...To the surprise of 0 people who are at all in touch with the reality of software development today.

#arXiv, the granddaddy of #preprints servers, has announced it’s splitting from Cornell U., its long-time host institution. Leaders there explain why. #OpenAccess #ScholComm @science.org www.science.org/content/arti...

Esolang-bench esolang-bench.vercel.app is making the rounds, but I think their claims are deeply misguided. Being a worse programmer in Brainfuck than in Python says absolutely nothing about whether LLMs memorized the right solutions. *Anyone* would be a worse programmer in these languages.

Interesting developments

For the first time, researchers have simulated nearly every chemical reaction in a living bacterial cell go.nature.com/4bp0jOy

“When AI Discovers the Next Transformer” Full Interview on YouTube: youtu.be/EInEmGaMRLc Robert Lange (Sakana AI) joins Tim Scarfe (ML Street Talk) to discuss Shinka Evolve, a framework that combines LLMs with evolutionary algorithms to do open-ended program search.

Probably the coolest thing I've seen this decade: eon.systems/updates/embo... An embodied fruit fly connectome running on a computer!

Updated my Mirage app hallucinator to support Gemini 3.1 Flash-Lite. It renders much faster now, almost usable. (video has just my typing sped up) youtu.be/qiMwqIq1_hI github.com/danielmewes/...

Introductions Gemini 3.1 Flash-Lite 🔦, a huge step forward on the boundary of intelligence beating 2.5 Flash on many tasks. x.com/OfficialLoga...

results from running my search for interesting neural cellular automata overnight

Today we're open-sourcing the Darwinian Evolver, a near universal optimizer for code and prompts. Our method uses LLM-driven, open-ended evolution to improve code. It also improves reasoning: We only *barely* missed out on hitting a new high score in ARC-AGI-2 (public) :) imbue.com/research/202...

Big if true! 😂

Lots of interesting findings about the effects of AGENTS.md files on coding agents in this paper. Surprisingly, there was very little lift from having one, and LLM-generated files significantly hurt performance. arxiv.org/abs/2602.11988

Short musings on "cognitive debt" - I'm seeing this in my own work, where excessive unreviewed AI-generated code leads me to lose a firm mental model of what I've built, which then makes it harder to confidently make future decisions simonwillison.net/2026/Feb/15/...

This seems like a great use of today's LLMs in maths and science: Use the model to come up with flaws in an argument or find issues with a theorem. youtube.com/shorts/8s4W5...

As jobs begin to shift with AI, I think there will be increasing numbers of people feeling like Aditya (former CTO Dropbox, early Facebook engineer, etc). People are more psychologically resistant to major changes than you might expect, but that doesn’t mean it will be easy to reconstruct meaning.

Tesla used to be S3XY. Now it's just... OC3Y? (Optimus, Cybertruck, 3, Y)

A thought provoking special episode from Machine Learning Street Talk youtu.be/pO0WZsN8Oiw?...

My two favorite insights from www.anthropic.com/research/ass... : 1. The assistant persona axis already exists before post-training. 2. I really like how they were able to plot the persona drift over the course of a conversation

100% youtu.be/FyStPztM7eo . Things are really off in our current society. (video by @hankgreen.bsky.social )

As someone who still finds typing on a touchscreen incredibly slow and frustrating, I'm very excited about the Clicks Communicator. It's basically a BlackBerry! clicksphone.com/communicator Even comes with a notification LED!

Has anyone tried the Nyxt browser? nyxt.atlas.engineer It's cool that it's written in Lisp and comes with a Repl. I'm running into some slow UI updates though.

GPT 5.2 (thinking on high) seems to be buggy. I'm seeing reproducible issues where it starts glitching out in the middle of its response with garbage tokens and then ends the message. Highly reproducible for me every 10th-20th prompt or so.

Gemini 3 Flash beats Gemini 3 Pro in both ARC-AGI 2 (33.6% vs 31.1%) and in SWE-bench Verified (78% vs 76.2%). 1/4th the cost and faster. This is very very impressive. Is there still a reason to use 3 Pro? blog.google/products/gem...

LLMs can generate arbitrary apps on the fly. Without ever writing source code. I wrote about this here: wp.me/pdyMeH-bd This was really fun to play around with. You can try it out yourself using github.com/danielmewes/...

I do think it's most likely that this was indeed due to teleoperation. But I also think it would be funny if this behavior got accidentally trained into the model due to operator-gathered training data that didn't get trimmed correctly. :D

Poetiq's methodology on top of Gemini 3 & GPT 5.1 exceeds average human performance on ARC-AGI-2! This is huge. Only caveat is that they evaluated on the public set - it might have been used in post training of Gemini 3? Looking forward to see private eval results! poetiq.ai/posts/arcagi...

