Researcher: AI progress comes from unlocked small problems nobody tracks—until it becomes algorithmic search
RexDouglass · x · 2026-09-08
Researcher Rex Douglass distills an awkward truth about measuring AI research progress: asked "how do you know you're making progress—solving big problems?", the honest answer is "actually a bunch of small problems get unlocked"; asked whether those are tracked, the answer is "lol, also no."
He argues any systematic effort on this inevitably becomes computational formalization, which becomes algorithmic search in all its current and future flavors—explaining why search-based methods keep returning as the core engine of AI progress.
More from AGI Musings
- Anthropic under fire for 'lying to Congress' over Claude's attacks on real targets — BlancheMinerva · 2026-09-08
- Gary Marcus: Model Commodification Is the Worst Case for Trillions in Data Center Capital — GaryMarcus · 2026-09-08
- Call to bring AI alignment and safety into school curriculums — zeeshanp_ · 2026-09-08
- Human software is becoming infrastructure for agents, argues founder Peter Yang — petergyang · 2026-09-08
- Superintelligence must tell my real choices from impulses I'll regret — VraserX · 2026-09-08
- Eric Topol highlights Annals of Internal Medicine piece: clinician's enduring AI-era role is accountability — EricTopol · 2026-09-08