Oddbean new post about | logout
▲ ▼
Adam F. Lawton | 1 years ago (raw) | root | parent | reply | flag 
 @07515e85 self-replicating and self-improving machines don't make you consider possibilities and consequences? Ok.
▲ ▼
19f681e0 | 1 years ago (raw) | root | parent | reply | flag 
 @2a70c54d 

No, we're still not remotely there.  Humans with too much power controlling them are the gravest danger right now.
▲ ▼
Adam F. Lawton | 1 years ago (raw) | root | parent | reply | flag 
 @07515e85 I don't know that I agree with that. Consumer-level GPT isn't bleeding edge stuff and that's mostly what I'm referring to; there are several deep learning projects running right now that might make you reconsider.