tlön, uqbar, orbis tertius is one of borges’ most celebrated short stories, and for good reason: the sheer density of novel and thought-provoking concepts borges manages to lay out in just a dozen pages is extraordinary. for deep memetics researchers, of particular interest is the story’s final section, the so-called “postscript”. in it, the narrator relates how an ambitious secret society of intellectuals called “orbis tertius” puts together a sprawling encyclopedia of a fictional world named “tlön”, with the intent of creating a new world. then, they try to meme it into existence, basically, through an ARG/viral marketing campaign. tlönian artifacts suddenly appear (or are planted) throughout the world, a journalist discovers (or is fed) the encyclopedia and publishes it, and a world still reeling from the chaos of war becomes enraptured with the order of tlön. children in schools are taught the language and history of tlön as though it is their own, entire fields of study are reformed in line with tlönic thought, and the narrator laments as the world is reshaped in the image of tlön.

a thought: what if the internet is our encyclopedia of tlön? this is the question i finally arrived at recently while considering twitter. twitter is fascinating because it seems to punch far above its weight in terms of political and cultural influence, despite the fact that it has a fraction of the userbase of other sites like facebook, instagram, tiktok, youtube etc. how is this possible? there is a pretty simple explanation: almost everybody who works in media is a prolific twitter user, especially journalists. from what i have read, there is a large portion of twitter that has basically become a social club for journalists and wannabe journalists, so if you want to make it in the industry it is practically a requirement to use it for networking.

but inevitably, hanging out on twitter all day means that what journalists see on twitter will have an influence on what they report on. this may not be entirely a bad thing, because it just so happens that they share the platform with many “newsworthy” individuals: politicians, celebrities, “intellectuals”, the kinds of people who might have scoops or be scoops themselves. there’s a sinister side as well, however, because of course there’s hordes of political extremists and miscellaneous trolls/schizos sharing the site too. this is where journalists can easily stumble into a classic trap, believing that the map (twitter) is the territory (america, the world, the people, the “real world”, etc.). so in their peripheral vision (the trending tab, perhaps) they see all these schizos and extremists getting worked up over some insane issue on twitter and assume “this must be something important that people are really concerned about”. they forget that objectively, twitter users are a small minority outside twitter, and there’s probably some sort of self-selection bias at work too that makes them more likely to be deranged than your average person.

then, the journalists decide that said insane issue is worth writing an article about, since it must be important because it’s trending on twitter. the article about the insane issue gets published in the mainstream media, where normies see it and spread it around, sharing it on facebook and the like. it must be a real issue and a Big Deal, they reason, because the news is supposed to publish stuff that’s a Big Deal. thus they come to believe it is a Big Deal and now everyone else is seeing it and talking about it and arguing about it and now politicians are making speeches about it because it’s clearly something that people care about. in that manner it really does become a Big Deal.

you may have heard of all the recent research showing that political views are becoming more extreme, more polarized. originally, extreme political views were a defining feature of the online world, of twitter, of tlön. one cannot help but wonder, is our world becoming tlön?

when you gaze into the abyss…

let us consider another example: the much-maligned youtube “algorithm”. like nearly every “machine learning” model, it can be thought of as a black box (or perhaps an abyss). it takes as input all of your viewing data and other metrics (that many are kind enough to provide), performs inscrutable operations on vast quantities of numbers in matrices (the “model), and then spits out recommendations. sure, theoretically you can peek in and see all those numbers in their neat matrix rows and columns, but what each one signifies (if anything) is completely incomprehensible, so it might as well be a black box. it is not at all like looking at the source code for a program.

we call it the model, though, because presumably that is what it is doing: trying to model you, with all those numbers. you can think of that model as your tlönic doppelganger. another “map/territory” distinction must be made here: you’re the territory and the model is the map. thus, it is neccessarily going to be imperfect, there will be gaps, were you to actually meet your tlönic doppelganger they would be you but somehow off, uncanny.

the issue is, when the “algorithm” recommends you videos, it’s actually recommending videos not for you, but for your tlönic doppelganger. as you watch these videos meant for your tlönic doppelganger, something happens - you find yourself inadvertantly changing, becoming your tlönic doppelganger - …the abyss gazes back.

of course the algorithm will pick up on these changes, change its model of your in turn, and now give you recommendations based off your new tlönic doppelganger, and the hyperreal feedback loop continues - what results after many iterations is something that’s been observed and decried by many as “the youtube algorithm is radicalizing people!”. a similar effect to what occurs with twitter, just individualized and fully automated. if the world is becoming tlön, then there must also be tlönic people to dwell in it.