benchmarkmaxxing in AI is unexpectedly worse than benchmarks in crypto. all the best devs i know still need to vibe-eval i.e. daily drive something for a few days to get a sense of how good or bad it is. all the slop i see on twitter is just hurting the dev community.

silicon valley literally had everything

🚨 NEW DATA: DELAWARE HAS LOST A NET 28 PUBLIC COMPANIES THROUGH REINCORPORATIONS IN 2026, ACCORDING TO NEW RESEARCH FROM A CORPORATE LAW PROFESSOR. University of Nevada, Las Vegas law professor Benjamin Edwards published updated findings Friday tracking 71 identified corporate-jurisdiction moves in 2026, up from 59 in his September update. Texas and Nevada are competing closely to attract companies changing their legal homes. September alone saw nine identified moves, with activity continuing well beyond the traditional shareholder voting season. The research identifies major companies including DoorDash, Dell Technologies, Datadog and Samsara among those pursuing departures from Delaware. The findings also provide important context: Delaware has attracted 15 attempted inbound reincorporations and remains strong in new public-company formations, gaining 73 companies through IPOs. The 71 tracked moves include proposed transactions, not exclusively completed conversions. But the direction of reincorporation activity is notable. Delaware is experiencing a net outflow of existing public companies even while it continues to attract new listings. Research published October 16: businesslawprofessors.com/20…

JUST IN: Nathan Fielder's Elizabeth Holmes documentary, "You Can See Everything," debuts with a 99% Rotten Tomatoes critics score & audience score — one of the highest rated films of all time.

Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper for my agent harness (pi): Each conversation gets assigned a separate Tor circuit, your requests are masked with random cover traffic, zkAPI ensures you don't leak your identity through payments, and all prompts are locally rewritten (bring-your-own-local-model; I'm using qwen3.8-flash-next) github.com/vbuterin/zkapi/tr… My agent harness is wrapped with a bwrap sandbox (also experimenting with MicroVM) and wrapped again with torsocks, so everything it does goes through Tor. Here's what me asking it to do a simple research task looks like. It's getting quite usable! Though you definitely still pay quite a bit in increased latency, especially when starting a new conversation. Also, Tor is far from optimal here for privacy; the cover traffic mechanism is okay against passive adversaries but much weaker if the LLM provider is colluding with the guard and actively manipulating timing. This kind of task really requires a more specialized mixnet, probably take the tradeoff of a higher bandwidth multiplier and built-in constant cover traffic. Also also, much cheaper zkAPI deposit and withdraw gas costs are coming to mainnet soon.

We processed an $8m single-day redemption out of nOPAL today. Liquidity working as intended. That’s 6.7% of TVL. DeFi will have inflows/outflows, and while RWA redemptions aren’t instant, well structured assets will deliver every time. Set realistic expectations and meet them.

Do a kickflip!

Holy shit y’all are not ready for tonight’s Tesla Roadster demonstration. There is a massive ramp 😂