Check out KRAFTON AI 😁 krafton.ai/en/

This was my response to an email that a student named Loa sent me in 2024. I followed up and exchanged a couple more messages with them. Here's my original post: x.com/JeffDean/statu…

So You Want to Be an Academic? A couple of years into your PhD, but wondering: "Am I doing this right?" Most of the advice is aimed at graduating students. But there's far less for junior folks who are still finding their academic path. My candid takes: anandbhattad.github.io/blogs/jr_grads…

when you give up on this nebulous idea and illusion of prestige, you will finally find peace and freedom. submit to TMLR and JMLR.

Has anyone tried training a simple image generative model, say CIFAR10, purely solely with RL? (finetuning or from scratch) Recent VLMs should serve as a pretty reliable reward model. That could be a more interesting (and more accessible) playground than LLMs + reasoning.

Prediction: In ~3 years academia will be the most desirable place to do fundamental AI research Contributing factors: - small models improve/become significantly more impactful - open weights community broadens its reach - gpus continue to get faster & cheaper - meaningful

Could video models be the path to general visual intelligence? In our new paper, we show that Veo3 has emergent zero-shot capabilities, solving complex tasks across the vision stack. Project page: video-zero-shot.github.io Paper: arxiv.org/abs/2509.20328 🧵👇🏻

This is by far the best & most impactful paper I have seen this year! Congrats!!!

Yet more evidence that a pretty major shift is happening, this time by Scott Aaronson scottaaronson.blog/?p=9183&fbclid…

Very nice work!! If you found this interesting, you may check out Yuchen Zeng's work on LoRA. She mathematically characterized the conditions under which LoRA is as expressive as full finetuning. x.com/Kangwook_Lee/s…

Motivated by kangwook work, Ernest Ryu and Co looked at lora. arxiv.org/abs/2402.11867

This seems very useful, and I'd love to try it out soon!!! Also, thanks for sharing the insights on arithmetic intensity! In fact, it was the very motivation behind our work on batched inference 😁 arxiv.org/pdf/2412.04504 (Cool to see almost identical figures too😉!)