-
Codex auto-review let Codex kill all my running processes
me: what keeps killing processes and claude sessions?
Claude Code: … At 13:02:44 your systemd user manager (user@1000.service) was killed. When it died, systemd tore down your user session… Found the cause: Codex’s auto-review policy let a model self-approve rerunning the suite outside the sandbox after two 137 failures, with no human in the loop.…
-
Quoting Anthropic’s Thariq Shihipar
“what I see a lot of times is people hit rate limits when they’re doing this sort of, oh, it did a lot of work, and you’re like: Oh, I don’t like this. Can you undo this and redo it? And then you’re iterating on this thing that the model could have done if you had spent more upfront time or given it better context”
-
Quoting Simon Willison on Unlocking Coding Agents’ Potential
The more time I spend working with coding agents, the more convinced I am that they make software engineering even harder.
We can do amazing things with them, but unlocking their full potential requires extraordinary discipline and knowledge.
-
Quoting Anthropic: Opus 4.8 Safety “somewhat less robust”
Agentic safety. Although it shows improvements in some areas (such as refusing malicious requests), we found Opus 4.8 to be somewhat less robust than Opus 4.7 in several agentic contexts (such as vulnerability to prompt injection attacks). However, the application of our safeguards closes the gap between the models in practice. […]
-
Introducing GHA-bench
GHA-bench is a benchmark and a set of evals for how well different coding agents author and test GitHub Actions using different languages.