Identify and explain (in a paragraph) the most surprising bit of information you learned.
I thought it was really interesting that there was a specific team at Deepmind studying what the code is doing for Alpha Go. The "psychology of AI" is a cool field, especially because we don't know exactly what embeddings algorithms are computing. The many activation functions and layers embedded could have some emergent behavior that we can't see, like racial biases, which could emerge in new predictions. So, this field will be incredibly important in rectifying any biases AI might learn from human-generated data which would have biases embedded in it.
The AlphaGo documentary discussed in the interview: https://youtu.be/WXuK6gekU1Y?si=L0-_RO7rOw_hz6Py