Nathan Young comments on Exaggerating the risks (Part 20: AI 2027 timelines forecast, benchmarks and gaps) | Reflective Altruism