Now
Current essay: The Imitation Game Was Never a Benchmark — weeks 1–5, ≤8,000 words.
Public research checklist:
- Judge recruitment message sent (day 1–2)
- Reading pass: Turing 1950 → draft §2 immediately after
-
experiments/a1-imitation-pilot/protocol.mdfrozen and committed before any data collection - Pilot run: 24 prompts × 5 categories, 3–5 judges, length-matched answers, randomized order
- Analysis + honest labeling as an exploratory pilot
Reading budget: 10–12 h across weeks 1–3. Next in queue: Bun’s Zig-to-Rust Rewrite Was Not a Prompt (the B2 ledger is already accumulating).
Last updated: 2026-07-19.