Simpledot learns every night.

Every answer produces outcome signals — which engine an independent judge picked, what our critic flagged, what users thumbed down. Each night those signals are distilled into a new policy that shapes the next day's answers. No query text is ever stored, only outcomes. This page is the public record of what Simpledot learned.

1. 2026-09-10Policy v11119 signals analyzed
   • Chat engine ranking updated to OpenAI, Anthropic, and Google based on 111 judge outcomes.
   • Code evaluations had insufficient volume to reliably rank engines.
2. 2026-09-08Policy v1080 signals analyzed
   • Prioritized OpenAI models for chat tasks based on 32 judge wins, followed by Anthropic (27) and Google (11).
   • Deferred code engine ranking due to insufficient sample volume (3 total runs).
3. 2026-09-05Policy v963 signals analyzed
   • Chat routing now prioritizes Anthropic and OpenAI engines, which tied with 22 judge wins each over Google's 7.
   • Code evaluations were omitted from priority updates due to insufficient sample size (2 queries).
4. 2026-09-03Policy v880 signals analyzed
   • Chat routing now prioritizes OpenAI, Anthropic, and Google based on 68 evaluation wins.
   • Code sample size was too small to reliably rank engines.
5. 2026-08-31Policy v728 signals analyzed
   • Observed truncated scripts in code answers; added a rule to ensure code is complete and fully executable.
6. 2026-08-25Policy v646 signals analyzed
   • Chat engine priority updated to rank OpenAI ahead of Anthropic based on 29 judge wins out of 34 scored comparisons.
   • Code engine prioritization deferred due to low sample volume (3 total interactions).
7. 2026-08-23Policy v560 signals analyzed
   • Chat engine priority set to OpenAI (32 wins), Anthropic (8 wins), and Google (5 wins) based on 55 judge evaluations.
   • Code task engine ranking omitted due to insufficient sample volume (5 total requests).
8. 2026-08-18Policy v430 signals analyzed
   • Chat answers now prioritize OpenAI engines based on leading judge win rates over the last 24 hours.
   • Code engine prioritization omitted due to insufficient query volume.
9. 2026-08-16Policy v354 signals analyzed
   • Chat routing now prioritizes OpenAI engines based on leading judge win share.
   • Code routing left unchanged due to insufficient sample size over the last 24 hours.
10. 2026-08-13Policy v2150 signals analyzed
    • Omitted engine priority updates as judge wins were concentrated entirely within OpenAI models.
    • No prompt rules or search hints were generated due to zero recorded defects and thumbs-down signals.
11. 2026-08-10Policy v1198 signals analyzed
    • Prioritized OpenAI engines for chat tasks based on top judge win counts over the last 24 hours.
    • Maintained empty prompt rules and search hints due to low overall defect and feedback rates.