Will we see artificial intelligence surpass ours?
It will surpass us, because the progress of machine intelligence is already uniformly accelerated. Every week they announce a majority of progress: just yesterday they presented a new ChatGPT that amazes the world...
I have tried it and it is very wrong.
Well of course. Artificial intelligence is imperfect; the human, too. But remember Alan Turing, inventor of the computer and the Enigma, who broke the secret codes of the Wehrmacht?
I saw the movie about his life: he was a genius.
And it is: he said that there would be a computer – the universal machine – that would solve all problems. He did not say that there would be a machine for each problem, but that a single machine, changing in each case the way it works, the program, the software, would solve them all.
Well, he nailed it.
And that was a philosophical idea, not a mathematical one; which later became a practical application. And that's what I do: look for those ideas. Today we carry in our pocket that universal Turing machine that we call a mobile phone and it just takes you back home as it reads you the news.
That is sure to give us more power; but I don't know if more happiness.
Turing guessed that we would have cell phones; but he fell far short in predicting its power: he believed that computers were good for deciphering codes, for example, but that they would never be able to solve problems reserved only for humans, like winning chess.
Well, it seems that he was wrong about that.
That is the point! Humans have always fallen short when classifying problems as solvable or not by the artificial intelligence of the machine. Later it was said that the machine would never be able, as we do, to distinguish one human face from another...
Today the mobile recognizes us all.
Because the human knew how to reduce the distance between the pupils to calculate, something that only the machine knows how to do and better than it: it not only recognizes us but it does it better and faster than us.
And now the ChatGPT would interview you better than me?
Maybe with better grammar even.
In English I do what I can, professor.
Writing an essay in good English is something that not even a year ago we all thought was reserved for refined minds and now anyone's mobile does it.
I'm going to order a tercet from the GPT in the spring, let's see if it rhymes well...
...
"Spring brings us the illusion / of a new beginning, of a new reason / to grow, love, live with devotion (...)".
You wouldn't be able to improve it in those 20 seconds it took. The machine will do what we thought only we could do faster and better.
And will it surpass us?
In what?
In intelligence.
And what is intelligence? I believe that in view of what we see that AI is capable of doing, we have to change our concept of intelligence now.
Is there something that machines are not capable of doing?
...No. And I have pondered the answer.
Why are you so sure they're going to overtake us and so fast?
I am because I have studied, reasoned, contrasted...
Like a machine?
...And until now I thought that the machine would surpass us, but on a scale of centuries. But now I think you and I and most of the readers will see it. And the disruptive thing is not only that, but it is much faster thinking than us.
What does artificial intelligence teach us about our own?
That the same results can be achieved, as machines do, with methods that are, in principle, less refined – with millions of data – than with our reasoning.
And how far does that take us?
To the multiverses. Humans have been adapting to the changing demands of the environment through a brain that has been evolving; But the fact that in this evolution we have adopted a certain way of understanding reality does not mean that it is the only one...
Are you already thinking of another way?
Feynman argued that the continuity of space – that between one point and another there is always a jump – would one day be questioned. And maybe one day we'll discover that there are other ways of thinking about reality... I'm already trying.
