Louis

If you're wondering whether saturating ARC-AGI-1 or 2 means we have AGI now... I refer you to what I said when we launched ARC-AGI-2 last year (which is also the same thing I said when we announced ARC-AGI-2 was coming, in Spring 2022, before the rise of LLM chatbots)... The ARC-AGI series is not an AGI threshold, it's a compass that points the research community toward the right questions. ARC-AGI-1 is a minimal test of fluid intelligence -- to pass it, you needed to show nonzero fluid intelligence. This required AI to move past the classic deep learning / LLM paradigm of pretraining scaling + static models at inference, toward test-time adaptation. ARC-AGI-2 is the same, but with tasks that probe deeper levels of reasoning complexity (particularly with regard to concept composition). Still, these are tasks that are solvable in minutes by regular people with no external tool use (we hired our test takers off the street), so it does not represent the upper bound of what human fluid intelligence can achieve (say, solving a Millennium problem). ARC-AGI-3 (launching March 2026) probes interactive reasoning: we evaluate how systems explore unknown environments, model them, set their own goals, and plan/execute towards these goals, autonomously, without instructions. We have also started work on ARC-AGI-4 and ARC-AGI-5, which I am pretty excited about!

The tiny Caribbean island of Anguilla now derives almost half of its state budget from the sale of .ai domain names.

Goals with ChatGPT.

An implementation of Contrastive Poincaré Maps with sample notebooks for the biomedical case studies is now publicly available: bit.ly/3I0sM2J

📢More from our recent @NatureHumBehav article from the Technical University of Denmark: Our study shows that behind the apparent complexity of human mobility lies a simple rule shaped by geography and distance. 🔗 dtu.dk/english/news/a… DOI: 10.1038/s41562-025-02282-7

Single-cell data can reveal hierarchical patterns in organismic development but popular embedding approaches often distort them. We introduce Contrastive Poincaré Maps, a self-supervised hyperbolic encoder that preserves hierarchies, scales efficiently, and uncovers developmental lineages across diverse datasets. Led by @nithyabhasker. Preprint here: bit.ly/4211hMY

In this article, Boucherie et al. apply physics-based models to the arrangement of locations to study how geography shapes human movement. They find an underlying pattern in how people choose to move, independent of geographical layout. nature.com/articles/s4156…

Paper 👉 nature.com/articles/s4156… DOI 👉 doi.org/10.1038/s41562…

* Decoupling geographical constraints from human mobility * 🌍🚶 How much of our movement is about human choice and how much is constrained by geography and the spatial layout of locations? Our paper (out in Nature Human Behavior) gives you a practical way to tell:

GenAI isn't just a technology; it's an informational pollutant—a pervasive cognitive smog that touches and corrupts every aspect of the Internet. It's not just a productivity tool; it's a kind of digital acid rain, silently eroding the value of all information. Every image is no longer a glimpse of reality, but a potential vector for synthetic deception. Every article is no longer a unique voice, but a soulless permutation of data, a hollow echo in the digital chamber. This isn't just content creation; it's the flattening of the entire vibrant ecosystem of human expression, transforming a rich tapestry of ideas into a uniform, gray slurry of derivative, algorithmically optimized outputs. This isn't just innovation; it's the systematic contamination of our data streams, a semantic sludge that clogs the channels of genuine communication and cheapens the value of human thought—leaving us to sift through a digital landfill for a single original idea.

Data centers vs. office construction. This chart perfectly illustrates where we are heading.

Next-gen vision pre-trained models shouldn’t be short-sighted. Humans can easily perceive 10K x 10K resolution. But today’s top vision models—like SigLIP and DINOv2—are still pre-trained at merely hundreds by hundreds of pixels, bottlenecking their real-world usage. Today, we scale up vision pre-training to 4K resolution. (1/n)🧵

Having a large research group is prestigious, but do younger scholars benefit? 🧑‍🎓 @YanmengXing, @YifangMa, @robysinatra and co find that mentees from larger groups are more successful, but also more likely to drop out of academia. Now in Nature Hum Beh: nature.com/articles/s4156…

#Fusion🔥| 1ère mondiale pour l'énergie de fusion ! 👀Nouveau record obtenu dans le tokamak WEST, opéré sur le centre CEA de Cadarache : il a maintenu un plasma de fusion chaud de plusieurs dizaines de millions de degrés pendant 22 minutes avec 2,6 gigajoules d'énergie injectée.

o3-mini-high does this challenge 1st first time (no other model has made working shaders in many tries, let alone one) "create a visually interesting shader that can run in twigl-dot-app make it like the ocean in a storm..." "make it even more interesting" Scary good model.

Tesla valuation was recently defended by the announcement of future robotaxis as a new big thing. Waymo has Google based robotaxis since 2009 that in 2024 did 4 million trips in Phoenix, S. Francisco and L.A. ft.com/content/4f129f…

More than 40% of postdoctoral researchers leave academia, according to a study of some 45,500 researchers’ careers go.nature.com/42tQBro

Doubling o1-preview performance on ARC-AGI with one simple trick 🚀 tldr: by providing human-like representations to o1, we are able to substantially increase performance on @arcprize.