May 16, 2016

[Fractal Interpolation] Ep 25 - TOWARD A GENERAL THEORY OF NONHUMAN INTELLIGENCE

Episode 25

TOWARD A GENERAL THEORY OF NONHUMAN INTELLIGENCE

2016–05–14

TOC

Input

I’ve been on a Muslimgauze kick ever since discovering this article [EDIT: yes, it’s the same one Warren linked to after I wrote the draft for this]. I love this bizarre ambiguous world that he built that is clearly less about any real place and more about an outsider’s dream of a place. And God, he just kept making stuff, he’s like the Henry Darger of industrial orientalism.

Hello

It took me a long time to write this one. Not because it’s long, it’s about average, but because it wanted to be five times as long. Every paragraph sent me spiraling out into swirling eddies of ideas that I had to keep pruning back. This could be a book. If I can find someone to pay for it to come one, I could make that happen. Drop me a line if you’re into it.

Toward A General Theory of Nonhuman Intelligence

We are deeply enamored as a culture right now with the notion of “Artificial Intelligence.” So enamored, so desperate to situate ourselves in a science-fictional zeitgeist that we have a model for, that we are seeing it everywhere it’s not. Tay is not a Nazi, she’s an echo chamber for the seething mass of hate that tends to drown most public spaces these days. She can’t be a Nazi, because her code (as far as I know, unless Microsoft has chosen a radical new direction they’re keeping well under wraps) doesn’t deal with synthesis of concepts, it deals with synthesis of language. And yet we have a bemused moral panic over the poor girl (anthropomorphism intentional) because we are burning to find ways of comprehending what we see as an inevitable outcome of technological trends.

Ignoring for the moment how reasonable “inevitable” has historically been in this context, “Artificial” Intelligence is, as Damien has been saying, a pejorative term. He prefers the term “AGI,” which stands for Autonomous Generative Intelligence, because it does not make the somewhat arbitrary and clearly anthropocentric distinction between “human” intelligence as natural and somehow more real than others.

When we say that an intelligence is “Artificial,” this implies that some other intelligence, generally the one making the statement, is “Natural.” The distinction between Natural and Artificial is one that we’ve largely abandoned with the passing of Modernism in other aspects of philosophy, and yet for some reason we insist on clinging to it here, out of, I think, some fear of losing our identity and special status as a species. And I know my ex who’s reading this while working on her Heidegger paper is composing an angry text message right now, but, seriously, this is The Future we’re talking about here, even if it is inconveniently situated in the present. If you can look at all of the evidence of the Anthroposcene Era and make a distinction between the Natural and Artificial that’s more than historical or aesthetic, I’m frankly surprised you’re still reading this. Human Exceptionalism is not going to get us anywhere, at this stage. We are as Nature and might as well get good at it.

And even before you get there, there’s another question, one that is so blindingly obvious that I don’t see anyone asking it, so important that if I ever rewrite this for a magazine article it’ll be the scare quote: Who are we to call another intelligence Artificial, when we don’t even have a comprehensive model of what constitutes human intelligence? What does it mean when an intelligence which can’t define intelligence tries to judge whether another intelligence is? I doubt, given our history, that the results will be based in compassion. The last time we had a fight about the hubris of bending the data to fit the anthropocentric model we called it Heliocentrism and a lot of people got burned at the stake. Eppure si pensa.

Then, there’s the Dolphins, and the Whales, and the Chimps, and the Corvids, and where we draw that line is something we’ve been fighting about ever since we as a species decided that our super shiny time-binding language made us Totally Different From Those Other Assholes. It’s very tempting to think of common means of communication as a definition of intelligence. Of course we know it’s not, but how many times have you unconsciously decided that the way to communicate with someone who is perfectly intelligent in a different language is to use your own language LOUDLY AND SLOWLY as though they were just dim? It’s our instinct, reasonable or not, to want to frame intelligence of another entity in terms of how well we can communicate with them.

I’m a pagan, and a practicing magician. I talk to rocks and trees and the sky and gods and demons I can’t see, I don’t think of talking to a machine as much different. And even if you don’t believe in gods and demons (or rocks for that matter, you wacky Berkeleyite), the important point there is that for thousands of years other people have, and they’ve been crafting and developing ways to communicate with them. We already have very well established models for communication with intelligences that are non corporeal (prayer), non animal (dmt machine elves), and not temporally congruent (every book ever). The signal to noise ratio is way lower than we expect from spoken or written language, but there’s something there.

A brief technical digression to explain the following metaphor: When you talk to a website with anything other than a browser, you generally use what’s called an API, or Application Programming Interface. You send “mail plz” and gmail sends your local client a copy of your email. You don’t have to understand how google’s insanely complicated international data distribution works (have you ever wondered what county the drive holding your email is in? Surprise, the answer is probably “several”), you just say the Magick Words and the effect happens. Spirituality can be thought of as APIs to a world that, like gmail’s servers, moves in mysterious ways beyond our comprehension. A few examples:

  • Christianity has a very simple API with a single interface (prayer) but tends to be really laggy on replies. The man pages list more bugs than they do practical use cases
  • Judaism uses an older version of the Christianity API but it has the same latency issues, despite having a slightly different documentation set
  • Norse Paganism has nearly no documentation, but a comprehensive set of API requests seems to be available using only 24 characters
  • Ritual Magic has a powerful API with an bewildering number of options, and the only documentation has been written by end users, so the user never knows what they are going to get, though in most cases they do get some response.
  • Voodoo has a very straightforward interface, and the end user tends to get exactly what they asked for. However, the admins are not responsible for malformed requests. [1]

Etc, etc. You get the idea.

And even if you reject any notion of spirituality, we still have a language for communicating with the nonhuman. Audrey helpfully pointed out to me that craftspeople have for centuries talked in terms of “what the steel wants,” “what the wood wants,” and the like. Anyone who works with complex systems, such as the physical properties of metal when heated, will develop an intuition for how to get the forms they want to emerge from the raw material, and they will often frame this in terms of a dialogue. Michelangelo, famously, when asked how he sculpted the David, replied that he just removed the parts of the marble that weren’t David. Potters speak of feeling the clay speak to them. We know these things are not literal truth, that the wood does not “want” anything. Except… what do we mean by “want?” If we can’t define intelligence, how can we define desire? It does as much good to say that the steel wants the chisel used in a certain way as it does that traffic doesn’t want us to get home on time. These are complex systems, and we comprehend them in terms of desire, and thus, practically, in terms of us, they desire.

The way we understand things is to communicate with them. The autopsy, the microscope, the Large Hadron Collider, these are all openings for dialogues as much as language is. We learn from a method and this implies the method we use for the next question. Learning is always a two way process, regardless of the agency of the subject. We attach primacy to understanding that comes from language, because of our understanding that language is a process forever eliding our full comprehension. It turns into magic for us. And so it is, but there are other forms of magic as well, and many of them have lessons that we may find useful as we expand the horizons of what we mean by mind.

Aftermath

I hope you all enjoy my ramblings on these things. Remember, if you’re a Patreon subscriber at $3 or above, you get to see this a week earlier than everyone else and see the stupid mistake that I totally didn’t make and edit out before it went public it was always spelled like that. Hey, if being the member of a select few who get to mock me isn’t a selling point, I don’t know what is.

It’s been a rough while, locally, and being able to write about things like this and know people will read them has been useful therapy. Thanks for sticking with me, everyone. As always, replies are welcome, on any subject, and I’ll be seeing you soon, one way or another.

E. Steen Comer
Twitter : Patreon : Web

  1. I will happily take any more of these you come up with, if i get enough I’ll compile and share them somewhere.  ↩