Thought-Provoking Conversation
Kanjun and Dylan delve into the encoding of goals through datasets and the inherent biases within data, sparking a discussion on the recruitment of resources into shallow value measures. Their exchange sheds light on the complexities of unbiased data representation and its implications.In this clip
From this podcast
Generally Intelligent
Episode 10: Dylan Hadfield-Menell, UC Berkeley/MIT, on the value alignment problem in AI
Related Questions