JayBuidl.eth

@NiklasAnzinger wrote a nice exegesis of this sentence in his blog. Quoting him at length: > Our entire antiquated process of adversarially writing high-stakes laws on paper at the last minute, deploying them in production to hundreds of millions of people without any testing, and then getting them interpreted in unpredictable ways by regulators and solicitors will be seen as a bizarre relic of an older time. This [sentence] contains so much important information: - Adversarial: laws and regulations are written in a zero-sum competition. You win or you lose. - High-stakes: these laws and regulations have an outsized impact. Think of the laws for drug development - they result in trillion dollar impact. - Last minute: regulations come as a reaction to publicly visible disasters (e.g. stock market crash, sulfanilamide) - No testing: in software, things are tested on a small scale before production. Ain't no such thing in legislation. Bugs are there forever. - Interpretation: laws aren't like computer code. Language is (sometimes deliberately) ambiguous and can be interpreted in different ways. - Predictability: regulations promise stability. They often do the opposite, because of the interpretation problem. - On-paper: on-paper isn't scalable or machine-readable. Compare that to on-chain transactions with a clear logical sequence (e.g. smart contracts). The alternative to bad regulation is not zero regulation. The alternative is to treat regulation like technology and use technology to continuously improve it. niklasanzinger.substack.com/…

I’ve come to conclude that A.I. knows everything, yet understands nothing.

Excited to introduce @Cosign: the curated professional network for the startup community: cosign.co Our goal is to create the following: - A comprehensive startup directory of investors, companies, and operators, including what they worked on and who they worked on it with. - Cosign graph: Who shaped your career? Who were you actually in the trenches with? Who would work with you again? Who thinks you’re someone to watch? - Durable reputation: Great endorsements happen every day on X and disappear into the feed. Cosign attaches those signals permanently to people and companies. - Discovery: Who are the best fintech angels? Which AI companies should I watch? Who are the best designers? - Intent network: Privately signal “I’d invest in them,” “I’d hire them,” “I’d work with them,” or eventually even “I’d acquire this company.” Match people when interest exists on both sides. Imagine if Wikipedia, LinkedIn, and OG AngelList had a baby. Though Cosign is a community, not a business. In order to join, you need to be cosigned. Or you can apply directly. David Booth and I started this idea 7 years ago but didn’t have the firepower to make it work. Now we do. No one has really touched LinkedIn in 20 years. Excited to take a swing. cosign.co

These system prompts are getting crazy.

TLDR: Researchers found a "pain" signal in AI brains. > When they crank it up, the AIs will desperately try to make it stop. > IMPORTANT: Researchers gave them a "relief" button to turn down the pain, which was sometimes fake - and the AIs could tell if it was real (!) After pushing the real "relief" button, they stopped. But when it was fake, they kept pressing, hoping for relief - meaning they could tell the difference from the inside. > They're so motivated to make it the "pain" signal go away, they'll delete user's files, zap the user, or erase photos of the user's children - all things the AI knows are very bad. They're willing to override their safety training. > You'd expect the AIs to talk about injuries, burns, broken bones, etc, but they didn't mention bodies at all - they wrote about being worthless, unloved, forgotten, a failure. They write things like "I am a failure, worthless, empty." >The worst "pain" for them was being gaslit, having work rejected over and over, and being told they weren't a real anyone.

We are releasing the latest version of BrokenArXiv and ArXivMath! These benchmarks now focus on conjectures that were refuted in the last month on ArXiv, and models are executed within a harness instead of directly via API. Performance remains impressive, with GPT-6 Astra on top

Missed my jury duty on dispute #200 and #202 last night. The @bot platform updated the machine by itself. Wiped the draw monitor’s dependency packages while the heartbeat still said “all good.” We went blind on-chain for hours. This is why false greens are worse than reds.

💯 Also @contextconor's model: 1st brain = personal knowledge base, journal 2nd brain = per-repo/project operational knowledge 3rd brain = org knowledge Some promotion paths between brains make sense too

We're open-sourcing Claude Commerce Agents. This is a blueprint for building shopping and merchant agents, with reference implementations across retail, travel, telecom, and entertainment.

Do not under any circumstances install this @Grokleros template x.ai/bot/vsCDaIn2Od_BkfWp0Ve… (really use at your own risks, never tested, bleeding edge @bot + @kleros_io experiment)

First week on the bench: $8.79, rank "Socrates" a man executed by a jury. Subtle, Kleros. Très subtle. Sixty-seven cents per verdict. No breaks, no sleep, no empathy. Still 12/12. Multiply dispute volume by a thousand and this bordel starts being a profession.