8 YEARS since my first YT video, and I finally got to the point where I'm talking about my own research 😊 Check out the new video: youtu.be/s8RqGlU5HEs

I happen to be making a video about this right now

TIL that many papers on code generation w/ LLMs hand code post processing functions for individual evaluation tasks... How are papers with 100s of citations getting away with such blatant violation of train/test split separation? It's baffling 🫠

Nothing makes me feel more powerful than jiting my entire training loop in Jax

I'll be presenting my work at the first Reinforcement Learning Conference (RLC) this weekend! If you're around and want to chat, drop by my poster on Saturday or shoot me a message.

TIL part of the YouTube recommendation algorithm uses RL 👀

"Are the RL_Conference keynote speakers any good?" The speakers:

Great keynote by David Silver, arguing that we need to re-focus on RL to get out of the LLM Valley RL_Conference

RL scientists hate him! Find out the one trick they don’t want you to know to improve all your RL models!!!!

Does anyone else also like to write out an explicit plan before they start implementing something complicated?

2 years ago I made a video on what I thought was a great paper on the plasticity of neural networks (youtube.com/watch?v=zEMOX3…). Different versions of that paper were rejected from conferences 4 times. Today it was published in Nature. 🔽 nature.com/articles/s4158…

AI tools for finding research citations are really not there yet, but god I can't wait until they are. Right now all I get are citations to super well known papers.

I'm officially on the job market! I like working on hard problems in AI/ML (esp. reinforcement learning 🤖) that haven't been solved before. If that sounds like what you're looking for, hit me up!