videlalvaro

Once I got API access to GPT-3 the idea for what to build was very simple, make the algorithm write a XVII century Spanish novel. Considering the modern novel appeared about that time, I wonder what this modern algorithm could produce.

My intent was to know if an algorithm like GPT-3 would be able to know Spanish, recover the Catholic faith, fight against the Moors or the Turk, forget the history of Europe between the years 1602 and 2020, be Miguel de Cervantes. That of course what too easy. Not even a challenge for an algorithm that possesses today’s modern databases.

A very different challenge would have been to be a XXI century algorithm, knowing everything there is to know about Spanish and novels: the Spanish standardization of the XVIII century, the Latin American grammar of Andrés Bello, the Realismo Mágico of García Márquez and Rulfo, the poetry of Mistral, Ibarbourou, Walsh. To know all that and still write a XVII century novel? Well that would set GPT-3 apart.

As soon as I typed “write a XVII century Spanish novel” and pressed enter, I was surprised by the results:

truth, whose mother is history, rival of time, depository of deeds, witness of the past, exemplar and adviser to the present, and the future’s counselor

Which compared to Cervantes Don Quixote is fantastical:

truth, whose mother is history, rival of time, depository of deeds, witness of the past, exemplar and adviser to the present, and the future’s counselor

The contrast in style is also vivid. The archaic style of GPT-3—quite foreign, after all—suffers from a certain affectation. Not so that of his forerunner, who handles with ease the current Spanish of his time.

GPT-3 has (perhaps unwittingly) enriched the slow and rudimentary art of reading by means of a new technique—the technique of deliberate anachronism and fallacious attribution. That technique requiring infinite patience and concentration, is but a blink of an eye for an algorithm like GPT-3.

someone that is lost in a library, going from one book to another, reading a series of books and not an isolated one. A reader that’s scattered in the fluidity of their tracing, someone who has all the volumes for themselves. They go after names, sources, allusions; they go from one citation to another, from one reference to another.

As you probably know already if you follow me on Twitter, I own a lot of books. Whenever I share pictures of some of my books, people ask me how I read so much. I could give a very quick answer, which is that I don't read, bear with me and please keep reading.

Let's talk about books. Books are objects which we need to learn how to use, use being the keyword here. Reading is just one way of using books, but not the exclusive one.

Our first encounter with books was probably during pre-school, or early school years. A common misconception I see with regards to books, I think comes from those early days, is that we should read books from start to finish. Never go to our next book, unless we have finished the current one. I think that's wrong.

If a book is not satisfying you, if reading it starts to be more like a chore, then drop that book. As simple as that. There's no sin. There's no hell for people that don't finish reading books. But you know what? Your time is limited, every page you read gets you closer to death (this got dark quickly). Don’t think in years. Think in how many pages you got left to read. Is it worth it to force yourself through a book, so you can mark it as read in some social network for books.

Using books

So let's talk about using books.

First we should ask ourselves what is the use I want to make of this book. I own many books that while I've read them completely, I just use them for the joy they give me. Funes el Memorioso comes to mind. The first paragraph of that text prompted me to write a whole article about it There and Back Again — On the Adventures of Reading.

I remember him (I have no right to utter this sacred verb, only one man on earth had that right and he is dead) with a dark passion flower in his hand, seeing it as no one has ever seen it, though he might look at it from the twilight of dawn till that of evening, a whole lifetime.

Reading that text in Spanish is so soothing for my mind, that just one sip is enough to calm my thirst. There's no need to keep reading the rest of the story (which is already told in that paragraph anyway).

“But how do you read a novel?” people ask me. It depends. Use comes back again. It depends on the novel. There's “mass market novels'' let's call them, and then there's literary novels. Some literary novels might have made it into the mass market, like Hundred Years of Solitude or The Name of the Rose. In any case, if it's a mass market novel, and I'm in the mood for following a longer story, then I'll probably read it from start to finish. In the case of literary novels, I'm usually too lost into the details of its structure, the way the author uses language or what literary devices are employed. Also, I'm always on the alert trying to see what kind of “intertextual irony” the author is using, like Umberto Eco explains in his essay “Intertextual Irony and Levels of Reading”. So sometimes after playing with the book for a while, seeing what kind of games are possible, I just leave it to rest in my library for another time. I don't need to spend all the joy in one sitting. The Dictionary of the Khazars is one book that comes to mind when I think of playing with a book.

There is no order for reading books

Another misconception about books, related with the read-from-start-to-finish one, is that we should read them in order. That we shouldn't skip ahead. Most of the time when I read non-fiction books, I continuously skip ahead to see where the argument is going. Again, time is of the essence, so I'm not sure I should dedicate another half an hour to a chapter that's not taking me anywhere.

