After six years riding the AI agent wave, I've decided its time to take the leap and build something new.
Mashing together some ideas from my time spent across crypto, AI and consumer data rights. Going to post more about what I'm building and what I'm noticing. Watch this space
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
Arthur Hayes reveals how one dinner with @naval at Token 2049 had him buying millions of zcash:native:
"He started talking to me about how the ring signatures of Monero are not as strong as you think. They'd be the events successfully deanonymized by Japanese law enforcement."
"Right then, I bought a few million dollars worth of Z cash. And what I noticed was only two brokers would make me a price in Z cash, because, you know, we don't do privacy coins. If I can't get it, then I want more of it."
@CrlynBuild Described the concept to claude showed it a few examples I liked including my own hand-drawn rough sketches, a few examples I didn't like with reasons why.
Once it got closer, I asked it to take what its learnt about my preferences into a prompt for Nano-banana etc. And iterate.
@Jia_builder I remember the Paul Graham had a post about always getting the .com and just getting more creative with word pairs. Didn't stop me going with .app though
https://t.co/xGqCn8CjAZ
@paulsavage It's not absolutely essential to have a .com domain. But on the other hand there's also no excuse for not having one. When founders are attached to a name with a marginal extension, the reasons are usually just status quo bias and lack of imagination.
@RichardSSutton Simulation is a great candidate. A lot of services work with fancy names like scenario planning, war-gaming or digital twin should be computation rather than curation
@fin465@garrytan Sounds intuitive but it shocking how much harder it gets once you have separate product marketing, demand gen, field marketing, sales and BD teams
Predictably, the challenge at scale is mainly a humans getting in the way of other humans