Why AI benchmarks fail, how safety gets measured wrong, and what real evaluation should look like with Dr. Sean McGregor
Dr. Darryl Wright joins us to discuss his background and mechanistic interpretability.
