, parameters are the numbers that determine how the vectors mentioned in the previous section get transformed as they pass through the model. They control things like: how much attention should this word pay to that word? How should this concept relate to that concept? What patterns are important and what patterns are noise?

Every connection between neurons in the neural network has a parameter (a weight) that controls the strength of that connection. A 7 billion parameter model has 7 billion of these connections. A trillion parameter model has a trillion. Each one was tuned, bit by bit, over trillions of training examples.

A model with more parameters has more dials to tune, which means it has the capacity to learn more subtle and complex patterns. A small model might learn that “the cat sat on the mat” is a common pattern. A large model learns that too, but it also learns that the sentiment of a paragraph shifts when you use the word “however,” or that a question phrased politely tends to expect a different kind of answer than a blunt one. The larger the model, the more of these subtle relationships it picks up.

More parameters generally means a smarter model, though it’s not the only factor. Training data quality, architecture decisions, and fine-tuning all matter too, and we’ll talk about that in a bit. But all else being equal, more parameters = more capacity to learn complexity.

The tradeoff is resources. Every parameter takes up memory. Running a model means loading all of those parameters into RAM (or GPU memory) and doing math on them for every single token generated. That’s why bigger models need more expensive hardware, cost more to run, and generate tokens slower.

You don’t really need to remember the exact numbers or know how things work precisely for this stuff.

The takeaway is: parameters = the model’s capacity to learn complexity.

How do LLMs actually generate their responses?

When you type a message to Claude or ChatGPT, here’s roughly what happens:

1. Your message gets converted into tokens (numbers)

2. The model processes those numbers through its neural network (the billions of parameters)

3. It predicts the most likely next token (number)

4. That number gets added to the sequence, and the model predicts the next one

5. Repeat, one token at a time, until the response is complete

This is why you see the text appear word by word when the AI is responding. It’s generating the response in real time, one piece at a time. It doesn’t write the whole answer and then reveal it. It is figuring it out as it goes.

This is also why the same prompt sometimes gives you different answers. There’s a degree of randomness (called “temperature”) built into the selection process. The model doesn’t always pick the single most likely next token. Sometimes it picks the second or third most likely option, which sends the response in a slightly different direction.

On most models you can also adjust this setting and request the model to use more of the less-standard responses. This is helpful if you’re doing something like creative writing or anything really where you need outside-of-the-box thinking. For anything that requires facts and exactness, low temperature models tend to perform better.

And what do these models really “know”?

LLMs don’t have a database of facts that they look up. They don’t search through a filing cabinet when you ask them a question. Instead, the knowledge is embedded in the patterns of their parameters. The model learned that certain facts tend to appear in certain contexts, and it reproduces them when the context calls for it.

This is why LLMs sometimes make things up. The AI community calls these “hallucinations.” The model isn’t lying. It’s generating text that seems like the most probable continuation of the conversation, and sometimes the most probable-sounding thing isn’t true. It’s predicting, not recalling.

This is one of the most important things to understand about LLMs. They are optimized to produce text that sounds right. Not text that is right. These two things overlap a lot of the time, but not always.

Rule of thumb: the more obscure or specific the fact, the more likely the model is to get it wrong or make it up. If you ask about well-documented topics that appeared frequently in the training data, the model is pretty reliable. If you ask about niche topics, recent events, or specific numbers, then verify the output.

What is training vs. fine-tuning?

Training is the initial process where the model reads all that text and learns the patterns. This is expensive and time consuming. Training a frontier model costs hundreds of millions of dollars in compute alone (this phase is sometimes called pre-training because it happens before any further refinement).

The result of pre-training is called a base model. Base models are smart, and they know a lot about language, but they’re weird to talk to. If you ask a base model a question, it might continue your text as if it’s writing a Wikipedia article, or generate random forum posts, or complete your sentence in a direction you didn’t expect. It doesn’t know it’s supposed to be helpful. It’s a text prediction machine, but not a conversational assistant like we’re used to engaging with like with chatGPT etc.

Fine tuning is what turns a base model into something useful. It’s a second round of training, done on a much smaller and more carefully curated dataset. This is where the model learns to follow instructions, answer questions, have conversations, and generally behave the way you’d expect a chatbot to behave.

There are a few different types of fine-tuning worth knowing about:

• Instruction tuning is where you train the model on thousands of examples of “here’s an instruction, here’s the correct response.” This teaches the model to follow directions instead of completing text randomly.
  

• RLHF (Reinforcement Learning from Human Feedback) is where humans rate the model’s outputs, and the model learns to produce responses that people prefer. This is a big part of why modern chatbots feel natural to talk to. The model learns things like “be concise when the question is simple” and “acknowledge uncertainty when you’re not sure” from human preferences.
  

