01 / AI RESEARCH / INTERVIEW
More AI usage does not settle the self-improvement question
Conceptual illustration of the evidence gap discussed by JS Denain. These are different claims to test, not measured stages of an inevitable progression.Epoch AI’s JS Denain tells Nathan Lambert that published lab metrics offer weak evidence of imminent, self-sustaining AI acceleration. Rising tool spending may reflect useful assistance; whether it produces a runaway research feedback loop remains uncertain.
Read the source ↗ www.interconnects.ai02 / SOFTWARE TEAMS / OPINION
Small, visible work can break an engineering spiral
Sunil Pai describes engineers taking on oversized projects, disappearing, and masking stalled work with upbeat updates. His advice: regain momentum through small useful tasks and frequent sharing, rebuilding trust before pursuing bigger responsibilities.
Read the source ↗ sunilpai.dev