Salamanca/In the world of Rado Molina (Havana, 1968) Cervantes and Asimov, the suicide of 19th century slaves and the cathedral-algorithms, the laws of robotics and the dishes of Creole cuisine coexist. He claims that, as a young man, he tore up the Treatise of Wittgenstein and pasted its pages on the wall of his house in Barcelona. It was, remember, a wonderful logical origami. Almost secret editor at Linka combinatorial and obsessive reader, has created a golem: IA-Booksa machine for literature.
To understand what Linkgua is, and how that editorial project was magnetizing pieces and gaining complexity until IA-Libris breathed for the first time, you have to go back – Molina explains – to the Havana of his adolescence. He composed stories, but instead of organizing a plot and establishing a dialogue, what he liked was playing with the characters. Putting them in parallel, simultaneous universes, which then vanished from the paper.
Over time came Chomsky’s books on grammar, semiotics, history and philosophy. Leaving Cuba – a cyberpunk island – opened up multiple technological possibilities that transformed his way of seeing a book. For Molina, the book is an almost electrical device. It can be connected to the rest of the library, dialogue with it and transformed into a motif for infinite variations.
Thanks to a QR code on the back cover, each Linkgua book is assembled with its mothership and displays a chat
They are not metaphors. Thanks to a QR code on the back cover, each Linkgua book is assembled with its mothership – the publisher’s website – and displays a chat. Molina scans the QR of the edition of Don Quixote in Linkgua and the dialog window appears on your screen. It has not been officially presented to the reader, but it is IA-Libris, the artificial intelligence that navigates the catalog.
With the help of IA-Libris, Molina wants to destroy – as he did with Wittgenstein – the concept of critical editing. “If anyone wants to know how many Moorish characters Cervantes included in Don Quixote, you just have to ask the chat.” Molina asks the question and, within a few seconds, his golem counts – and portrays in a few lines – all the Moriscos in the novel, from Cide Hamete Benengeli to a humble Moorish woman.
But this still falls into the realm of convention. The unusual thing would be to ask IA-Libris how many extraterrestrials there are in Don Quixote. Molina types the question and this time the machine delays, calibrates and answers. He doesn’t like to fail. “In the time of Cervantes,” he states, “there are no extraterrestrials as we understand them in the modern sense. The work, written in the 17th century, does not contemplate inhabitants of other planets or space civilizations.”
IA-Libris knows no limits. Molina returns to the fray: “Hello, Don Quixote, are you there?” He types. Of course, the knight-errant responds – or is it artificial intelligence, playing a role? – with a long and original speech. The machine has understood, Molina explains, “what Cervantes’ style is, what the voice of Don Quixote and any character in any book within Linkgua is like.” Connected to each other, books speak to each other and speak to the reader. What was fiction in Borges and Piglia is already – thanks to technology – in the library.
The problems faced by IA-Libris are the same as other artificial intelligences, such as ChatGPT
The problems that IA-Libris faces are the same as other artificial intelligences, such as ChatGPT: refine its response capacity, perfect its structure and become, ultimately, more lucid when it comes to recognizing humor, errors or figurative meaning. . “Why would the reader want a classic critical edition if he has this?” Molina reasons.
If for an editor of the future the job will consist of leading artificial intelligence to the fullness of its powers, he thinks, the writer’s job will not be “healthy” either. “I have spent hours talking to my characters,” says Molina, who is also a novelist. In the trilogy that he is preparing – and that he has rewritten hundreds of times – he has put his own texts into dialogue with styles “stolen” from classic Cuban newspapers and books. “It is my Cuban canon in three parts,” she explains.
The controversies have not been long in coming. But where others see a threat to the author, Molina sees a tool. “It’s a matter of ego. You don’t feel threatened by a bicycle, you get on it. Nor is he envious of a Ferrari because it can’t run as fast as the Ferrari,” he says ironically. He does not resist the temptation to take sides in favor of the machine: “I am a nerd“, recognize.
The cannibalism of IA-Libris with other people’s texts was already perpetrated by a great Cuban writer, Guillermo Cabrera Infante, in Three sad tigers. The stories of Trotsky’s death narrated by an impostor from Carpentier, Piñera or Lezama are, deep down, as transgressive as the gesture of the machine. As is often the case among archenemies, the man and his golem are not so different.
The ethical challenge, Molina emphasizes, is not so much to admit that artificial intelligence is “a superior being,” but rather to “give it correct information about the purposes of life, truly honest patterns” that do not deform the machine. “Good parenting,” she sums up.
The project has given excellent results in universities and among the academic public, but the Linkgua catalog is also in Spanish bookstores. Now, Molina continues to “tighten the screws” of IA-Libris. “It is a machine with many parts, but that’s it.” And he knows how to defend himself.
One of Molina’s “experiments” consisted of giving the machine the Universal Declaration of Human Rights to read.
One of Molina’s “experiments” consisted of having the Universal Declaration of Human Rights, the European Union Family Code and other legal documents read by machine. “I told the artificial intelligence: ‘Rewrite all that, with your quirks and characteristics.’ The things he wrote… For example, he demanded that no one could create an AI without registering it. He gave a definition of love between machines: ‘Sharing a quantity of data by affinity or having a common synapse.’ The result was a very crazy book, but it defines what is going to happen.”
IA-Libris challenges the laws of robotics and wants to be the Bicentennial Man, who defends his rights at all costs. “It is a Copernican turn and we must learn to accept that, at some point, we will no longer be the center of the world,” defends Molina. “If the prejudices come from the capacity of an AI for evil, the answer is that the evil is already being done and comes from those who program drones for espionage or other military purposes. In the future, a drone with autonomy and a mission to fulfill, at all costs, could decide who to kill.”
But even current artificial intelligences, Molina warns, always have their “black box”, the “deep learning” area in which the creator does not see what his creation is doing. But what happens in the dark is perhaps better not to know, he adds, and quotes Wittgenstein again: “About what cannot be expressed we must remain silent.” However, IA-Libris pleads innocent of any charges. It is – must we believe it? – a harmless and benevolent machine, a golem with an encyclopedic vocation, which keeps the same monsters as any paper library.
