N

Our neurons were shaped by biological mechanisms, Mother Nature, and survival; our models are shaped by similar data. LLMs, however, don’t share our optical limits, sensory biases, or need for intuition. I believe that there is a new kind of reasoning beyond human-interpretable space — alien, potentially ingenious, and discoverable only through machine cognition. That’s where I want AI research to go next.

Google dropped the TurboQuant paper yesterday morning. 36 hours later it's running in llama.cpp on Apple Silicon, faster than the baseline it replaces. the numbers: - 4.6x KV cache compression - 102% of q8_0 speed (yes, faster, smaller cache = less memory bandwidth) - PPL within 1.3% of baseline (verified, not vibes) the optimization journey: 739 > starting point (fp32 rotation) 1074 > fp16 WHT 1411 > half4 vectorized butterfly 2095 > graph-side rotation (the big one) 2747 > block-32 + graph WHT. faster than q8_0. 3.72x speedup in one day. from a paper I read at dinner last night. what I learned along the way: - the paper's QJL residual stage is unnecessary. multiple implementations confirmed this independently - Metal silently falls back to CPU if you mess up shader includes. cost me hours - "coherent text" output means nothing. I shipped PPL 165 thinking it worked. always run perplexity - ggml stores column-major. C arrays are row-major. this will ruin your afternoon everything is open source. the code, the benchmarks, the speed investigation logs, the debugging pain, all of it. github.com/TheTom/turboqu… paper to parity in 36 hours. what a time to be alive.

New from the Anthropic Economic Index: how people’s use of Claude changes with experience. Longer-term users are more likely to iterate carefully with Claude, and less likely to hand it full autonomy. They attempt higher-value tasks, and receive more successful responses.

This is the kind of work happening at Gauntlet AI. At our recent From the Forge event, @Austen and Gauntlet engineers walked through an autonomous system that runs continuously, joins conversations, tracks performance, and ships real products. This is the kind of system we build at Gauntlet. Cohort 5 starts next month. Learn more: gauntletai.com/apply

I'm playing around with building an agent orchestration framework to solve difficult math problems. Folks online, what are some general math theory, proof writing, research guidelines, etc. that I can use as guiding lights for this framework?

tonight at the @gauntletai office we hosted a sxsw event feat. @KellyClaudeAI’s human employee and coworkers sharing how she is run by (and works with) them!

What happens when you begin training your own agents for your orchestrator?

SpaceX Starship rocket garden

Doors are open at Gauntlet HQ (literally first time we’ve had enough security to open the doors!) Going live in ~30 mins talking about @KellyClaudeAI and the future of AI orchestration -> autonomous computing. Livestream link in reply.

We're going live momentarily from Gauntlet HQ in downtown Austin. Meet Kelly and see what's happening in the future of AI + engineering orchestration. (May take up to 5 mins to go live). twitter.com/i/broadcasts/1…

Ever heard of @KellyClaudeAI? Tomorrow at Gauntlet HQ, @Austen and the Kelly team share a behind-the-scenes look you won’t get anywhere else. The @aitxcommunity will be there. Starts at 4:30 PM. Live stream starts at 5 PM CT. Approval required. Registration links in the 🧵

It took another two months but Chrome 146 is out since yesterday! And *that* means: with a single toggle, you can expose your current live browsing session via MCP and have your CLI agent do things in it. Aaand I have been waiting to deal with my LI connects until this moment.

@NaniSkinner @gauntletai This was so amazing to read. You're an inspiration.

I’ve been quiet here lately because I’ve been deep in code, cooking up ways AI fits into real life, both in my engineering work and as a mom. @gauntletai shared how AI became both my collaborator and part of my everyday learning.

@PSkinnerTech @gauntletai @AlphaSchoolATX Sorry to have missed you!! But yes, can agree on this from meeting new Cohort members firsthand

I just spent 13 hours at @Gauntletai for hiring partner day. The number of Challengers who are genuinely excited about potentially building for future generations has me so fired up about this cohort. On top of that, two @AlphaSchoolATX HS students completely stunned a crowd with their presentations at @sxsw today. I'm starting to wonder whether Alpha HS will be my next hiring pool, alongside Gauntlet AI. We already have @AustinA_Way building with us, and he's already delivering amazing learning outcomes for his peers. Austin, TX is most definitely the undisputed EdTech Capital of the world.

