After Robinhood raised the gas limit and lowered fees for users on Robinhood Chain, chain revenues have declined significantly (green), while DEX volumes have remained strong (purple). It's solidly #2# behind Solana. Lots of folks pointing at this like it's a mistake. But it's clearly intentional. And if you actually believe in the growth of RWAs and on-chain finance, this is 100% the right answer. Robinhood is thinking a lot bigger than optimizing for short-term sequencing fees. Owning the ecosystem and bringing tokenized assets on-chain and increasing the GDP of the blockchain is the much bigger long-term bet, and that will only happen if capacity scales and fees stay low. They will make it up on volume. Bullish.

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here:

Many fields spend enormous energy on prestige allocation. Judging prize winners, refereeing papers, organizing seminars, deciding who gets tenure. Mathematicians are not highly paid; their compensation is largely in prestige and in the intellectual enjoyment of the work. Prestige allocation is very important. I am not demeaning this. Poor prestige allocation leads to warped incentives and a field of work getting benchmaxxed away from its underlying social goal. (China's cash-for-publications policy, which was ended in 2020, was a classic case of this.) It is appropriate for a field to spend of energy getting prestige allocation right, the same way it's important for a company to spend extra time and energy to make sure that capable employees are recognized and rewarded. So why are people reacting badly to the mathematics letter? Put aside the Navier-Stokes drama, we all agree that was scandalous. That didn't need a letter. This letter is about much more than that: AI Labs, you are breaking our prestige allocation structures. Who do we give prizes to? Who should get tenure? Who will give the seminars? How do we continue to motivate young mathematicians? We carefully calibrated these systems, and you are breaking them in the search for glory in a different prestige system than our own. Show some respect for the way we do things. The reason why tech people are reacting badly to this, I suspect, is because programmers—the backbone of the tech industry—just went through this. The programming prestige allocation has already been toppled. 10x programmers got displaced by 20-year-old tokenmaxxers and agent orchestrators. If you spent your career memorizing the minutiae of SQL query planners, that is now an encyclopedic party trick. Everything is changing, and the answer has been a resounding: get the fuck over it. Nobody else cares. Programming serves human goals, not programmer prestige hierarchies. It is painful. But I think this is the right answer. Programmers will survive, and so will mathematicians. Mathematics does not exist for mathematicians. It exists for humanity. If mathematics needs to adapt to AI, and if that adaptation will be painful and chaotic, if it will force the field and its social structures to rapidly reassemble, then the sooner the better. It is not technology's job to genuflect to our existing social structures. It has always and forever been the job of social structures to adapt to technology. Any answer to the contrary—to the printing press, to industrialization, to television, to the internet, and now to AI—correctly resides in the dustbins of history.

They don't really believe this, they're just saying this to get more donations They don't really believe this, they're just saying this to raise their next round from VCs They don't really believe this, they're just saying this to market their IPO They don't really believe this, they're just saying this for their next quarterly earnings call They don't really believe this, they're just saying this to get more people to tune into the next United Labs of America state of the union

We all know intelligence is spikey/jagged - but do we know _why_ it is? TL;DR is everyone is cheating in AI (and because we're releasing 🤫 soon, we're making precommitments to do better 💪)

So basically P(doom) is plummeting right before our eyes due to an X-risk fear social contagion, caused by a Frances Haugen-style corporate whistleblower microdrama in the weeks before a blockbuster IPO. I can safely say this one was not foretold in 2015-era LessWrong posts.

