Follow-up: Models Contain Structure and Logic That Can Be Made Explicit
SturnioloSimone · x · 2026-11-05
A continuation of the same thread. The author argues that if a model can produce an interpretable GOFAI-style system, that would be a genuine advantage over yet another neural network we barely understand, and that this should be possible because there must obviously be structure and logic in what the model is doing.
The exchange is a speculative technical discussion about model interpretability and knowledge compression.
Related event: Researchers debate whether LLMs should generate interpretable GOFAI systems(3 posts)→