• Domain-specific fine-tuning is where you take an existing model and train it further on data from a specific field. A hospital might fine-tune a model on medical records so it becomes better at clinical language. A law firm might fine-tune on case law. A company might fine-tune on their internal documentation so the model understands their products and processes. This is where things get interesting for businesses.

The cost difference between pre-training and fine-tuning is enormous. Pre-training GPT-5 or Claude from scratch costs hundreds of millions. Fine-tuning an open source model on your own data costs anywhere from a few dollars to a few thousand, depending on the size of the model and how much data you’re using.

This is one of the reasons open source models matter so much. You take a free base model like Llama or Mistral, fine-tune it on your specific data, and you end up with a custom model that understands your domain, runs on your own hardware, and costs nothing per query. That’s a big deal for businesses that process a lot of data and don’t want to send it to a third party API.

Why are some models better than others?

We covered this a bit already, but just to highlight and expand upon a few factors in a bit more detail:

• Training data quality. More data isn’t always better. Cleaner, higher-quality data leads to better models. If you train on a lot of garbage, the model produces garbage.

• Model architecture. How the model is structured internally matters. The Transformer architecture (introduced in a 2017 paper by Google researchers called “Attention Is All You Need”) is the foundation for all modern LLMs. There are meaningful differences in how each company builds on top of that foundation.

• Scale. More parameters and more training compute generally lead to better performance, up to a point. There are diminishing returns, and smaller models trained on better data sometimes beat larger models trained on worse data.

• Fine-tuning and alignment. How the model is refined after initial training makes a huge difference in how useful it feels to talk to.

• Context window size. How much the model keeps in mind during a conversation affects its ability to handle complex, multi-part tasks.

Model sizes: why some run on your laptop and others need a data center

As we’ve established, not all models are the same size. Parameter count varies hugely, and that directly determines what hardware you need to run them.

A rough rule of thumb: each billion parameters needs about 0.5 to 1 GB of RAM (depending on the precision/quantization). A 7 billion parameter model needs around 4-8 GB of RAM. A 70 billion parameter model needs around 40 GB. Frontier models from OpenAI, Anthropic, and Google have hundreds of billions to over a trillion parameters, and they require massive clusters of specialized GPUs that cost millions of dollars.

This is why some models are available to run locally on your own computer, and others are only accessible through cloud APIs. You pay per token to use GPT-5 or Claude because the infrastructure required to run them is enormous. But you download and run Llama 8B or Mistral 7B on a decent laptop for free.

There’s also a technique called Mixture of Experts (MoE) where a model has a huge total parameter count but only activates a fraction of them for each token. DeepSeek V3 has 671 billion total parameters but only uses 37 billion per token. GLM-5.1 has 744 billion total but only 40 billion active. This lets big models run on smaller hardware than you’d expect.

The quality gap between the smallest and largest models is real, but it’s also shrinking. A well-chosen 14B parameter model running on your laptop today can do a decent job of everyday and simple tasks even compared to the frontier models (although it might be slower).

The gap is most noticeable in complex multi-step reasoning, long creative writing, and tasks that require a lot of world knowledge. For everyday stuff like drafting emails, summarizing documents, or answering questions, the local models are surprisingly good.

And of course not all local models are created equal. GLM5.1 is a surprisingly good model that can run on a mac studio, which, while still very expensive ($5-10k+), pales in comparison to the millions of dollars the massive data centres cost to build.

How does knowing all of this help you?

I mean, hopefully you just find this stuff super interesting like I do! There’s value in knowledge, and in knowing how things work, even if you don’t really need to know in order to use them.

Some of that value comes from some changes you might make when using these tools.

When you know that the model is predicting the next token based on patterns, you understand why giving it more context leads to better outputs. You understand why being specific in your prompts matters. You understand why it sometimes confidently says things that are wrong.

When you know about context windows, you understand why long conversations sometimes go off the rails.

When you know about temperature and randomness, you understand why regenerating a response sometimes gives you something better (or worse). It’s a different path through the probability space. And knowing that you can adjust the temperature settings depending on the task can allow you to harness these tools in a way that is specific to your needs.

You also start to appreciate what these tools are and aren’t. They’re not search engines (though they now have search built in). They’re not databases. They’re not oracles. They’re pattern-matching machines of extraordinary sophistication, trained on a large percentage of humanity’s written knowledge (and then further trained / fine tuned with additional and curated human feedback).

That makes them useful.

It also makes them fallible in specific, predictable ways.

Knowing all of this should make you a better user, and give you more confidence in your prompting future.

Disclaimer: The content covered in this newsletter is not to be considered as investment advice. I’m not a financial adviser. These are only my own opinions and ideas. You should always consult with a professional/licensed financial adviser before trading or investing in any cryptocurrency related product. Some of the links shared may be referral links.