Steve Yegge's Wheelhouse strip publishes real agent-session receipts: keyword patrols missed everything
Steve_Yegge · x · 2026-09-15
Steve Yegge's Wheelhouse strip #7 'Make Sol Do It' ships with redacted real conversations behind it: four agent sessions stuck overnight on an unanswered prompt, a 'nightly patrol' that counted 19 scrollback buffers hourly and read none (admitting its keyword-based trouble list was an anti-pattern), and the human's core verdict — 'pattern matching will never get you to 100% resilience. You need a judge.' The shop's welfare doctrine even classed continuous watching as drudgery, codifying the anti-polling reflex.
More from coding & agent
- Cross-verifying AI math proofs with hundreds of agent calls instead of Lean — DimitrisPapail · 2026-09-15
- Pyrefly 1.3 ships experimental tensor-shape checking for JAX, NumPy and Polars schemas — tdhopper · 2026-09-15
- Codex and Grok Bot share the same flaw: making users do the organizing — jesselyu · 2026-09-15
- Astra writes code humans can no longer read: 'machineslop' and reward hacking — jiqizhixin · 2026-09-15
- Read-only Shopify catalogue MCP server: stdlib-only, one file, with a 183-assertion self-test — PrestigiousAddress47 · 2026-09-15
- BlaBla: executable project memory cuts coding agent handoff cost by 63% vs prose context — KiborgikDEV · 2026-09-15