There have been a few works that showed signs of LLM introspection since I wrote my article. Though the consensus still seems to be that it's quite unreliable. amongai.com/2024/12/24/l...

This was a tough but necessary decision - I posted my own notes on this here, from the perspective of a current PSF board member simonwillison.net/2025/Oct/27/...

I have to admit the mechanism behind cable impedance and impedance matching never really clicked for me, despite being a licensed radio amateur for 20 years. ...until I watched this video by AlphaPhoenix - it's such an amazing visualization of what's going on in a cable! youtu.be/RkAF3X6cJa4?...

Claude Haiku 4.5 outperforms Sonnet 4 in some coding benchmarks (such as SWE-bench Verified). This is exciting, since it's 1/3 the price. However, in my actual use, I've found it to be a bit underwhelming compared to Sonnet 4 (not to mention Sonnet 4.5). What has been your experience with Haiku?

This is so true. I think almost nobody is talking about symbolic GOFAI these days, so I'm not concerned about that. But all of ML being re-branded to AI lately, while AI has simultaneously been made synonymous with generative AI in many places, has led to so much confusion.

The age of single-serving, disposable (simple) software is here. I now frequently have LLMs write software for me that I use exactly once. Often to convert some data or create visualizations, but also one-off new features to add to some open-source application that I only want to use once.

new blog post! why do LLMs freak out over the seahorse emoji? i put llama-3.3-70b through its paces with the logit lens to find out, and explain what the logit lens (everyone's favorite underrated interpretability tool) is in the process. link in reply!

Today we're releasing Sculpture, an AI coding tool that combines async agents with the ability to collaborate with agents locally. It also comes with built-in verifiers to automatically check the quality of AI written code. More to come! imbue.com/sculptor-ann...

I used AI vibe coding tools to port Anthropic's API client to Common Lisp: github.com/danielmewes/... It was a very quick and fun mini project that taught me a thing or two about Lisp. Use it at your own risk.

I feel like each of Anthropic's three post mortems is missing some key explanation step in their root cause descriptions. www.anthropic.com/engineering/... 🧵

My Hulu / Disney+ subscription will be preempted indefinitely.

As AI systems keep getting better at very hard problems while getting more opaque, the way that we work with AI is shifting shifting from being collaborators who shape the process to being supplicants who receive the output. I discussed what that means. www.oneusefulthing.org/p/on-working...

Meta trained a special “aggregator” model that learns how to combine and reconcile different answers into a more accurate final one, instead of relying on simple majority voting or reward model ranking on multiple model answers.

The funny thing about the prediction that AI would be writing 90% of all code by now is that the prediction's failure distracts from the fact that AI adoption in code writing is actually extremely high, it was already over 30% in December, 2024 according to one measure, with large economic impact.

I remember seeing this graph and thinking it was about the influence of training data on LLM responses as well. Quite misleading.

I think we're starting to see diminishing returns from LLM pre- and post-training. The limitations of today's LLMs are unlikely to just disappear with the next bigger model. This is not all bad: we can start focusing on how to work around those limitations and how to put current LLMs to work.

Just experienced phantom breaking for the first time in my Kia with HDA2 (Highway Drive Assist) and it made me sentimental about the exciting times of early Tesla Autopilot 😢

Thanks to the genius of a well-thought-out tariff policy, manufacturing trucks (for the US market) in Mexico now incurs lower tariffs than manufacturing them domestically. 🤦 www.reuters.com/business/aut...

What if we could evolve AI models like organisms, letting them compete, mate, and combine their strengths to produce ever-fitter offspring? Excited to share our new paper, “Competition and Attraction Improve Model Fusion” presented at GECCO 2025 (runner-up for best paper)! arxiv.org/abs/2508.16204

Study finds that polarization and echo chambers on social media may be inherent to the medium and not due to algorithms. www.science.org/content/arti...

Impressive result: 99.9% on AIME 2025 with both open models and 85% fewer tokens! The trick? Measuring confidence in the reasoning traces as they are being generated, then ending low confidence traces early. jiaweizzhao.github.io/deepconf/

It seems like there is not enough of a policy response to the existence of self-driving cars, with 57M miles of data, Waymo’s autonomous vehicles experience 85% less serious injuries & 79% less injuries overall than cars with human drivers 2.4 million are injured & 40k killed in US accidents a year

One thing I've learned is that when Americans call something "European style", it's 100% not like anything I've ever seen or eaten in Europe. I think it's just an attribute used to mean "a bit different from how we usually do it in America"?