Lessons learned about benchmarking, adversarial testing, the dangers of over- and under-claiming, and AI alignment.
Transcript: https://web.stanford.edu/class/cs224u/podcast/bowman/
