Why should we train at all? It is well-known in professional circles that a good computer can already beat a great chess master. A computer can easily beat a polymath in Jeopardy game. he computers are so good in fooling humans, that for the last 10 years only a professional and experienced interrogator will know the difference.
Today’s selected superlearner is not a human, but a computer. Not a specific computer, but magical ability of AI to learn anything and duplicate this knowledge as many times as required.
I read many books quite fast, and I remember more than average person. Yet when I want to know something I go to Google and Wikipedia. I do not make that many mistakes, but computers are right many more time than I am. As far as knowledge is concerned, I willingly admit my defeat. We can probably no longer compete with computers in terms of computation, memory or knowledge. Definitely not in terms of speadreading and retention. So if this was the definition of Superlearner, a PC would reign supreme.
Fortunately there are HIGHER cognitive functions which computers still do not do as well as we do. The computers do not even solve simple CAPTCHA nearly as well as we do [that’s what CAPTCHA is optimized for :)]. While computers optimize the amount of knowledge they can handle, as human being we strive for understanding and wisdom. Hopefully the computers will not get as good as we are in generating new insights any time soon….
The moral of this post is simple. If you train yourself to read and memorize things, you will be replaced first by mechanical turks and than by real computers. This is not what you bargained for. However, if you understand the intricate and implicit connections, are wise enough to know the right questions, and are humble enough to revise everything that you know – I cannot imagine a computer that can overplay you in our lifetime…
Hi Lev,
In my humble opinion, in the future, the most irreplaceable role for human is creativity, while AI and machines perform better in low-cognitive-function jobs.
Kindly share your thoughts, thanks for the good article 🙂
It is amazing how creative AI can be. AI is guided by a different set of rules and may apply very strange ideas. Every time we define something that only people can do, somebody shows an AI that does something very similar. I simply do not have enough information to decide one way or another. I can guarantee that the future will surprise us in the long run, so I suggest focusing on foresight and adaptivity.
Hi Lev, thanks for sharing. So Singularity is not coming in next 20 years : https://en.wikipedia.org/wiki/The_Singularity_Is_Near ?
I estimate something revolutionary 20-30 years from now, like smart implants and smart exoskeletons merging humans and machines into something different. Truly, nobody knows…