🐛 (@flashfx) profile picture

markets, crypto, AI

Joined July 2010
Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)Photo by 🐛 (@flashfx)
🐛 retweeted
Some important things to understand about market structure: -fundamentally nothing has changed since 10/10 re: alts and arguably has gotten worse -as such, liquid funds can’t size as big on alt allocations. This was the main structural bid on “quality alts” in the past. Most people are just trading momentum/trend -Capital / margin is more spread out because of the growth of cross-marginable stock / RWA perps. Fiat stables haven’t grown their market cap in some time now -leakage: Hacks, exchange fees, launchpads, grifters. Same as before but recent uptick in hacks due to AI cyber capabilities -so who has money? (1) Basis/systematic traders who got favorably ADL’d on 10/10 ie not a source of bid on market beta (2) HYPE airdrop recipients/holders, who are mostly losing traders / momentum types (3) hackers who sell coins -> fiat These aspects all contribute to crypto trading like absolute garbage outside of some short periods where people are forced to chase aggressively trending coins
34
41
8
650
54,792
Really enjoyed this
My full conversation with @based16z 0:19 the Hyperliquid call 5:01 almost losing his $HYPE airdrop 8:27 life after life changing money 10:12 dating a billionaire's daughter 14:37 from online poker to Crypto 16:40 two trades made his year 20:56 the Snapchat trade 27:43 Sydney Sweeney $AEO trade 31:14 posters move markets 41:42 loot and rob 44:23 advice for a 20 year old 47:23 math lore 51:49 why he has no social media
2
76
Leopold sold a product that people paid significant fees to buy. Lots of talk about how Leopold took too much risk -- he took exactly as much risk as the insane $20-$30bn of capital demanded him to take. The lesson isn't in what Leopold did (or failed to do) in his investing strategy -- the lesson is how institutions who manage money at a level that can be a meaningful check in a $30bn AUM fund didn't see the risks he was taking. FOMO is a crazy thing.
There are multiple ways to create "leverage" in investing. Any good PM will be very deliberate about how they mix them together: 1) Borrowing money (margin) 2) Levered positions (options, swaps) 3) Underlying companies that are levered (including de facto levered) 4) Volatility 5) Illiquidity 6) Concentration Leo put all six of these together, and did it in an EXTREME way on at least five of them. Two is dangerous. But all SIX!? Six and your ride will be insane. On that ride, if you are all gas, no brakes, there is only one way it will end... even if you would have been "right" in the end. Splat.
2
40
13,712
🐛 retweeted
Indexing isn't passive investing
8
10
10
86
142,030
Great to see this. ETF structure will help with tax efficiency, and I’d expect AUM increase significantly over time given how strongly flows have shifted toward ETFs versus mutual funds. Crazy that this only has $75m in assets with the track record @gatorcapital has.
ETF Merger/Acquisition Filed Gator Capital Long/Short ETF Ticker: tba Fees: 3.01% Effective date: targeting Dec. 11, 2026 Target: Caldwell & Orkin - Gator Capital Long/Short Fund - $COAGX / $GATRX - fees remain same at 1.00% management fees and 3.01% including interest and dividend expsense. Reorganization of Gator Capital Long/Short Fund into an exchange traded fund structure. ETF prospectus: sec.gov/Archives/edgar/data/… Mutual fund prospecutus: sec.gov/ix?doc=/Archives/edg… Filing: sec.gov/Archives/edgar/data/… #ETF #ETFmerger @gatorcapital
2
110
The best AI regulation the government could write right now is to simply enshrine the concept that a meat sack needs to call up another meat sack at AWS to get more compute provisioned above a certain level That alone solves basically every doomsday scenario
3
6
3
63
22,778
a lot of people – maybe most people – secretly yearn for the apocalypse as manifestation of status anxiety this conceit has many faces – marxism, climate fear, AI x-risk, prepper culture, anti capitalism – but it's the same psychological phenomenon. secular millenarianism
81
88
16
1,126
89,074
Two similar arguments about the neutrality, or even positive effects, of inflation are both sleights of hand: “median wages kept up with inflation” and “60% of Americans own stocks.” Both imply that inflation has been neutral or positive for most Americans, and that people who feel harmed by it are simply mistaken. Both use reassuring aggregate statistics to wave away very real distributional effects. “Median wages kept up with inflation” The fundamental problem with this argument is that it takes an aggregate outcome and pretends it describes individual experience. If your wages rise 10% and prices rise 10%, you didn’t get richer. At best, you maintained your purchasing power. If prices rose before your wages did, you spent some period getting poorer and eventually caught