Tweets

2024-08-19T03:35:58+00:00 | 🔗

It's still crazy how big the gap between 1st and 2nd can be in the olympics

2024-08-18T04:20:45+00:00 | 🔗

RT @sporadicalia: wait wait wait … is he gonna be right?? https://t.co/yTLj9rLfXg

2024-08-17T19:57:39+00:00 | 🔗

I mean comeon, a little too on the nose is it? https://t.co/5Xr5kEhzzT https://t.co/JZ4biNI7nc

2024-08-17T00:08:56+00:00 | 🔗

Big surprise, people can't train (with general blunt methods) AI smarter than themselves https://t.co/IBfwhUUuIS

2024-08-16T17:59:51+00:00 | 🔗

it's very unfortunate, they should start doing this again https://t.co/XF0YWzY7XT

2024-08-15T04:34:26+00:00 | 🔗

RT @Acyn: Jessica: Donald Trump has said he has a big beautiful health care plan and there are no specifics on it. And Kamala is being held…

2024-08-14T23:15:33+00:00 | 🔗

github down???

2024-08-14T20:52:52+00:00 | 🔗

@shorez Thanks for the hint! Wordle 1,152 1/6 🟩🟩🟩🟩🟩

2024-08-14T15:17:02+00:00 | 🔗

Thoughts on lmsys. Say there is a large cohort of human annotators giving feedback for RLHF and generating data. You train a model on that. Then you direct the same cohort of people onto lmsys. Your model will artificially perform better because it has the prefs of those ppl

2024-08-14T14:50:50+00:00 | 🔗

This is from the Tractatus, notably Wittgenstein would discard many of these ideas and disagree with them in his later work. Later on he would put forward how language depends on practice and custom, which notably language models do not have and can only be slowly accumulated https://t.co/2ZadYiPdos