4 ms·
The key issue is that Intelligence requires conceptual awareness of space time. Reinforcement "learning" is about statistics, very likely how our perception fa
by gibsonf1 3y ago
The key issue is that Intelligence requires conceptual awareness of space time. Reinforcement "learning" is about statistics, very likely how our perception faculties work. But that is just the input to intelligence, not intelligence itself.
- AbrahamParangi 3y agoWith whatever respect is due, what you're talking about a vague psychologically based theory of intelligence that actually has a very poor track record of prediction - meanwhile "dumb" statistical ML has no clear theoretical barrier regarding what it might not be able to do! Let me say that again, if you believe that there is a specific test that AI today will fail tell me it, and I will show you that it either passes it today or that we can be fairly sure it will pass it in some years time.