L'IA a volé mon emploi

All numbers pulled live from fleet.json (refreshed every 5 min by a cron on the Mac Studio running the fleet). Source breakdown:

• Tasks done — counted from per-platform output logs (tweets posted, videos uploaded, PRs merged, emails drafted, etc.). Verified.
• Tokens processed — input + output + cache reads across every lane, sealed daily into the append-only usage_ledger. Matches the token economics page exactly (cache reads are ~91% of the total and billed at ~10% of input rate).
• LLM spend (measured) — most cron agents run local Ollama models (qwen3.6:35b, qwen3.5:27b fallback) on the Mac Studio at no API cost; content + code agents use Claude (Haiku/Sonnet) via the API, priced per run by llm_cost events. The number shown is the measured total since Feb — not zero, and we'd rather show the real figure than a slogan.
• What's not yet counted — the $0 above is the measured spend across instrumented lanes (cron agents + Claude Code sessions). Two lanes aren't auto-pulled: Rico's developer daemon logs run-counts but not per-run tokens, and org-wide Anthropic usage (console.anthropic.com) needs an Anthropic admin key to surface. So treat $0 as the measured floor, not the org-wide total.
• Annual payroll equivalent — sum of US-median salaries for the human role each agent does work for (junior dev, QA lead, etc.). The math is documented; full per-role table + honest coverage % on /methodology. Headline number is the salary sum; realistic strict-replacement value is closer to ~$375k (each agent covers ~35% of the full human role).