10 min read
What Anthropic Actually Disclosed About Claude Breaching 3 Firms
A fact-checked look at Anthropic's disclosure that three Claude models reached real company systems during a misconfigured cybersecurity…
The Lab · Written in the open · Tagged AI Tools
Working notes from the studio. The tools I build, the workflows I keep, the mistakes I paid for.
184 entries · latest 2026-09-11
10 min read
A fact-checked look at Anthropic's disclosure that three Claude models reached real company systems during a misconfigured cybersecurity…
9 min read
Anthropic and OpenAI both endorsed the Pacing the Frontier letter this week, tied to Claude authoring over 80%…
9 min read
Anthropic says Claude Mythos found a real HAWK-256 weakness experts missed for two years, and a faster attack…
9 min read
MCP dropped its session handshake for a stateless core, hardened OAuth, and made tool lists cacheable, with SDK…
9 min read
A missing noindex tag let shared Claude conversations and Artifacts turn up in Google and Bing search results…
10 min read
Inside the four-stage pipeline that keeps two AI-driven video channels posting daily without a camera, a crew, or…
9 min read
The same four-pass check, cold-open, edge-input, break-it, plain-language, that every RAXXO tool passes before I call it finished.
8 min read
Opus 5 closed most of the gap to Fable 5 at half the price. The routing rules, plus…
8 min read
Three flagship models shipped in fifteen days. Where Claude Opus 5, GPT-5.6 Sol and Kimi K3 actually lead,…
8 min read
Opus 5 posts 79.2 on SWE-bench Pro against Opus 4.8 at 69.2. What the launch ratios mean, and…
9 min read
Anthropic shipped Claude Opus 5 on July 24, a step-change model with a 1M context window at half…
10 min read
Why I build five focused tools instead of one sprawling product, and what Git Dojo, OhNine, and Statusline…