Andrew Clay Shafer 雷启理 (@littleidea) 's Twitter Profile
Andrew Clay Shafer 雷启理

@littleidea

reading, riting and rithmetic
learn, do, do, learn
harbinger of boomshakalaka
neither dev, nor ops, and never the twain shall meet

ID: 14079705

linkhttps://twitter.com calendar_today04-03-2008 20:17:01

51,51K Tweet

13,13K Followers

3,3K Following

Andrew Clay Shafer 雷启理 (@littleidea) 's Twitter Profile Photo

2024, the baseline was chaotic, jumped out of an airplane, ran a marathon, made small progress on many fronts, but not sure I made the best of everything, parenting is hard, humbled, here's to another one

Andrew Clay Shafer 雷启理 (@littleidea) 's Twitter Profile Photo

A periodic task chatGPT has done successfully for months just broke in the most bizarre and inexplicable way, so we have that going for us. Same model, prompt, data… total gibberish output

Lenny Rachitsky (@lennysan) 's Twitter Profile Photo

WTF are evals? Evals are how you measure the quality and effectiveness of your AI system. They act like regression tests or benchmarks, clearly defining what “good” actually looks like for your AI product beyond the kind of simple latency or pass/fail checks you’d usually use

WTF are evals?

Evals are how you measure the quality and effectiveness of your AI system. They act like regression tests or benchmarks, clearly defining what “good” actually looks like for your AI product beyond the kind of simple latency or pass/fail checks you’d usually use
Hiten Shah (@hnshah) 's Twitter Profile Photo

You can’t outsource drive. You can only design for it. I used to think you could teach hard work. That if you said the right thing or set the right goal, people would suddenly start pushing themselves. But I’ve never seen it work that way. Not once. What I have seen is this: