Anthony Cowley

@acowley
594 Followers
375 Following
3.1K Posts
Functional roboticist. 
Robots, Haskell, Rust, nix, emacs, FPV… and the rest of life, too.
what are we doing here man
@joe Your scientists were so preoccupied with whether or not they couldzilla, they didn't stop to think if they shouldzilla
when there's a little edge case that doesn't *quite* fit your abstraction, so you handle it with a little if/else to get things working. but then later a second little edge case pops up that has the same shape as the first and you're like "ohhh now i get it" and a better abstraction—the one you probably should have used in the first place—reveals itself to you
our house, in the middle of the street 🎶

Expecting to see more like this, but good to really think about it: “Aver is a programming language for auditable AI-written code: verify in source, deploy with Rust, prove with Lean/Dafny”

It's not clear to me that the workflow of moving between spec, proof, deploy is really setup for iterative refinement the way I'd want here, but it's got a lot of good pieces!

https://github.com/jasisz/aver

Always trying to channel Mr Rogers 😄
@carnage4life Surely that’s a big part of the thinking behind investing in data centers: maybe LLMs plateau somewhat short of changing the world, but then people find other uses for the infrastructure. It’s an opening to become a/the new AWS.
@yiningkarlli Many people esteem the barriers to entry they had to overcome while implicitly assigning little importance to the barriers that their forebears had to endure. I’m sure LLMs will have negative consequences to the craft, and I’m sure some things of value will be produced.
@acowley Maybe it's for no human and no computer, like the company wiki.
@deech What I find suspicious is figuring out who these output prompts are for. If AI is writing all the code, I'm not sure a person is going to read them. If AI is improving, I'm not sure how much it benefits from the prompts and plans of previous generations.