last frontier labs batch for now bio and world models coming next - @_inception_ai (@adityagrover_, @StefanoErmon) - @ElorianAI (@AndrewDai, @yinfeiy) - @si_pbc (@_neelr_, @devanshpandey) - @adaption_ai (@sarahookr) - Sooth Labs (@rsalakhu) x.com/i/lists/20617784890369…

NitroGen just won CVPR Best Paper Honorable Mention!! We are making strides towards general-purpose embodied agents that master not only the real world physics, but also all possible physics across a multiverse of simulations. It’s been 4 years since MineDojo, our first embodied agent in Minecraft, won NeurIPS Best Paper. Congrats to everyone on the team!!

Imagine trying to type an email by blowing into a straw. For 8 million people in the US & Canada with upper limb impairment, this is their reality. @axionlabsinc has built a solution that restores dignity and independence with Click, a non-invasive BCI. Watch my full chat with @Reddy_XX and @TinTunjic at @southpkcommons below.

last frontier labs batch for now bio and world models coming next - @_inception_ai (@adityagrover_, @StefanoErmon) - @ElorianAI (@AndrewDai, @yinfeiy) - @si_pbc (@_neelr_, @devanshpandey) - @adaption_ai (@sarahookr) - Sooth Labs (@rsalakhu) x.com/i/lists/20617784890369…

I've been posting more actively lately and wanted to properly introduce myself for people who don't know me personally: • At 16 I finished school, and at 20 got my BSc in ML • At 21 moved to the Bay Area: I sent 200 applications after graduating, and got exactly one offer there 😅 • At 22 I came across @80000Hours and decided I needed to work on AI Policy • At 23 I got into the AI Governance Masters at @TU_Muenchen (top-20 CS unis in the world) • At 25 I volunteered on responsible AI reports with @AllTechIsHuman and AI Standards Lab • At 26 I joined @farairesearch, a leading technical AI Safety research org Now back to fundamental ML research and working on something new. More soon 😉

day 2/n The Platonic Representation Hypothesis by @minyoung_huh, @thisismyhat, @ssnl_tz, and @phillip_isola this one keeps generating research questions the more you sit with it the central claim is that any model trained on enough data essentially approximates some shared platonic space of meanings representations from different models (not necessarily from the one domain so could be LLMs, vision models, whatever) converge toward this space and preserve the same relational structure between concepts so training a generative model is really just learning increasingly precise approximations of this underlying space (and different architectures produce same destination) very philosophical, very worth reading!

been building this list for a week now to follow the people behind the labs (as these are two very different feeds) this batch is particularly interesting because almost all of them are betting on something architecturally non-standard: - @SakanaAILabs (@hardmaru) - @liquidai (@helloiamleonie, @maximelabonne, @xanamini) - @flappyairplanes (@aidanmantine) - @magicailabs (@EricSteinb) - @ndea (@mikeknoop, @fchollet) - @_doubleAI_ (@YoavLevine) full list: x.com/i/lists/20617784890369…

my dream angel investor: - human (or fund, no ego) - check size: $25,000 to $1,000,000 - says yes and actually wires - signs the safe without redlining it - doesn't want a board seat - makes the intro before i ask - doesn't ghost after "keep me posted" - reads my monthly updates - writes the check this quarter, not "next fund" - already a paying aesty subscriber (annual, obviously!)

been building this list for a week now to follow the people behind the labs (as these are two very different feeds) this batch is particularly interesting because almost all of them are betting on something architecturally non-standard: - @SakanaAILabs (@hardmaru) - @liquidai (@helloiamleonie, @maximelabonne, @xanamini) - @flappyairplanes (@aidanmantine) - @magicailabs (@EricSteinb) - @ndea (@mikeknoop, @fchollet) - @_doubleAI_ (@YoavLevine) full list: x.com/i/lists/20617784890369…

continuing with robotics, a field close to my heart 🤖 - @SkildAI (@pathak2206) - @physical_int (@hausman_k) - @RhodaAI (@startupjag, @GordonWetzstein)

Hands on AI Engineering! I open-sourced a collection of 50 hands-on AI engineering tutorials. It features step-by-step projects and tutorials on: • AI Agents and Multi-agents • RAG (Agentic, Vision, and Local) • MCP AI Agents • OCR Apps • Voice AI Agents • & so much more 100% free and open source. 1k Github stars I've shared the link in the comments!