Think Strategy 🎐 @ThinkStrategy3

@paulg Constraints and challenge to value system is lethal combination for young minds.

"An Introduction to Flow Matching and Diffusion Models" is a set of MIT lecture notes for the course "Generative AI With Stochastic Differential Equations" (2026) that provides a clear introduction to the mathematics behind modern generative AI. The notes discuss flow matching and denoising diffusion models as core techniques behind many advanced generative systems, with references to models such as Stable Diffusion 3, FLUX, VEO-3, and AlphaFold3. They develop the mathematical foundations of generative modelling, covering topics such as sampling from probability distributions, ordinary and stochastic differential equations, Brownian motion, diffusion processes, flow matching, score matching, classifier-free guidance, architectures for image and video generation, latent spaces, autoencoders, and discrete diffusion models for language generation. What I particularly appreciated is the teaching style. The notes first build geometric and probabilistic intuition and only then derive the complete mathematical formulations. The result is a treatment that is rigorous, visual, and remarkably approachable. This is probably one of the best freely available resources for understanding what is actually happening under the hood of diffusion models from a mathematical perspective. diffusion.csail.mit.edu/2026/docs/lect…

"Transformers" by Daniel Jurafsky and James H. Martin is one of the clearest and most mathematically grounded introductions to the Transformer architecture I have ever read. Chapter 8 introduces the Transformer as the standard architecture behind modern large language models. What makes this chapter particularly interesting is its step-by-step presentation of the underlying mechanisms: contextual embeddings, self-attention, query, key and value vectors, scaled dot-product attention, multi-head attention, residual streams, feedforward layers, layer normalization, masking, and the parallel matrix formulation of attention. In particular, the treatment of attention as a weighted sum of contextual representations is especially valuable. The chapter first develops an intuitive, simplified view of attention and then gradually derives the full formulation using the Q, K, and V matrices. This approach makes it easier to understand what is actually happening inside the architecture from an algebraic and matrix-based perspective, rather than simply viewing the usual block diagrams. I think it is an excellent resource for anyone interested in understanding how Transformers work from linguistic, mathematical, and computational perspectives. web.stanford.edu/~jurafsky/slp3…

Good luck with New chapter @NoamShazeer . Reasoning will get better now @sama @JeffDean

Noam Shazeer @NoamShazeer

2 weeks ago

I’m excited to share that I’ll be joining OpenAI and look forward to working with the exceptional team there. It was a difficult decision to move on. I’m incredibly proud of the amazing team at Google and everything we’ve built together. It has been an honor and a pleasure to

985 869 16K 9.5M 2K

@elonmusk @JeffBezos @zerohedge Ad astra per aspera

I always found it puzzling how language models learn so much from next-token prediction, while video models learn so little from next frame prediction. Maybe it's because LLMs are actually brain scanners in disguise. Idle musings in my new blog post: sergeylevine.substack.com/p/language-mod…

Good luck with New chapter

Andrej Karpathy @karpathy

2 months ago

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

8K 11K 150K 27.7M 14K

@antoniolupetti superficial abstractions

🎉 The Agents in the Wild: Safety, Security, and Beyond workshop @ICLR2026 is less than a week away! Join us April 26 in Room 204 A/B, Riocentro, Rio de Janeiro! 🌴 Safety and security for AI agents — both foundational and emerging challenges — demand serious attention. Researchers and practitioners are mobilizing: ▪️ 151 papers accepted ▪️ 161 reviewers (58% industry, 42% academia) ▪️ Up to 800 participants expected ▪️ Incredible engagement on a topic that clearly matters. The schedule: 👇

saying the unsaid

Sam Altman @sama

2 months ago

feels like a good time to seriously rethink how operating systems and user interfaces are designed (also the internet; there should be a protocol that is equally usable by people and agents)

2K 767 12K 1.6M 2K

How @elonmusk fixed Starlink: “Starlink was a mess. It was 10X too expensive and they were building 1/10 of how many they needed. Elon’s like I've had it. This is now the bottleneck. I'm fixing this. He grabs a team of engineers that he trusts and they fly up to Seattle. They fire the entire Starlink leadership team. They sit down in a war room and they start running the algorithm. •What is the first principles of satellite design? •How simple can we make this thing? •Why does this exist? •Why are these two things so far apart? •Why do we need this much energy? •Why do we need this manufacturing process? And over the course of a few months they make a two order of magnitude leap. These people had never encountered this design before, but just by applying the algorithm and working with maniacal urgency towards this extremely high design bar, they created this product that's now —if it was a standalone business —would be worth tens of billions of dollars [or more].”

David Senra @davidsenra

3 months ago

My conversation with @EricJorgenson, author of The Book of Elon (@elonmusk). 0:00 Book Reveal 0:39 Build Useful Things 2:19 Engineering Talent Edge 4:26 Wired for War 6:47 Tip of the Spear 8:47 Burn the Boats 13:13 Facing Fear 15:16 Origin Story Myths 18:19 Know Business A to Z

106 468 3K 5.2M 4K

@rukidut Thank you. something good will come out of it

We shape our codebases and afterward our codebases shape us

@PyTorch This is how it started

@YouTube What's up

Boris created Claude Code. His point here is important - when AI handles the code generation, the engineer's value shifts to the decisions above the code: 1. what do we build? 2. why? for whom? 3. and how it all fits together. The bottleneck was always judgment, taste, and systems thinking. AI just made that more obvious.

Boris Cherny @bcherny

5 months ago

@big_duca Someone has to prompt the Claudes, talk to customers, coordinate with other teams, decide what to build next. Engineering is changing and great engineers are more important than ever.

244 651 9K 1.2M 1K

Your taste is your evals.

Garry Tan @garrytan

5 months ago

AI just moves the excuses out of the way. You don’t get to say “I can’t edit” “I don’t know how to code” or “I’m not technical” Intelligence is officially too cheap to meter. Your taste is your evals. Your desire to win is your agency. Time to build. garryslist.org/posts/ai-didn-…

95 103 865 39K 418

@asharoraa Define "delete" what happened? Phone crashed? New phone?

@venturetwins AI : Hold my beer.

@alc2022 Service Provider Product/Enterprise