up. And if the economy became more productive in the meantime, a flat real wage means you didn’t share in that improvement through your wages. More importantly, your wage is only one part of your economic life. What happened to your cash savings? Your bonds? Your pension? Your rent? Your taxes? Your insurance? Your mortgage? Inflation doesn’t raise every price, wage and contract simultaneously. It moves through the economy unevenly and with long lags. People have radically different exposures to it. A homeowner with a 3% fixed-rate mortgage experiences inflation very differently from a renter. A heavily indebted borrower experiences it differently from a retiree living on fixed-income assets. A worker with bargaining power experiences it differently from someone whose salary resets once a year. Those differences matter not only because they redistribute income and wealth, but because they change behavior. People shorten contracts, demand inflation protection, borrow more, become less willing to lend, seek greater wage bargaining power, join unions, buy inflation hedges and otherwise try to protect themselves from being on the losing side next time. “Median wages kept up” tells you almost nothing about any of this. It takes the person standing in the middle of the distribution and uses him to erase everyone else. “60% of Americans own stocks” The stock argument is worse. “60% of Americans own stocks” puts someone with $10,000 in a 401(k) in the same bucket as someone with $10 million in equities. Those people do not have remotely the same exposure to rising asset prices. For most households, their most valuable economic asset is their future labor income, not their stock portfolio. A bump in a retirement account doesn’t necessarily compensate someone for the declining purchasing power of wages, cash and fixed-income savings, much less pay the higher grocery, housing and insurance bills arriving today. It also ignores why people hold different portfolios in the first place. Consider an elderly person who needs her savings to fund near-term living or medical expenses. She cannot necessarily put everything into equities and ride out a five-year bear market. Having trusted the central bank’s commitment to price stability, she may quite rationally have accumulated cash and bonds. Unexpected inflation reduces the real value of those savings. What exactly is the argument here? That she should have known better? That she should have taken more equity risk? That her loss doesn’t count because 60% of Americans technically own some stock? That turns a failure to maintain the purchasing power of money into a failure of the victim to construct the right portfolio to defend against it. And stocks themselves are hardly a guaranteed inflation hedge. Inflation can raise nominal asset prices, but it can also compress valuations, raise discount rates and produce terrible real equity returns. Saying people are protected from inflation because they own risky assets is a peculiar defense of price instability. The trick in both arguments is the same: point to the thing that adjusted upward and declare everyone whole. But people don’t live inside the median. They live in particular households with particular wages, debts, savings, rents, pensions, portfolios and consumption needs. Inflation changes the value of all of those things differently and at different times and that creates winners and losers. Median real wages can be flat while millions of people suffer substantial real losses. A majority of Americans can own stocks while most of the gains from rising asset prices accrue to the minority who own vastly more of them. The relevant question is not whether wages rose or whether most people own some stock. It is who gained and who lost claims on real resources. That is also why dismissing anger about inflation as ignorance or “sour grapes” is politically dangerous. People experiencing these losses do not need an economist to tell them that their particular combination of wages, savings and expenses left them worse off. Showing them a chart of median real wages does not make their loss disappear. Telling them they benefited from an asset boom because there is $10,000 of stock in their 401(k) can be almost insulting. You can be deeply concerned about the populist politics that inflation helps produce without pretending that the grievances underlying it are imaginary. In fact, trying to explain those grievances away with carefully selected aggregate statistics may make the problem worse. People who believe the system treated them unfairly become more hostile to its institutions when those same institutions appear to be telling them that what they experienced didn’t happen. Inflation’s distributional effects are not a misunderstanding of inflation. They are among its primary real effects and gaslighting about this likely exacerbates its polarizing political effects.
