back to main landing Tweets ⌂

Tweets

2024-08-27T20:23:29+00:00 | 🔗

Why do we think that the jump in intelligence is from like normal high school/college students to Nobel Prize winners???? Clearly, it is from rocks to life or from monkeys to humans.

2024-08-26T20:18:16+00:00 | 🔗

Prompt-caching is an llm waiting to be fine-tuned

2024-08-26T17:43:56+00:00 | 🔗

If there's one thing I learned, it's that transfer learning sounds a lot better in principle than it is in practice. It is very difficult in practice and is much more about precision and curation. I do not think throwing reams of video data at transformers makes them better. https://t.co/B9J6318bHQ

2024-08-24T13:05:43+00:00 | 🔗

yup! https://t.co/3GgrGL1NgV

2024-08-24T03:48:27+00:00 | 🔗

“If you want to fool magicians, you’re going to fool them with some hundred-year-old mathematical principle.” I feel the same way about revolutionizing AI, although I would say "decades-old economic principle" instead

2024-08-24T03:32:36+00:00 | 🔗

Sentiments like this are so weird, the admin class will only get bigger..... irreversible trend https://t.co/Kmmz2jyoZJ

2024-08-24T03:00:47+00:00 | 🔗

I purposefully avoid using stuff like Cursor or Devin not because they aren’t magical or powerful (they are), but because the type of work I want to do and where I know my strengths are… should not require tools like that.

2024-08-23T18:55:36+00:00 | 🔗

Love this P(DOOM) framework!!! https://t.co/93SGUeORin https://t.co/8KQBHZvVMR

Tweet image for 1827057115313172985

2024-08-23T14:54:21+00:00 | 🔗

@j0hnparkhill Lots of education…

2024-08-23T14:25:27+00:00 | 🔗

The AI/ML efforts I'm most optimistic about are the ones that aim to collect more DATA, not more compute and better architectures. https://t.co/4uM6CV4fao

Where else should I get information from? contact me somehow if I'm missing something, I really appreciate new interesting sources of information.