Raw LLM Responses

Hi! I’m a PhD student in behavior analysis and a Board Certified Behavior Analyst. I enjoyed this episode. Human behavior is predictive but much more complicated than the statistics currently used by cognitive-goers. In fact, you don’t need too much group information about others (as a molar or zoomed out analysis is not too helpful for predicting behavior). Analysts can predict behavior by a more molecular (zoomed in and highly individualized) discrete analysis. Very generally speaking, if I know some of your own history of behaviors, and current situation, I’m likely to predict your behavior in an array of situations. Behavior is lawful like all other sciences. Therefore, it goes beyond math and enters contextual sciences that consider not just evolution of species (selectionism) but societal evolution (selectionism in terms of behaviors that work are those that continue) and biology as well (continuously receiving feedback for selectionism from the environment). I doubt that a truly successful algorithm can be created without a PhD in behavior analysis collaborating with other PhDs in the aforementioned sciences. Math is needed but not in the statistical group data that we currently use. It’s a bit more complicated than that. Indeed AI is necessary. I recently shared with my class a theory for this. I’d like to teach AI to enter into derived learning through recombinative generalization principals to learn more about predicting behaviors. It’s pretty cool. 😎