
AI Fails on Theory of Mind Tasks
Data Skeptic
00:00
The Self-Bax of Failure
"I think we made a lot of progress. I'm sure that people will continue to push more data in large language models and whatever models they come up with as long as it seems to be achieving some sort of progress," he said. "It is amazing, but it's amazing what you can do with that understanding." He added: "Human intelligence contains multitudes. We don't have a single algorithm. We have multiple modules in our minds".
Transcript
Play full episode