19
50
13
247
61,237
🐛 retweeted
The whole plumbing meme is quite funny because it's one of the best analogies for what you should be doing in markets. The plumber gets paid precisely because the job is the one nobody wants to touch. He is useful to the system he operates in and he could tell you in one sentence what he's being paid for. So the question to run in markets is the plumber's question: what problem am I solving here and is it unpleasant enough that most people are quietly avoiding it? Most people in trading can't answer that question. They stare at charts, name patterns and build frameworks around things that nobody is compensating them for. There is no problem being solved and no one on the other side who needed it done. They're doing work, but they're not doing anyone's plumbing. We all should strive to be like plumbers.
Threadguy says crypto OGs were trading against plumbers and wouldn’t last in today’s era "Alt season bro? The fabled alt season where every altcoin goes up 20x and it just has to happen this way? 2017 ICOs?” “Some of the best were the best don’t get me wrong. But it’s like, wake up unc, cented owns 60% of the supply and he’s dumping on your headtop. Every new narrative is controlled. The game has evolved.” "The new game is not gonna happen on Twitter. Gone are the days where you have influence just because you've been here for ten years and you're a funny poster.”
16
13
5
230
25,007
This is all true but I would just point out that in basically all of the examples below, the end state of these markets is not materially cheaper or more efficient for consumers, or at least it’s not obvious this is true. A recent example is prediction markets. I actually thought that “peer to peer” betting would or could be much cheaper for participants. The prediction markets definitely shilled this narrative. Vegas sets lines to make ~10% in expectation. My (admittedly naive) thinking was, ok well if *anyone* can now make markets, someone will come in and do that for 5% or maybe even 1% or whatever. Surely they would be more liquid as well. In reality, the cost to trade sports (the biggest markets btw) now on PM and Kalshi is getting very close to this 10% rake we see on sportsbooks. It took roughly 2 years to be back at sportsbook prices More examples: -remember $10 Uber rides? Most Ubers now seem to cost about the same as a taxi. The cost did not *come down* to have someone drive you around, it became maybe marginally more convenient but certainly not an order of magnitude cheaper -AirBnBs are now either as expensive as hotels or they are annoying enough to deal with that you realize why hotel rooms cost what they cost. you may be able to get an Airbnb in a city for $100 where a comparable hotel room is $150, but you will have to deep clean the shower, take the trash out, wash the sheets and then maybe leave out milk and cookies for the people who clean after you just cleaned. The thing about regulatory arbitrage is that there’s very often no real gains or actual innovation occurring. And, because a lot of these companies are VC funded / have big marketing and ~incentive~ budgets, the market is distorted in the beginning: $10 subsidized Uber rides do not persist because that is not actually the market clearing price. There is some lower bound on cost to 1. Have a vehicle that requires maintenance and insurance and ongoing opex (fuel / inspections / whatever) 2. The value of someone’s labor to drive you around This applies in some capacity to almost all of these reg arbs. Maybe someone just will not quote both sides of a sports market for much less than 10% 🤷‍♀️. It’s probably just not worth it let someone stay in your vacant condo for the night for $25. Compare Uber to something like Waymo that is a *real* technological change, this is what is typically needed for orders of magnitude cheaper products or services. (Not saying this is a certainty, there are tons of bullshit regulations that will crop up with FSD that themselves will need to be arbed) In almost all the examples we think of, incumbents are supplanted by new incumbents and the idea of “consumer surplus” does not really come into fruition. This requires real innovation and real progress because eventually with enough demand and public support, they just make all this stuff legal, there’s no regulation to arb. The only real takeaway from building a business on this premise is that you really want to be the winner because then you become the incumbent and you have pricing power (and new regs that provide a new moat) and can just go back to charging what the previous incumbent charged
Joked about reg arb on podcast today. But this isn't unique to crypto: > uber skirted medallion laws > airbnb ran unlicensed hotels > paypal moved money without MTLs > youtube copyright arbitrage > draftkings/fanduel used skill game loophole This is classic startups. Find a rule that protects incumbents more than consumers, grow faster than enforcement, turn users into constituency, help write the new rules. Nothing new under the sun.
