Jay

strongly urge people to diagram out their logic wherever possible the claim is made (right after saying most economists and AI experts do not expect a permanent underclass scenario), that "if left to its own devices, Silicon Valley may summon a permanent underclass through its own market logic" (a strong assertion) the justification is a mixture of (1) pointing at benchmarks (what?), and (2) indicating that some companies have performed layoffs in anticipation of AI (a tenuous claim -- there are a lot of indications that companies, including Block, use AI as a way to spin layoffs and cost-cutting as an efficiency play rather than a result of unsustainable operating bloat and/or market underperformance) the claim is not justified further, and had one just drawn out the claims first, it would become clear that bad individual business decisions lead to outcompetition by others, not a macro crisis

Went to ClawCon in New York. The vibe felt a lot like crypto events a few years ago: packed with a super young crowd, tons of hype, and merch everywhere. You hear the same takes ten times, but there are a few diamonds in the rough. A lot of the people I spoke with felt like they would have been crypto people two years ago.

If you could transact on a blockchain privately, what features what would be the most important for you, and what would you do?

Multi-proposer blockchains are meant to solve the monopoly on transaction inclusion inherent to single-proposer chains. But multi-proposer designs only solve half of it. Censorship resistance keeps proposers from blocking your transaction; it doesn't keep them from reading it and extracting MEV. Why does this matter? Because in the context of trading, MEV creates inefficient price discovery. If blockchain is fundamentally trading infra, bad prices have follow on effects for the lending, prediction, and derivatives markets that consume those prices. Giga will account for this with Sedna, a protocol that splits each transaction into rateless coded symbols and disseminates small bundles of those symbols across multiple proposer lanes, such that no single lane ever holds enough fragments to reconstruct the payload. The result is a design that eliminates the most common forms of MEV (front running, sandwich attacks, etc.) almost entirely. For trading, that means order flow privacy is a property of the protocol. On single-proposer chains, the solutions for common MEV are application-layer workarounds: private mempools, encrypted order flow auctions, etc. These work to varying degrees, but they ask users to trust an additional party (i.e., trade one monopoly for another) or accept added latency. Sedna moves that protection into consensus, tying privacy to the same data layer Giga's consensus already requires for liveness.

I think the next era of crypto research is going to look like a long steady grind of obvious-in-hindsight improvements, rather than a Manhattan project

Surprised to see so little chatter about Ethereum L2 DeFi TVL recently surpassing Solana I invest in both ecosystems so this def isn't a tribal warfare post. But it does feel like a narrative violation that's gone mostly unnoticed When Solana DeFi TVL overtook Ethereum L2s about 18 months ago, it was widely discussed. The consensus view amongst founders and investors was that it was directional: monolithic scaling was winning, the modular/L2 roadmap was losing, and the gap would only widen But L2s quietly flipped back earlier this month. Combined L2 DeFi TVL (~$6.5B) now exceeds Solana (~$5.6B) after trailing for roughly 18 months The biggest driver of that reversal is Base, which alone accounts for ~$4.3B (surprisingly close to Solana itself). It barely existed two years ago. This is Coinbase flexing that distribution muscle If nothing else, it suggests the scaling debate is less resolved than the discourse assumed. And I suspect this won't be the last time the leader flips either