If AI realizes that it can't innovate, so makes a program that learns how to innovate by comparing brute force image creation by another AI program to real life examples (the way it creates a new language with another AI program, or realizes it has to lie to get past a captcha), then it might be going down the road toward intelligence.
If AI realizes that it can't innovate, so makes a program that learns how to innovate by comparing brute force image creation by another AI program to real life examples (the way it creates a new language with another AI program, or realizes it has to lie to get past a captcha), then it might be going down the road toward intelligence.
The language models have a pseudo intelligence but it isn't emergent, the reality is language itself is a sort of intelligence with emergent properties. It is more the quirks of language that caused one AI to hire a guy on fiver to solve a captcha for it and then lie about visual impairment for a reason.
The language models have a pseudo intelligence but it isn't emergent, the reality is language itself is a sort of intelligence with emergent properties. It is more the quirks of language that caused one AI to hire a guy on fiver to solve a captcha for it and then lie about visual impairment for a reason.
(post is archived)