E185: @variational_io - Zero Fees, More Assets, No Funding Rate @variational_lvs co-founded a hedge fund at 20 that Digital Currency Group acquired a year later, then walked away from that exit to build Variational that now sits second only to Hyperliquid in open interest and daily volume. His argument: crypto rebuilt exchanges before it built brokers, and that architectural gap is what's holding on-chain trading back. 0:00 Intro 2:20 Merch, Travel & Being Back in Singapore 5:13 Who Is Lucas Schuermann? 5:24 How Lucas Thinks About Building Trust 6:03 Starting College at Age 12 8:26 How Do You Actually Skip That Many Grades? 9:45 Making Friends Six Years Younger Than Everyone Else 10:40 Are You Still a "Weirdo" at 30? 11:40 Starting a Hedge Fund at 20 - "Hubris of the Young" 14:09 From Paper-Reading Group to a Real Fund 16:50 Acquired by Digital Currency Group at 21 18:59 Why Sell the Fund to DCG? 19:54 Sponsors: Variational & @Bitwise 20:47 Did You Regret Selling That Early? 21:51 The Money DCG Made Off Their Work 22:38 Starting a Market Making Firm 24:46 Explain Variational Protocol to Your Mom 26:49 Biggest Mistake Building Variational 28:13 The "Holy Sh*t, We Built Something Massive" Moment 29:54 Reason #1#: Zero Fees 32:59 How Do You Trust the Price With Only One Market Maker? 34:33 Are Trading Fees a Scam? 35:29 How Much Do Fees Actually Cost Traders? 36:33 Reason #2#: Asset Selection & What "Liquidity" Really Means 40:20 Why Hasn't Crypto Solved This Already? 42:39 Where On-Chain Order Books Break at Scale 46:04 Sponsor: @KASTxyz 46:52 Why Wouldn't a Hyperliquid Trader Switch to Variational? 49:28 Who Gets Destroyed When Wall Street Liquidity Arrives On-Chain 51:18 Reason #3#: Swaps 52:34 The Real Problem With Perps: Funding Rates 55:18 Roger's BitMEX Story: "Holy Sht, I Can Print More Bitcoin" 58:44 The Hard Lesson: Wiped Out by Funding Fees 1:00:38 Why Perps Are Easy but Options Aren't 1:01:13 What Is a Swap, Really? 1:03:20 How Swaps Beat Perps 1:05:28 Why Raise $50M in May? 1:07:30 Sponsor: @JupiterExchange & @ethena 1:08:14 Does VC Funding Help or Hurt a Project? 1:09:46 Variational vs. Hyperliquid: The Architecture Difference 1:11:44 Why Didn't Hyperliquid Build for Hundreds of Assets First? 1:13:14 Staying Relevant After the Points Program Ends 1:15:11 The "AWS of Liquidity" Analogy, Revisited 1:16:47 What Hyperliquid Gets Wrong About the Future 1:17:48 What Variational Could Get Wrong 1:19:06 Who Wins Long Term: Binance, Coinbase, Hyperliquid, or Variational? 1:21:39 Bitcoin Price Outlook & Crypto Adoption in 2026-27 1:23:59 Variational's Endgame 1:25:51 The One Thing to Remember 1:27:18 Closing Thoughts

Last week the boys got dunked on, so this week they brought on the FOMO co-founder. @paulerlanger says a fully transparent social graph is why FOMO became Robinhood Chain's biggest app. Timestamps 00:00 Intro 01:06 Paul vs. the uncs 03:37 From dYdX to FOMO 05:14 Tarun's weekend in the trenches 08:40 Transparency vs. copy trading 12:27 FOMO's distribution power on Robinhood Chain 17:52 Where memecoin trading destroys value 21:10 Churn, theses, & traders as the next celebrities 27:08 Shaming, sidewallets, & clans 32:35 Hunter Biden's $LAPTOP token 37:15 Stock coins, bond ETF fees, & how long the meta lasts 42:52 AMC's CEO vs Robinhood's tokenized stock 47:51 One-to-one backed equity & earnings calls as TV shows 52:20 OpenAI, Anthropic, & the Navier-Stokes drama 58:49 Is math research over? 🔥Stay updated with all the latest hot takes by following and subscribing to @_ChoppingBlock and @unchained_pod! 🎥 YouTube: 🎧 Spotify: 🍎 Apple: 🎙 Podcast Home:

Paul sees stock-paired memecoins as community. Haseeb sees what happens when the cycle turns. Paul: "Memecoins with stock tokens, I think are great community building aspects for these companies." "You saw what happened with GameStop. So now if you have this community around this memecoin, finally, people are actually caring about your company." "There's public company CEOs that have joined FOMO in the past week and now they're following the coins on FOMO." Haseeb: "Right now we're in part one of the cycle. Part one of the cycle is the fun part. That's the part where numbers go up." "When the cycle ends is when everybody's down bad. Everybody's upset. Everybody starts talking to, what's that, Burwick Law or whatever. That's when the knives come out." "People will absolutely do that for just a $10 million memecoin rug pull." Paul: "The difference is they're not launching the token and rugging it. It's like the token's already created."

Haseeb says earnings calls are bifurcating, a document dump for the agents and entertainment for everyone else "I like the take of stockholder communication bifurcating. At the moment of your earnings release, there's a document dump for the agents, for all the trading firms and hedge funds and anybody doing really sophisticated trading. You ingest that all immediately in some PDFs." "And then for the humans it's like a TV show. You get on your Twitch stream and it's entertaining and it's colorful." "We're moving to a world where the old school literal conference call is going away, because who is that for? Professional traders are not going to be using that anymore. So it's either financial entertainment for people who are vibe trading, like, I think it's time to buy, I like the CEO's confidence. And then you've got the agents making all the actual high frequency, large money financial decisions." "There's nothing in the middle anymore for these very serious conference calls. They'll basically become an artifact of the past pretty soon."

Last episode we got called uncs. Then Tarun spent the weekend trenching and came back converted. (I remain unrepentant.) We needed a true believer to explain what was going on, so FOMO CEO @paulerlanger joined us to make the case for social trading reviving the trenches ⤵️

age of research for post-training ended a few months ago: "the marginal return of engineering the data and environment pipeline substantially exceeds that of algorithmic novelty in post-training"

Excellent primer on the economics of token caching, and why fancy token model routers aren't as economical as you'd think. As workloads lean more agentic (way more tokens in than out vs chat), caching dominates as the most important input into token costs.