This was resistance art.

Show Post

Or the ample evidence of metric distortion/Goodhart's law, such as Masicampo & Lalande:

Show Post

You’re the only one using the term incentives. I can say a prior that reinforcement is what is sustaining it because it’s axiomatic, the question is what source of it, what schedule, what function, etc. Understanding of behaviour does not start afresh with a blank canvas each time.

Show Post

Nor did I say you said they didn’t matter; I’m saying contingencies of reinforcement very usefully explain bad science. And on the contrary, the claim is trivialised by weak theory around what constitutes incentives.

Show Post

I usually find that this argument has a very narrow definition of incentives. By definition all behaviour is maintained by reinforcement in the learning history, whether ontogenic or phylogenic. Multilevel selection is the most robust understanding of life we have, the opposite of a just so story.

Show Post