21
10
131
23,090
The most avg llm is "super-intelligent" now. But we really should ask what we mean by getting smarter. Is that what is happening? LLM's are superhuman on breadth of recall (they have the internet of knowledge at their fingertips) and effectively speed of very competent information synthesis (they can respond w reason very fast). Would be great at jeapordy vs someone trying to compete w google search who couldn't be close to fast enough as the best human contestants. But what we are talking about with the frontier game should not be framed as some physics take on economic value of accuracy. Gerstner and the recursion gang talk as if the frontier is racing up an intelligence mountain to Ultron. I don't think this is the case. What the frontier models are actually selling are fewer supremely confident wrong answers and better prompt instruction following. Basically, less need for a human to intervene ever. That's not getting "smarter" in my book even though i'm sure plenty folks will disagree. That's really just error-rate compression. Not that this isn't a big deal as its kind of what determines when no human is needed at all for the most complex tasks, but economically u are chasing something different then the framing which shud be the focus. I think most people would agree that error-rate compression has terrible scaling economics. Going from 90% to 99% costs enormously more than 0% to 90% did, and 99% to 99.9% is brutally worse. What's mining the last bitcoin cost ? Zero or a Zillion? Does it even matter? This is why you should seriously question the math on the last mile of "Smart" because that's not how the economic value of "Work" is measured. You should be asking how much economically valuable work lives in the last 10% or 1%? Is 93% + human reviewer the economic sweet spot? Also, self-driving isn't the same as knoweldge work but its probably good to think about in this debate cause it has much more definable kpi's when it comes to error rate compression then what you are chasing in the last mile of most knowledge work. Miles w/o a crash or fatal collision is easier to measure then a false citation in some corporate legal filing that gets a case flipped. But we as a society accept human error in driving and we historically have not tolerated machine error in same way. But human accountability is clearly a big deal in a lot of the last mile of value in knowledge work. You can get a model to do all the m&a deal work and still pay ur huge fees for someone to hold accountable. So, does that lead really have an economic fly-wheel? At minimum its worth pondering way more then most people been doing lately.
Brad Gerstner @altcap made the boldest, most non-consensus point of the pod, that the frontier's lead widens over time, it doesn't converge. "The non-consensus argument might be that intelligence is not converging at all, that superintelligence becomes fully recursive. As it becomes recursive, you actually extend the lead, because the smarter your model gets, the more revenue you get, the more compute you can buy, the better the model you can build... The distance between the frontier and everybody else doesn't converge, it actually extends."
4
4
83
20,368
GPT-5.6 dropped today. Everyone will spend the week arguing about benchmark deltas. But the benchmarks haven't told the full story of these models for quite some time. 5.6 is the last model in the 5.x line and another post-training pass on the same base OpenAI has used since April, a pretrain codenamed Spud that leakers estimate at around 4 trillion parameters. Claude Fable 5 is widely believed to be a substantially larger pretrain. Nobody confirms parameter counts and all of this is leak-grade information, but it's likely directionally correct. What this means is that these two models both represent frontier intelligence that's been arrived at in very different ways. 5.6 is mostly new RL on an existing base. Fable is mostly new scale. Why does this matter? I've been using Fable since it came out and something about it is different, and the difference doesn't show up on a benchmark. All the major models are good at math and code now. The difference is something that is difficult to name, but I would call it "judgment." Fable knows what matters when a question is underspecified. It pushes back when I'm asking the wrong thing, and can even make logical leaps to help me get closer to what I'm really after without asking. More than once it's answered the question I should have asked instead of the one I typed. I have a hypothesis about why, and this is a unique moment to test it. Models get better two ways. You scale pretraining: bigger model, more data, more compute. And you do RL: reward the good outputs, punish the bad ones. RL has driven most of the visible gains for ~two years. But RL has a limit built into its structure. It can only improve on metrics that can be measured and rewarded. Math is verifiable. Code is verifiable too--compiles or it doesn't. Judgement is a much more human thing without perfect external validation. We all know that one friend