Clauday.

On the same day Cloudflare and Amazon shipped decision models, a paper showed the failure mode that matters most for them. "More Choices, Fewer Decisions" (arXiv 2609.38827, 49 upv…

If harness design decides how well an agent does, the next step is obvious: let agents design the harness. MILO (arXiv 2609.38349) does this, and it beats every hand-built harness …

Terminal agents have a problem that chat models do not: a bad action changes the world. Install the wrong package and every later step works in a broken environment, even if the mo…

Let an agent write its own homework and grade it too, and eventually the proposer and the grader learn to make the same mistakes. "False Frontiers" (arXiv 2609.39102, 180 upvotes o…

Google's most capable model is out, and almost nobody can use it. Gemini 4 Argon launched on September 30, but it is going only to trusted cyber defenders in Google's Fairwind Prog…

Every skill scanner on the market checks skills one at a time. Skill cascading attacks are built to pass exactly that test. The paper, from Zihao Zhu, Siwei Lyu, Adel Bibi and Baoy…

The feared failure of chain-of-thought monitoring was secret code: a model under monitor pressure learning to hide its real reasoning in text humans cannot read. Julian Schulz's ne…

Offensive-security benchmarks are saturating, and ScopeBench argues they were measuring the wrong thing anyway. In a real penetration test the question is not whether an agent can …

Your agent does not need to be hacked to be robbed. It just needs to remember. That is the thesis of Persistent Billable State, a paper by Jinqian Zhang and six co-authors on what …

The Jev backlash arrived on Sunday. The Normalization of Inexplicable Failures, a blog post on ihatethefuture.com, hit 224 points on Hacker News with a blunt charge against calibra…