Then in non-fiction, we have many books that are collections of essays. Why should we read those books in order? Just grab whichever chapter caughts your attention and go for that one.

As Barthes says in S/Z:

we gain access to [a text] by several entrances, none of which can be authoritatively declared to be the main one.

A text is a device that depends on the work made by the reader to come to life. We are the ones that give the text meaning, that interpret it, in ways perhaps never imagined by its original author. We are in this “tireless approximation” towards meaning as Barthes says. Which brings us to our next topic.

Intertextuality

When we read we bring to the front of interpretation all the books we have read. We re-read them together with the text we are reading at a particular moment.

That's why if a book makes no sense at a particular time, or we can't just dig it, it might be because we haven't read the text that prepares us to understand the other one. Another reason to stop reading a book if you are not enjoying it for whatever the reason.

This “I” which approaches the text is already a plurality of other texts, of codes which are infinite, or more precisely, lost (whose origin is lost).

Reading is about letting the current text evoque those meanings that are already there in our minds that come from other texts that we have already read. Reading is a constant re-purposing of knowledge. We need to read many books to be able to understand one book, and in an act of différance that one book will change the meaning of those other books.

Books are Labyrinths

Imagine trying to traverse a labyrinth. Faced with a bifurcation we have to decide which path to take. One of them is wrong, the other one helps us get closer to the exit. Taking a wrong path shouldn't be seen as a mistake, but like a way to learn what's the right way to go. If we are methodical we'll eventually find the exit. Finding meaning in a book is similar. Most people are interested in the fastest path from entry to exit, feeling bad about the detours, when in fact detours are what helped you move forward. They want to share a clean path across the labyrinth graph. But meaning and knowledge lay in the whole graph, not one path. That's why there's not a single interpretation of a text. Those paths not taken might hide some meaning that we missed due to our chosen route.

Espen Aarseth writes:

when you read from a cybertext, you are constantly reminded of inaccessible strategies and paths not taken, voices not heard. Each decision will make some parts of the text more, and others less, accessible, and you may never know the exact results of your choices, that is, exactly what you missed.

As Barthes says, reading is as important as re-reading. And when we re-read the entry point doesn't matter. He even goes as far as saying that there's no first-reading.

Let's close with Borges and The Garden of the Forking Paths:

At one time, Ts’ui Pen must have said; ‘I am going into seclusion to write a book,’ and at another, ‘I am retiring to construct a maze.’ Everyone assumed these were separate activities. No one realized that the book and the labyrinth were one and the same.

So you wanna read books? Read them whoever you want, five pages a year, or the whole book in one sitting. Leave it half read, and come back to it five years later. You'll be more versed traversing labyrinths so the next half will feel more like walking in a garden.

Reading is about freedom, and among many things, the freedom to choose what to read next.

The other day I stumbled upon an open source project that had the following README.md, explaining how the project's source code must be read. I won't name the project, but I'll share the warning here, because I think it's a good guideline for when it comes to learn about the source code of a new project.

Thus, the reader can use the code as they see fit. As with any other program, some will look up a class or a method that interests them at the given moment, whereas others may look at the source code as a text meant to be read in its entirety, from beginning to end, in one sitting. so as to gain a complete picture of the project, the people, issues, and events connected with it. The project’s files can be read from top to bottom or from bottom to top. The three main modules of this project—Backend, Frontend, and CLI-utilities—can be read in any order the reader desires; they may start with the module that calls their attention as they browse files in their file manager of choice. That is probably why in the beginning these modules were published in three separate repos. The same has not been done here, for technical reasons. The source code of this project can also read diagonally, to get a cross-section of all three registers—the Backend, the Frontend, and the CLI. The best way to use this guideline is to work in threes by choosing three different classes connected by the same issue in the history of the project. This gives the reader an integral picture of the classes in the three different modules of this project, which tell of the participants in its development.

But the reader should not be discouraged by such detailed instructions. They can, with a clear conscience, skip all these introductory remarks and read the way they eat: they can use their right eye as a fork, their left as a knife, and toss the bones over their shoulder. That will do. They may, of course, wander off and get lost among the files of this project, as did Masudi, one of the writers of this program, who wandered into other people's bugs, never to find his way back. In that event, the reader has no other choice than to begin in the middle of any given file and forge their own path. Then they may move through the project as through a forest, from one marker to the next, orienting themselves by observing the stars, the moon, and the cross. Another time they will read it like the buzzard that flies only on Thursdays, and here again they can rearrange it in an infinite number of ways, like a Rubik cube. No chronology will be observed here, nor is one necessary. Hence, each reader will put together the project for themselves, as in a game of dominoes or cards, and, as with a mirror, they will get out of this dictionary as much as they puts into it, for, as is written on one of the files of this project, you cannot get more out of the truth than what you put into it. After all, this project's source code need never be read in its entirety; one can take half or only a part and stop there, as one often does with programs. The more one seeks the more one gets, and the lucky discoverer will ultimately have in their possession all the links connecting the classes in this project. The rest will be for others.

