When Will a Recursively Self-Improving AI Model Beat God of War?
imjustnewatai · x · 2026-08-06
A user envisions a benchmark where a recursively self-improving (RSI) model plays a complex single-player game like God of War, learning from repeated failures, remembering why, and eventually beating it from the open world. This reflects expectations for AI to learn autonomously and evolve through trial and error.
Related event: Researchers Propose Video Games to Test AI Recursive Self-Improvement(3 posts)→
More from AGI Musings
- Calling for Decentralized AI: From the Cloud Back to a $5K Local Personalized AGI — Dan_Jeffries1 · 2026-08-06
- Why AI Agents Struggle to Go Mainstream: The Bottleneck is Go-to-Market, Not Tech — Dan_Jeffries1 · 2026-08-06
- Practical AI podcast: Models, Harnesses, and Multi-Agent Systems explained, plus enterprise adoption tips — Practical AI · 2026-08-06
- Sabine Hossenfelder & Gary Marcus: Current AI Architectures Won't Reach AGI — skdh · 2026-08-06
- Approaching the AI Transition: Humans as the Next Bottleneck — aiamblichus · 2026-08-06
- Testing Recursive Self-Improvement in AI Through Video Game Benchmarks — imjustnewatai · 2026-08-06