Quote:
|
I fully expect that we will have true self aware machines in the next 2 years.
|
The Turing test is bad at predicting if an AI will become a SKYNET: Gpt3 could be able to pass the Turing test, but it would be incapable to DECIDE that humanity should be exterminated. If you do not feed some text first, it will not do anything.
Your #2 was vers interesting to read. I personaly think that the real danger will come when an AI will be able to think: fix ils ont objectives, decide and act on it. And all I am saying is that none of the models I know can do that. But hey, there might be one in à classified military facility somewhere to prove me wrong...