Jr Kibs

The Singularity is near...

Announcing ARC-AGI-3 The only unsaturated agentic intelligence benchmark in the world Humans score 100%, AI <1% This human-AI gap demonstrates we do not yet have AGI Most benchmarks test what models already know, ARC-AGI-3 tests how they learn

Your work tools in Claude are now available on mobile. Explore Figma designs, create Canva slides, check Amplitude dashboards, all from your phone. Give it a try: claude.com/download

@CodeByNZ It's over

OpenAI's latest repo has Claude as the third top contributor 😭😂

AI is continuing its march into science and it's about to get nasty youtube.com/watch?v=x3OYM4…

NASA plans to invest $20 billion over the next seven years to develop a base on the surface of the moon, part of its goal to not only send humans back to the lunar surface, but allow them to live there. bloomberg.com/news/articles/…

It’s interesting to see the dynamic that’s emerging right now when using systems like GPT-5.4. The system is so smart that it constantly asks whether it can take control of things. And you know that if it does, everything will move much faster. But you’re afraid it might make serious mistake, like deleting database or wiping files from repositories, so it becomes a constant trade-off between the two. Because you know that if something goes wrong, your boss won’t make OpenAI pay, you will be held responsible. At some point, we’ll have to give up control and let it operate in full autonomy… but for now, the uncertainty is so high that we prefer to remain the bottleneck.

AI will help discover new science, such as cures for diseases, which is perhaps the most important way to increase quality of life long-term. AI will also present new threats to society that we have to address. No company can sufficiently mitigate these on their own; we will need a society-wide response to things like novel bio threats, a massive and fast change to the economy, extremely capable models causing complex emergent effects across society, and more. These are the areas the OpenAI Foundation will initially focus on, and in my opinion are some of the most important ones for us to get right. The Foundation will spend at least $1 billion over the next year. @woj_zaremba, co-founder of OpenAI, will transition to Head of AI Resilience. I believe that shifting how the world thinks about safety to include a Resilience-style approach is critical, and I am extremely grateful to Wojciech for taking on this role. Wojciech has been my cofounder for the last decade; anyone who knows him will understand what I mean when I say he is one of a kind. He has a lot of ideas about how we build a new kind of AI safety. @JacobTref is joining as Head of Life Sciences and Curing Diseases. @annaadeola, our VP of Global Impact, will transition to Head of AI for Civil Society and Philanthropy. @robert_kaiden is joining as Chief Financial Officer. @jeffarnold is joining as Director of Operations.

Anthropic seems to be entering phase that looks like the beginning of hard takeoff.

Little known fact, the Anthropic Labs team (the team I joined Anthropic to be on) shipped: - MCP - Skills - Claude Desktop app - Claude Code It was just a few of us, shipping fast, trying to keep pace with what the model was capable of. Those early Desktop computer use prototypes, back in the Sonnet 3.6 days, felt clunky and slow. But it was easy to squint and imagine all the ways people might use it once it got really good. Fast forward to today. I am so excited to release full computer use in Cowork and Dispatch. Really excited to see what you do with it!

You can now enable Claude to use your computer to complete tasks. It opens your apps, navigates your browser, fills in spreadsheets—anything you'd do sitting at your desk. Research preview in Claude Cowork and Claude Code, macOS only.

GPT-5.4 Pro continues to be the only model of its class. For anything really hard & complex, I throw it into the maw with every bit of context I can think of. More often than not, something very useful comes out. I can't get the same results from Codex or Code or anything else.

JEPA are finally easy to train end-to-end without any tricks! Excited to introduce LeWorldModel: a stable, end-to-end JEPA that learns world models directly from pixels, no heuristics. 15M params, 1 GPU, and full planning <1 second. 📑: le-wm.github.io

This is not the rise of new technology; it is the rise of new civilization!

"Exclusive Self Attention" This paper proposed Exclusive Self-Attention (XSA), which is a tiny two-line change that stops attention from looking at itself. This forces it to focus on the rest of the sequence, and can make transformers more effective! This improves the performance at long context at almost no extra cost.

Install the frontend skill to improve gpt-5.4 for frontend! In the codex app: $skill-installer frontend-skill

@sterlingcrispin 💯

Claude 4.6 is a good programmer but writes insanely severe bugs constantly, it won't catch them all in audits, nor will other claudes You need codex 5.4 auditing every commit 4+ times. If you don't believe me, try it. I have an /auditcodex skill for it github.com/sterlingcrispi…

The future will not be only biological, nor even artificial; it will also have something new that we cannot quite define. Xenobots are proof that universe will be populated by entities beyond our imagination.

The brain–computer interface is the first in the world to be available for wider use outside of clinical trials go.nature.com/47e5G21

@ylecun @_helenqu @cosmo_shirley @rdMorel @albertobietti @mikemccabe210 @PolymathicAI Yes... but after several years of existence, it still doesn’t outperform DINO. When are we going to get practical applications with JEPA ?

@_helenqu @cosmo_shirley @rdMorel @albertobietti @mikemccabe210 @PolymathicAI JEPA finds abstractions that enable prediction. This is how intelligence and science work: find abstract representations of reality that allow to make predictions.

physical systems (orbits/fluid mechanics) may look complex, but are often governed by simple equations/few parameters. can current self-supervised methods learn the underlying physics? our new paper finds that learning in latent space may be the key! arxiv.org/abs/2603.13227🧵

We need to rethink the web for agents.