June 21, 2024

A final regular post for the year, with some new things and a bit of catch-up. New things first:

“How Citizen Surveillance Ate San Francisco”, by Lauren Smiley

It’s not hard to understand why people in a city where crime, homelessness, public drug use, and other urban problems are rapidly increasing would want video evidence, whether to show the police or simply to prove that the things they describe are real. But pervasive surveillance has its own consequences, which aren’t necessarily predictable.

Which is to say: Anything “caught” on camera is actually, on some level, set loose—often into realms scarcely envisioned by whoever filmed it…Video may seem like a means of control, but it can just as easily unleash chaos.



“Inside the Weird and Wonderful World of Miniatures”, by Scott Huler

A couple of years ago when staying in a hotel, we came across a competitive pogo-stick-riding event on TV. People on very tall pogo sticks (the ones used in competition are much larger than the ones people might have had at home) bounced around a kind of obstacle course, using steps and ramps and things to do flips and other tricks. It was pretty impressive, but the thing that really struck me about it was the sense that this is a thing— that there were clearly established standards and lingo and key players already firmly in place, and a community of people formed around this activity that I didn’t know existed.

This article is a little like that. I’ve seen miniature rooms before— I’ve been through the Thorne Rooms at the Art Institute of Chicago, which are mentioned here, many times. But I guess I thought of them as mainly for museums, as a method of illustrating different historical periods; I hadn’t given much thought to people making them for fun, although in retrospect it is not really surprising. I’d certainly never thought about miniaturist conventions or gallery shows, all of which exist and are growing in popularity. Anyway, the piece describes the work of masters and more casual hobbyists alike, and there are lots of great pictures of their work, which is meticulous and varied, running the gamut from historically precise gothic houses to grimy club bathrooms…and Jeffrey Dahmer’s refrigerator.



“Holly Herndon’s Infinite Art”, by Anna Wiener

I’m a big fan of Holly Herndon’s music, but even if it’s not your thing, she is someone worth paying attention to for the way she (and her husband, Matt Dryhurst) are thinking about the impact of artificial intelligence on artists. Rather than simply rejecting it, they have embraced the possibilities it presents, while trying to build new tools and platforms for artists to make use of it while also being fairly compensated. This piece is sort of a brief profile, but mainly a look at some of the ways they are doing this.



“The Hoffman Wobble”, by Ben Lerner

This is difficult to describe; it is (apparently) a mostly fictional account of someone trying to alter Wikipedia entries to counter certain ideological ways of framing news stories. It is based, to some extent, on something that Ben Lerner did in the past, but the specific facts are entirely, or almost entirely, made up. It is impossible to be sure about any of this, which the the point; the piece constantly and explicitly undermines its own credibility. It is also, in some ways, the best analysis I have read of how the way that information circulates online is corrosive of truth and trust and certainty.

For in the information wars, the culture wars, isn’t it fiction all the way down? To say you “believe” a particular fact is to signal membership in a tribe, to dispute the “facts” about, say, there being a video of Hillary Clinton and Huma Abedin drinking the blood of a child whose face they then skin and take turns wearing as a mask is to bring the wrong weapon to the fight.



I also wanted to include some articles and things that I missed from earlier in the year. Sometimes, these are things I meant to post about and somehow forgot; in other cases, there things that for whatever reason are only coming to my attention now.

First, the producer JLin apparently released a new EP back in September, featuring work from her collaboration with the ensemble Third Coast Percussion. That’s a combination that makes a lot of sense, since the drums are always one of the best things about her work.



“ChatGPT Is a Blurry JPEG of the Web”, by Ted Chiang

This is an analogy that may make no sense at first glance, but once Chiang explains it, you realize both how perfect it is, and how revealing. It explains, though not in a technical sense, why large language models like ChatGPT make the kinds of mistakes they do, and why they so often “hallucinate” or fabricate information: in essence, they are providing a compressed reproduction of the information they were originally trained on. The analogy also, counterintuitively, explains part of what seems so remarkable, almost magical, about these systems: that they seem to have some kind of understanding.

