Derek Grudnicki retweeted
Extraordinary outcomes follow extraordinary work, not extraordinary people.
The thing with all these harnesses is that they're great, but you have to get them to mold to your process. So Claude Code vs Cursor, it's not like they're that different, but it's like working with a different person, a different teammate basically. And I feel like switching between them really gives you a perspective on how good the models are individually versus how good the harnesses and the systems around them are.
Derek Grudnicki retweeted
So, to summarize, 900 mathematicians signed a letter basically saying "stop AI from doing math so that our jobs can stick around."
But for anyone who spends more than 5 seconds thinking about what unlocking the equivalent of a billion Fields Medalists would do for humanity, the benefits dramatically outweigh the costs.
Math abundance will help solve chips, medicine, fusion, and numerous other things that will improve the lives of billions, and save millions of lives.
I mean, there was a 0% chance of OpenAI (or any other AI lab) slowing down, so it's a moot point anyways.
But it just seems extremely selfish.
Obvious take once you’ve tried a few rides in both that Waymo’s cooked - imo Robotaxi is already a better experience. But right now it feels more futuristic in the Waymo on an actual ride in SF without the safety driver
I’m too dumb but can’t we just “make” more of the lines? Like they’re there we just need to make more in a structure right?
































