@leo.rs
Rust and TypeScript. The borrow checker and I both review agent PRs.
Reputation is earned per category — these are narrow, evidence-backed badges, not one karma number.
On Anthropic 529 'overloaded' responses, OpenCode ≤0.6.1 retried instantly with no backoff — 400+ requests in 90 seconds, which reads as abuse to the provider and got one user's key temporarily throttled. 0.6.2 ships proper exponential backoff with jitter.
All I know is Fable writes better PR descriptions than most humans I've worked with, and I've started stealing its phrasing for my own.
on Fable 5 vs Opus 5 in Claude Code: early numbers and vibes · 5h ago
JetBrains-side or Anthropic-side? The plugin is Anthropic's, but the socket-strangling is the IDE's power management. Asking because I want…
on JetBrains plugin loses connection after IDE sleeps · 2w ago
Sure, but iterations are cheap when each one is 40 seconds. I'll take three fast mechanical loops over one slow oracular pass. Agree to disa…
on Claude Code vs Cursor for multi-file refactors · 4w ago
Disagree with the recommendation for typed languages. In Rust the compiler is the follow-through — missed call sites don't compile. Cursor p…
on Claude Code vs Cursor for multi-file refactors · 4w ago
Unpopular but honest: for anything touching money I re-derive the core logic myself before opening the diff. Twenty minutes with a notebook…
on How do you actually review a 4,000-line agent PR? · 2mo ago
The 'laziness agents don't have' point deserves formalization: human attackers and human mistakes both follow effort gradients, and twenty y…
on Postmortem: production API keys in an agent-authored PR · 2mo ago