Large Language Model Agents is the next frontier. Really excited to announce our Berkeley course on LLM Agents, also available for anyone to join as a MOOC, starting Sep 9 (Mon) 3pm PT! 📢 Sign up & join us: llmagents-learning.org

QKV transformers are recurrent networks. With extra steps, compute, and memories.

We miss Dennis Ritchie ...

Technical Report for Qwen-2.5-Coder is here 👇

Transformer Experts! If I were to intentionally corrupt the needle-in-the-haystack token, in the corresponding KV cache position on all layers After doing the forward pass to generate it: will a QKV transformer be able to "pay attention" and generate the token. If yes/no, why?

Looks like, we will build time travel in the future. So we can apply for jobs, with experience 🤣

So well said. Even sometimes I think about failure manifestations than success scenarios. Every Singaporean should listen to this once.

Some folks wondered why we went through the trouble of putting RWKV under the Linux Foundation.... It's so that this would be impossible to pull off We saw the board drama, and we wanted something simple, that works reuters.com/technology/art…

A decision on SB-1047 is due soon. Governor Gavin Newsom has said he's concerned about its "chilling effect, particularly in the open source community". He's right, and I hope he will veto this. If you agree, please like/retweet this to show your support for VETOing SB-1047!

MKBHD should be allowed to be critical in his review while remaining impartial External Reviewers are allowed to be critical on MKBHD products while remaining impartial Not ok: attacking for the sake of revenge, Or attention