This study raises the interesting question of the interplay between trained biases in ChatGPT and the ability to coax this AI to transcend those biases.
The analogy with chess is apt — computers play chess but in ways different from us by being able to brute force their way through millions more positions.
Yet I continue to maintain that human intelligence is qualitatively different from artificial intelligence.
In a conversation with Lex Fridman, Magnus Carlsen betrays no sense of empathy for how his view that life is an accident might negatively impact others.
I had the privilege of speaking on intelligence metrics. Here are my slides. I hope soon to develop this talk into a proper peer-reviewed paper.