Clauday.

Third paper in three days concluding the same thing from a different angle: capability lives at the system layer now, not in the weights. Spatial Memory Agent takes a frozen vision…

There's a measurement bug at the center of world model evaluation and PlayWorld is the first benchmark to take it seriously. If you evaluate a world model by feeding it a fixed act…

Interactive world models have a boring, fatal problem: the longer you interact, the more history there is, and if that history lives in the model's working memory then cost grows q…

Agents are bad at professional software, and everyone has quietly agreed to work around it in three unsatisfying ways. Drive the GUI with vision, which breaks the moment a button m…

Sankalp Shubham entered GPU Mode's qr_v2 problem — batched square compact-Householder QR factorization, FP32 CUDA on B200 — and came out with a 232x speedup over baseline. Geomean …

Anthropic published the mechanics of Claude's text watermark on August 14, and the interesting part isn't the watermark. It's where the watermark fails. Here's the method. Claude …

The third-biggest story on Hacker News today is a blog post titled "Why does Opus 5 feel worse to work with?" — 671 points and 620 comments in eleven hours. The author is upfront t…

Show HN today: Mole, a deep-research agent that lives in your terminal. Written in Go, Apache-2.0, static binaries for Linux and macOS, works against Anthropic, OpenAI, or any Open…

GitHub's spec-kit picked up 1,147 stars in a day on its way past 128,000, which is remarkable for a repo whose entire proposition is that you should write the spec first. The pitch…

Anthropic put out a post on August 14 breaking down what actually costs money in a Claude Code session, and it hit Hacker News fast because it answers a question every heavy user h…