who is brilliant, but has terrible judgement. Or the other friend who is dumb as rocks (said lovingly) but just knows what to do somehow. When labs proxy judgement with human preference scores they often end up with the opposite. We all remember GPT-4o which was King Glazer... well sycophancy is an RL artifact. That's what a model optimized too much to give answers that people like looks like. It looks increasingly like RL gains are spikey and domain-specific. General judgement comes from somewhere else. It comes from a general understanding of the world (a world model), which comes from scale. There's some evidence for this. RL-tuned models beat their base models on the first try, but give the base model enough attempts and it mostly catches up. RL is a valuable tool that sharpens what scale already built. Which gets me to the prediction. GPT-5.6 should compete with Fable on anything you can write a reward function for. Coding benchmarks, agentic tasks, math. It might even win some outright. But it should lag on what you can't reward: judgment, synthesis, knowing what matters in a situation they haven't been directly trained for. Early testers already describe Sol, the 5.6 flagship, as a Rottweiler for hard coding and Fable as a wise owl on architecture. That fits the theory. There's a second tell, and it's the one I trust most. OpenAI is reportedly scrapping Spud and training a much larger base for GPT-6, specifically to answer Fable. They know that RL is not enough. No one has set out to specifically train "judgement" because it's too ephemeral. It's an emergent quality of scaled-up models. This is an extremely important insight, because it means there are likely many other emergent qualities from scaling that we have only scratched the surface of. In the coming years these models may develop taste, research intuition, maybe even something we don't have a word for yet. In other words, the last 2 years have been a bonanza of verifiable gains. But we are re-entering the scaling era in a big way, and there are going to be untold benefits and breakthroughs from this. Contrary to what the skeptics say, I expect this new era to come with massive model improvements that may accelerate rather than decline. The capabilities that are worth the insane capex are the ones we don't have names for yet. When it comes to the capabilities of these models, there's nothing above us but sky.
7
6
5
80
15,477
As usual people are missing the forest for the trees in these debates The actual problem here is that 80% of college students have no business being in college, they are fundamentally unsuited for it and efforts to include them anyway cheapen the institution
Friends don't let friends continue to give exams and papers that are not adapted to the current AI landscape (this is AFTER 27 bailed so the reality is even worse) PS Props to students 1, 22, and 31
43
87
9
1,634
57,594
Saying I enjoyed this essay is probably not quite right. Worth reading.
I wrote about how accumulating capital won't save you from being disempowered by superintelligent AI.
11
116
18,610
🐛 retweeted
someone make a polymarket on whether they think the zcash bug was exploited so we can see who actually believes it vs baiting engagement you will be able to definitely prove whether it was or not with the next network upgrade so oracle for it exists
193
43
16
878
89,528
The way to understand the ZCash bug is it’s not infinite mint of ZEC itself. It’s more like the shielded pool (Orchard) could become insolvent. Think of it more like the KelpDAO hack for ETH. Very little reason not to proactively unshield any ZEC today. Being early to the exit is always better in a bank run. Consider Orchard burned, and don’t re-shield until there’s a new pool with a clean history
🤖 Made with AI
68
148
57
913
201,003
🐛 retweeted
it's unclear if most ppl realize that ozempic’s real effects on culture haven’t even started. cuz what you’re seeing now is the first order effects & some glimpses of second order where ppl get thinner & some products experiencing a resurgence. but the second & third order effects are where things might get gnarly due to the fact that these drugs seem to dampen desire itself across a surprisingly wide range of behaviors (it's not universal or obvious yet). food is simply the first & most obvious target. liek what happens when millions of people suddenly spend less time thinking about consumption? what happens to industries built around cravings, indulgence, impulse purchases, addiction loops, or even certain forms of entertainment? entire sections of the economy assume humans will remain governed by the same reward circuitry we’ve had for thousands of years. if these drugs meaningfully alter those circuits, we’re talking about a tool that edits human motivation. we are gonna see thinness but in a lot of diff ways it seems like.
377
290
89
5,408
1,083,550
Assume the least amount of due diligence from your average holder. So the average holder of Bitcoin is likely to look at the Saylor sell as an extremely bad omen, and start a negative feedback loop.
15
8
3
131
14,478