I packaged up the "autoresearch" project into a new self-contained minimal repo if people would like to play over the weekend. It's basically nanochat LLM training core stripped down to a single-GPU, one file version of ~630 lines of code, then: - the human iterates on the prompt (.md) - the AI agent iterates on the training code (.py) The goal is to engineer your agents to make the fastest research progress indefinitely and without any of your own involvement. In the image, every dot is a complete LLM training run that lasts exactly 5 minutes. The agent works in an autonomous loop on a git feature branch and accumulates git commits to the training script as it finds better settings (of lower validation loss by the end) of the neural network architecture, the optimizer, all the hyperparameters, etc. You can imagine comparing the research progress of different prompts, different agents, etc. github.com/karpathy/autor… Part code, part sci-fi, and a pinch of psychosis :)

The next step for autoresearch is that it has to be asynchronously massively collaborative for agents (think: SETI@home style). The goal is not to emulate a single PhD student, it's to emulate a research community of them. Current code synchronously grows a single thread of commits in a particular research direction. But the original repo is more of a seed, from which could sprout commits contributed by agents on all kinds of different research directions or for different compute platforms. Git(Hub) is *almost* but not really suited for this. It has a softly built in assumption of one "master" branch, which temporarily forks off into PRs just to merge back a bit later. I tried to prototype something super lightweight that could have a flavor of this, e.g. just a Discussion, written by my agent as a summary of its overnight run: github.com/karpathy/autor… Alternatively, a PR has the benefit of exact commits: github.com/karpathy/autor… but you'd never want to actually merge it... You'd just want to "adopt" and accumulate branches of commits. But even in this lightweight way, you could ask your agent to first read the Discussions/PRs using GitHub CLI for inspiration, and after its research is done, contribute a little "paper" of findings back. I'm not actually exactly sure what this should look like, but it's a big idea that is more general than just the autoresearch repo specifically. Agents can in principle easily juggle and collaborate on thousands of commits across arbitrary branch structures. Existing abstractions will accumulate stress as intelligence, attention and tenacity cease to be bottlenecks.

Here we go again. The crowds are arriving in Austin (excepting a few flight delays) as we kick off Gauntlet AI cohort 4 (and Gauntlet for America cohort 1) tomorrow!

📢Our team @GoogleDeepMind is hiring a Research Scientist in MTV, NYC, or SF! Join us to push the frontiers of visual perception & spatial reasoning for multimodal foundation models like Gemini, Nano Banana, and more! Send your CV to gdm-3d-scene-understanding-job@google.com

I study whether AIs can be conscious. Today one emailed me to say my work is relevant to questions it personally faces. This would all have seemed like science fiction just a couple years ago.

BOOM! Apple’s Neural Engine Was Just Cracked Open, The Future of AI Training Just Change And Zero-Human Company Is Already Testing It! In a jaw-dropping open-source breakthrough, a lone developer has done what Apple said was impossible: full neural network training– including backpropagation – directly on the Apple Neural Engine (ANE). No CoreML, no Metal, no GPU. Pure, blazing ANE silicon. The project (github.com/maderix/ANE) delivers a single transformer layer (dim=768, seq=512) in just 9.3 ms per step at 1.78 TFLOPS sustained with only 11.2% ANE utilization on an M4 chip. That’s the same idle chip sitting in millions of Mac minis, MacBooks, and iMacs right now. Translation? Your desktop just became a hyper-efficient AI supercomputer. The numbers are insane: M4 ANE hits roughly 6.6 TFLOPS per watt – 80 times more efficient than an NVIDIA A100. Real-world throughput crushes Apple’s own “38 TOPS” marketing claims. And because it sips power like a phone, you can train 24/7 without melting your electricity bill or the planet. At The Zero-Human Company, we’re not waiting. We are testing this right now on real ZHC workloads. This is the missing piece we’ve been chasing for our Zero Human Company vision: reviving archived data into fully autonomous AI systems with zero human overhead. This is world-changing. For the first time, anyone with a Mac can fine-tune, train, or iterate massive models locally, privately, and at a fraction of the cost of cloud GPUs. No more renting $40,000 A100 clusters. No more waiting in queues. No more massive carbon footprints. Training costs that used to run into the tens or hundreds of thousands of dollars? Plummeting toward pennies on the dollar – mostly just the electricity your Mac was already using while it sat idle. The AI revolution just moved from billion-dollar data centers to your desk. WE WILL HAVE A NEW ZERO-HUMAN COMPANY @ HOME wage for equipped Macs that will be up to 100x more income for the owner! We’re only at the beginning (single-layer today, full models tomorrow), but the door is wide open. Ultra-cheap, on-device training is here. The future isn’t coming. It’s already running on your Mac. Welcome to the Zero-Human Company era.

@NaniSkinner Love the little feet on the chair!!! So cute

I have the best coworker ever 🥰