Self‑improving AI systems can rewrite the designs that govern their own behavior, yet most existing approaches focus improvement on a few parent agents and their evolutionary experience, leaving the broader population underexploited. We find a sizable gap between population coverage—the set of tasks solved by at least one evolved variant—and the performance of the best individual agent. This gap indicates that useful capabilities and diagnostic evidence emerge across the population without being consolidated into the leading agents. These observations motivate the view‑from‑nowhere principle: an agent should be improved by interpreting its behavior through population‑wide patterns of success, failure, and prior self‑modifications. We instantiate this as Population‑Centered Evolution (PCE). PCE distills task records and modification histories from the entire population into a population block that captures recurring failures, complementary successes, and the effects of previously attempted modifications, then uses this block to guide improvements to a selected parent’s harness. Under matched models, initial harnesses, and evolution schedules, PCE achieves 72.1% success on SWE‑bench Verified and 96.7% on Polyglot, improving over a strong group‑evolution baseline by 3.2 and 5.9 percentage points respectively. In our evaluation, PCE reduces the gap between population coverage and best‑agent performance by one‑third on SWE‑bench and closes it entirely on Polyglot. These results suggest that treating the population as evidence—not merely as candidates for selection—offers a complementary principle for agent evolution.
Review