Sam Gershman joins us for a discussion on both AI, and questions in Philosophy of Science. 
In the podcast we talk about:

1) Whether observation and interpretation can be 'theory neutral'.
2) How auxiliary hypotheses can 'buffer' our beliefs.
3) A Bayesian framework for evaluating our auxiliary hypotheses.

3) The aspects of human cognition that AI currently lacks.
4) Advancing AI, and the AI alignment problem.

Sam is a Professor within the Psychology department at Harvard, where his lab works on questions in learning, memory, and decision making from a computational cognitive neuroscience perspective.