Perplexity: auto skill config cut tool-call failures by 11% in production test
perplexity_ai · x · 2026-11-20
Perplexity shared results from its production randomized experiment on autonomous skill configuration:
- Tool call failure rate dropped 11% across all Computer activity, and 7% excluding the system's own background runs
- Medium- and high-dissatisfaction threads per user dropped 7.6%
- Cost and response time showed no measurable change
Related event: Perplexity Rolls Out Auto-Configured Personal Skills to All US Users(4 posts)→
More from coding & agent
- Browserbase uses agents to patch, build, and deploy 40M lines of Chromium code in under a day — blaizedsouza · 2026-11-20
- Cognition launches Devin Bots: Exa runs training and CI with agent teams — marvinvonhagen · 2026-11-20
- Eval-driven development: pick your model, retrieval, and agent design by testing, not guessing — blaizedsouza · 2026-11-20
- A bookmark-worthy evals resource pack: general, agentic, skills, and eval-driven development — blaizedsouza · 2026-11-20
- Open-source Harness Engineering course: 14 lectures, 8 projects on making coding agents reliable — blaizedsouza · 2026-11-20
- Google Cloud outlines 5 research-backed patterns for agent orchestration — rseroter · 2026-11-20