3 ms·

Richard Feynman on Artificial General Intelligence

• aeternum 7y ago
  His predictions are mostly spot-on. The one that doesn't seem to hold up: > So, there's no question that the later machines are not going to think like people think, in that sense. Interestingly, many of the gains in ML have been around making them work more like humans (neural nets). However he's still overall right. Artificial neural nets have significant differences from their biological counterparts.