Microsoft's Horvitz on Clippy: Bayesian Prototype vs. Shipping Product
erichorvitz · x · 2026-08-20
Eric Horvitz shares the backstory of Clippy, revealing that the Microsoft Research prototype used Bayesian inference to predict user needs. The shipping product differed significantly due to scaling challenges, using a lighter inference variant and removing autonomous initiative. The infamous proactive behaviors were actually driven by a legacy rules engine, not the AI prototype.
More from Companies & People
- YC's codebase hits 4M lines, adding 2M and 350+ internal AI tools since March — garrytan · 2026-08-20
- Clarification: Opinions are personal, not company representative — chris_j_paxton · 2026-08-20
- OpenAI Researcher Departs to Focus on Formal Verification of Frontier AI Systems — HanchungLee · 2026-08-20
- Elorian Hiring: Building AI Systems for Native Visual Reasoning — dhruv2038 · 2026-08-20
- 200+ Interviews: The 5 Most Useful Questions — yangyi · 2026-08-20
- Beijing robot output up 2.3x in H1; Zhiyuan A3 wins China & EU certifications — 创业邦 · 2026-08-20