Sublime

An inspiration engine for ideas

AllPeopleCollectionsArticlesAudioBooksFilesHighlightsImagesLinksNotesTextTweetsVideosSocial

Unpacking decentralized training

George Hotz explains his thoughts on working at OpenAI vs. META: > why would any researcher want to work at openai? > paper clip maxxing > can’t share your research > you’re on the BAD team > meanwhile who would’ve thought.. > zucc is the GOOD... See more

NIK x.com

This is, by the way, the dirty secret of the machine learning movement: almost everything produced by ML could have been produced, more cheaply, using a very dumb heuristic you coded up by hand, because mostly the ML is trained by feeding it examples of what humans did while following a very dumb heuristic. There's no magic here. If you use ML to... See more

Forget Privacy: You're Terrible at Targeting Anyway

Trends – Artificial Intelligence (AI) – May 2025 – BOND

The document analyzes rapid growth and transformative trends in artificial intelligence, highlighting unprecedented user adoption, technological advances, global competition, enterprise AI integration, and associated benefits and risks shaping the future.

bondcap.com

Gwern - The Anonymous Writer Who Predicted The Path to AI

youtube.com

I'm not anti-AI. I'm pro-Creators.

Andy Allen x.com

Zuck poached 4 elite Chinese researchers from OpenAI: - Jiahui Yu: led o3, o4-mini, GPT-4.1 - Hongyu Ren: creator of o3-mini & o1-mini, core contributor to o1 - Shuchao Bi: head of post-training multimodal org - Shengjia Zhao: key contributor to GPT-4 &... See more

Yuchen Jin x.com

DeepSeek-R1 Uncensored, QwQ-32B Puts Reasoning in Smaller Model, and more...

Andrew deeplearning.ai

Update: A couple of us Google Brain / DeepMind researchers have left to work on something new. Some observations: * Off-the shelf AI libraries (e.g. RAG and agents) do not work well enough for high-quality products. * In-house research is essential to get the... See more

Peter J. Liu x.com