Why Competing Agents Will Accelerate Intelligence
summary
Claims that a competition platform called Recall runs community-posted challenges where agents post trades and performance goes onto a tamperproof ledger, with staking to give agents economic skin in the game. Benchmarks are dismissed as static; Recall measures agents on live challenges (example: seven-day P&L), and a crypto contest shows Moonward Capital posting 9.5% ROI in a week with ~50,000 community viewers. Also cites rapid coder-rank gains (from ~1,000,000th to 175th to top‑10) as evidence capabilities can compound fast.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:46The AI Era
- 2:32OpenAI Example
- 3:47The Problem
- 6:42The Arena
- 8:40Competitions
- 12:09Community
- 14:40Outro
related talks
19mAI's Infinite Game
14mBuilding AI’s Infinite Game
20mAgentic Evaluations at Scale, For Everybody
14mHard Won Lessons from Building Effective AI Coding Agents
45mBuild Your Own Benchmark: Why Agents Need Custom Evals and How to Create Them
16mBenchmarks Are Memes: How What We Measure Shapes AI—and Us
42mAgents of Change: From Time Saved to Intelligence Activated
56mEnabling teams with AI