The Problem With Too Much Training Data

The problem would be that there wouldn't be enough training data. They put people in an FMRI scan and then for 16 hours, they had them listen to a podcast. The question was: Would the AI be able to infer what they were thinking as they were watchingThat movie? Okay. So let me give you some examples. Here's one. I got up from the air mattress and pressed my face against the glass of the bedroom window,. expecting to see eyes staring back at me, but instead finding only darkness. That's pretty darn close. Whoa. It probably was not the SGU, but it should have been.