I think there’s a simpler explanation. Imagine what it would look like if ChatGPT were a lossless algorithm. If that were the case, it would always answer questions by providing a verbatim quote from a relevant Web page. We would probably regard the software as only a slight improvement over a conventional search engine, and be less impressed by it. The fact that ChatGPT rephrases material from the Web instead of quoting it word for word makes it seem like a student expressing ideas in her own words, rather than simply regurgitating what she’s read; it creates the illusion that ChatGPT understands the material. In human students, rote memorization isn’t an indicator of genuine learning, so ChatGPT’s inability to produce exact quotes from Web pages is precisely what makes us think that it has learned something. When we’re dealing with sequences of words, lossy compression looks smarter than lossless compression.



Early Computer Art in the 50s and 60s, by Amy Goodchild

Exactly what it says on the tin, this is a fascinating overview of the earliest efforts to use computers to make art. While, as you’d probably guess, these artists were certainly working with more limited technology than people today, the limits aren’t necessarily— or at least no only— the ones you’d expect. It’s not just a matter of memory and processing power; it’s also that early computers simply lacked the tools for input and output that we are now used to. Instead of of moving a mouse to draw a line, which immediately appears on a screen, these artists were entering code, and then waiting at a plotter to see what that code got translated into, visually. One result of this is that a lot of these people were interested in computers or math or engineering first, and from there ended up investigating how they might be used to make art; many of them worked at places like Bell Labs. Similarly, early pioneers in electronic music, like Karlheinz Stockhausen, were effectively doing electrical engineering as much as composing. The first synthesizers didn’t look like pianos, but like a bunch of knobs and wires soldered together on a board; fundamentally, they were just oscillators producing a current, and then a bunch of circuits and resistors to modulate that current or introduce overtones before it was translated by a speaker into audible sound. (Really, that is still what a synthesizer is; all of those elements are now just packaged in ways that make them look more like “regular” musical instruments). Goodchild describes one of the artists saying that

while we can use the computer to simulate traditional artistic methods, it is more interesting to use the computer in entirely new ways, developing its own aesthetic language instead of using it to simulate physical media.

The thing is that for a lot of the early artists, this was really the only option. The technology could simulate traditional methods poorly, if at all, so using them at all meant coming up with new ways of doing things. That in turn meant that the results would also look like something new and different.

It’s also striking that these artists, and the people around them, were having debates about automation and creativity that are more or less the same ones we’re having today around ChatGPT and other generative AI systems. At what point does using a machine to alter or facilitate the making of art shade into letting the machine make the art for you? Is following an algorithm incompatible with being creative?



“Hyperdimensional Computing Reimagines Artificial Intelligence”, by Anil Ananthaswamy

I often encounter articles about new or developing technologies that I don’t really understand, but still feel like I see why the idea is new or exciting. Essentially, the idea here is that if you want the kinds of artificial neural networks that systems like ChatGPT are based on to recognize or process more kinds of inputs, you have to make them more complex, and more complexity means more expense, power consumption, and potential points of failure. Hyperscale computing allows you to record and categorize more kinds of inputs without increasing the complexity of the system in the same direct way, by categorizing concepts as strings of numbers representing points in a multidimensional space. This is probably a bad analogy for any number of reasons, but I’m thinking of what they’re describing here as a little like the difference between bitmap/raster graphics and vector graphics, the latter of which can store a much more precise record of an image, despite creating much smaller files that are less prone to error and easier to manipulate. The promise is artificial intelligence that is much more efficient; I guess whether that is good or bad in the long run remains an open question. It would also be much more transparent, eliminating, or at least reducing, the so-called “black box” problem that plagues current AI systems— and that is very definitely a good thing.



“The Internet is for 12-Year-Olds”, by Max Read

I don’t know if I entirely buy the premise here, but it does make a lot of things make more sense.

Because the audience online so wildly over-samples 12-year-olds relative to the population, and because all social platforms work like highly competitive marketplaces, you are constantly being disciplined into creating content that is essentially, though not explicitly, for 12-year-olds.



And that’s it for now. I’ll be doing some kind of end-of-year wrap-up shortly, with my own best-of-year lists. Hope everyone has a happy holiday season.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.