@maximelabonne Sounds amazing!
Btw, I was meaning to ask you and always forget:
Any good book or guide on how to train LLMs from scratch? Like the full lifecycle, with some good level of detailing. (Assuming compute is granted)
@MrBeast@nikitabier, this is the proof that your algo prioritizes what keeps people hooked and active, not some kind of “useful” content.
Why the hell would someone want this in my feed. Engagement farming shouldn’t be prioritised this much by your algorithm.
Fix that bro
@AnthropicAI So now every single MD first lines will be “take it easy, chill out and don’t stress or panic about anything. You’ve got this I trust you”.
Search the importance of the “feeling” neurons and deactivate them if they don’t hurt benchmarks.
@HKydlicek@HKydlicek found no reference to FinePDFs-Edu in there
Cite: "data blend consisting of approximately 70% post-training data and 30% pretraining data from the parent Nano v2 recipe"
But https://t.co/hAkzp4zeM3 (paper from Sep 2025) recipe doesn't mention FinePDFs-Edu either
@Kimi_Moonshot@cursor_ai@Yulun_Du, were you not complaining about @cursor_ai not even approaching to you guys on this?
Or they already reached out after they got exposed and offered a large sum?
Don’t sell your soul, man