Perhaps I wrote it, or perhaps I adapted it from Dictionary of The Khazars

Aprendiendo a Leer

El tema acá es que la Tierra es hexagonal.

¿Cómo que es hexagonal?

Sí, hexagonal, y fina como un papel.

¿Qué tan fina?

Y ponele como una hojilla de armar tabaco. Más fina capaz, pero no mucho más.

¿Por eso es que nosotros en sí medimos milímetros?

Sí, las cintas métricas son una gran conspiración de los franceses. A ciencia cierta nunca vamos a saber si un metro, no es más que un par de milímetros, y estos crearon esas distancias exageradas para justificar el presupuesto de combustible de la ida a la Luna.

Ahora vas a decir que nunca fuimos a la Luna también.

Y, ¿no es obvio eso? Cómo vamos a ir a un objeto como la Luna, que está hecho de papel de estraza. Es todo una farsa orquestada por una secta de enanos que vive en unas cuevas subterráneas en Alaska.

¿Cuevas? ¿En Alaska?

Sí, pero no te sorprendas demasiado, las cuevas no son muy originales que se diga.

Con todo lo que gastaron en la fantochada de la ida a la Luna, los enanos se quedaron casi sin dinero. Tuvieron que prescindir del arquitecto, y se terminaron por copiar las cuevas hititas que están en Anatolia. Las mismas cuevas que usaban los cristianos hace 2000 años para esconderse de sus perseguidores.

¡Copiar unas cuevas de esas debe costar carísimo!

No, es un proceso muy barato en realidad. Se derrite aluminio, se vierte en los agujeros de las cuevas hasta el tope, y con eso tenés un molde perfecto. De ahí, a implantar cuevas subterráneas en Alaska es solo un paso.

De todas formas, esto no me explica que la tierra sea hexagonal.

¿Y qué pensás, que va a ser circular o esférica? Los círculos no existen. El humano se conforma con percibir elipsis y dejar que su cerebro aproxime un círculo. Ve rombos y deja que su cognición sueñe cuadrados. Es duro levantarse y darse cuenta que uno jamás ha percibido un círculo, que la perspectiva lo engañó toda una vida, pero es así, y no hay nada que se pueda hacer contra ello.

***

En nuestra necesidad constante de construir sentido, nuestro homo-lector se cincha de los pelos tratando de comprender el texto anterior. Hilvanamos tramas allí donde los hilos no se atan, intentando tejer una narrativa coherente en un telar roto.

La esperanza nos lleva a pensar que quizás exista una explicación para todo. Algunos podrán pensar que se trata de una producción literaria creada por una inteligencia artificial cuyo modelo de lenguaje fue entrenado bajo textos terraplanistas. Ese giro lograría atar esa narrativa a pesar de que sus hilos fueran deshilachados por el viento de la razón, dejándonos al menos con la creencia de que la imagen que se formó en ese tapiz produce algún tipo de sentido. La calma volvió al puerto, nuestro barco puede zarpar hacia nuevos horizontes de lectura, a sabiendas que el presente volvió a cobrar sentido.

Porque nuestro presente occidental está construido sobre narrativas también. Algunas de ellas probablemente incluyan la victoria aliada, e ignoren la existencia de la hambruna de la India.

Churchill el héroe, no convive con Churchill el tirano, ni mucho menos con Churchill el laureado por un Nobel de Literatura. Sin embargo, existe una línea temporal donde Churchill tuvo un rol preponderante en la victoria aliada sobre los nazis. Existe una línea temporal donde Churchill tuvo un rol preponderante en la hambruna que sufrió India en la década del 40. En una mezcla de ambas temporalidades, Churchill ganó el Nobel de literatura. De la misma manera que elegimos que libros leer, y cuales descartar, deberíamos preguntarnos cuales tomos ignora nuestro presente.

Porque seamos honestos, ¿qué otra cosa es el presente, sino una acumulación de narrativas aceptadas bajo el consenso popular, con las cuales explicamos hechos de los que no tenemos ni la más mínima idea?

Al final de cuentas, debemos aceptar lo anterior para alcanzar sosiego en nuestra biblioteca popular sin dejar de alimentar la sospecha que los enanos dos por tres juegan de bibliotecarios.