Sosa Escudero is the author of “Borges, Big Data, and me.”
The literature of Jorge Luis Borges is almost endless because it tends to infinity, because whenever you reread it you find something new. The experience is infinite. And something similar happens with artificial intelligence and big data. Walter Sosa Escudero is a professor at the University of San Andrés and links the Argentine writer with algorithms in his book Borges, Big Data, and me.
He is also an economist and the author of two other books published by Siglo XXI: Big Data and What is (and what is not) statistics? The writer passed by the Leamos-Bajalibros stand at the Buenos Aires Book Fair and spoke with Belén Marinone about infinity, about the endlessness of this novelty called artificial intelligence.
It may interest you: What Netflix should learn from Borges
-What is the difference between Big Data and artificial intelligence?
-In the midst of the revolution, spending a lot of time defining what is artificial intelligence, what is machine learning, what is statistics, big data, is not seeing the most important thing, which is what they have in common. In the long run, the universe of artificial intelligence is greatly affected by the universe of data. This revolution, for example, of ChatGPT has to do with the fact that there is a tremendous amount of text to analyze, which is data, and behind these things that use data there are statistical, algorithmic methods.
-Then it is very difficult to talk about the difference between one world and the other. The Big Data world is all about this big data thing; the world of artificial intelligence, by exploiting some rules. Suppose you put Big Data Machine Learning to learn to play chess. What he’s going to tend to do is look at a bunch of games and see if he can infer rules, optimal roams. The world of artificial intelligence is going to take the little rules of chess and try to exploit them to the extreme. That is, one tries to exploit the rules, the other tries to induce, see if he learns them.
[”Borges, Big Data, y yo” puede comprarse en formato digital en Bajalibros clickeando acá]
-In other words, the invincible Russian chess players could perfectly lose against the machine forever. Now there are no chances of winning.
As a human, no. Precisely these robots, for this type of thing, have a processing capacity that humans do not have. Think of simpler things than playing chess. I can read a book in an afternoon. Well, the Bot can read millions of novels in an afternoon. Read in quotes.
-Where do you get the data from to be able to answer me and to be able to talk to me, for example, ChatGPT?
-Reduced to its essence, ChatGPT is symbols in, symbols out. That is, you put symbols, letters, words, spaces and a punctuation, you press enter and ChatGPT puts symbols. I put “my mom…” and stop. What would we do at the Book Fair? We do a survey: after “my mom me”, what’s next? Does she Love, hit, scrutinize, pamper? And we make a ranking, “pamper” wins: my mom pampers me. Well, ChatGPT does that, but with a huge database. The scope of algorithms is so powerful that I can do much more sophisticated things than “my mom told me…”.
It may interest you: The statistical trick that the police use to estimate how many gangs there are in California
-When interacting with ChatGPT, there is a barrier between what is real and what is virtual or what is not real. The same thing happens in literature when we ask ourselves how much is fiction, how much is reality. How could we relate these worlds?
-The achievement of ChatGPT is that you don’t realize that you have asked a robot or a person because it picks up the idioms. But if I put it to distinguish if this was done by Borges or a common person, ultimately, ChatGPT works for all of us who are not Borges, it does not work for genius If you give it a set of instructions to use an electric coffee maker and You tell him to write an article on how to use an electric coffee maker, the GPT chat is going to do spectacular. Now, if you tell him to write you El Aleph, it’s a much more sophisticated program. In the old days, how did you tell if a student had cheated? Because he wrote weird. In other words, a student who writes in a way that is not consistent with how he behaves. Well, the ChatGPT now behaves like a person. Not like Borges, not like some robotic thing. But as one of us.
Technology companies dispute the developments of Artificial Intelligence. REUTERS/Dado Ruvic/
-Should we be afraid?
-I am 57 years old and I think it is the tenth time that I have been told that I am going to be replaced by a robot. And I am waiting. Obviously this is going to bring about some change, but when you look at the history of humanity, this is neither the first nor the most dramatic time that something like this has happened. When the genetic revolution appeared, at the end of the 1960s, we thought we were going to be surrounded by mutants and the same thing happened that happens now: we wanted to prohibit or intervene. And it took us some time to realize that there are things about genetics that are very good and others that are very bad. The answer is yes, we have to be afraid, we have to be amazed, but no more than what we have done with the advent of agriculture, the printing press, the internet.
It may interest you: How Artificial Intelligence will change Google search results
But Artificial Intelligence is not so infinite, nor so vast, nor so omnipotent. For example, you cannot predict the future: how much the dollar is going to be worth, who is going to be the next President, how a soccer match is going to turn out.
“It is not a problem of data and algorithms, nor of economists, nor of political scientists,” explained Sosa Escudero, “it is that the social thing is strategic, it is complex, it is interactive. The machine should learn what we realized long ago: that certain phenomena are essentially unpredictable.”
And, of course, that Artificial Intelligence cannot write like Borges.
How Artificial Intelligence Will Change Google Search Results What Netflix Should Learn From Borges