Affect as a Service

Below is a text generator based on the GPT-2 language model. This generator was trained on the Artforum archive's reviews, previews and critics' picks. You can read our documentation on how to build a GPT-2 text generator here.

You can access the generator at any time by clicking on the + symbol in the upper right hand corner of the header.

Below is a conversation that took place on Discord between July 19 and July 20, 2020. The conversation has been edited for clarity. The participants are:

Paige K. Bradley, a writer and former editor of Artforum.

Geoffrey Mak, a writer and contributor to Artforum.

Andrew Thompson, who has not written for Artforum.

Cade Diehm, who has not written for Artforum.

  1. Andrew Thompson on Jul 19th 2020
    • let me give some background on this whole thing
    • I scraped all of Artforum almost a year ago now initially as a way to do some kind of work on the nature of art writing, and awhile later Jules got in touch with me because he had used the somewhat recently released GPT-2 model for text generation on a Pitchfork dataset we had on the site
    • GPT-2 is the model created by OpenAI that was like this "too dangerous to release" Natural Language Processing model that they then decided to release anyway lol
    • It became notable early on for being able to produce uncannily human-like text unlike basically any model that was previously used
    • so we decided to apply the model to the whole Artforum dataset, which stretches back from the first Artforum piece back in what like 1960-something to I think January 2019
    • so that's what the generator on the page is
    • and what I wanted to do was basically display the generator to people both at Artforum and not and get some sense of their reaction to it, to the nature of the texts it's producing, etc
  2. Paige K. Bradley on Jul 19th 2020
    • I have some questions and observations based on the texts it generates and their potential relationship to Artforum's archive, but I'm curious first if you could say more about the initial motivation for this project: to explore the nature of art writing. Did the idea come about as an impulse to see what the GPT-2 would consider art writing? What's your own history with Artforum as a reader, etc. I'm interested not necessarily in asking "do you like/not like Artforum," moreso what is your level of interest/engagement with their material. I do think it's notable that no other publications' outputs were part of this.
    • Does that choice reinforce the industry assumption that what Artforum produces is what writing about art looks like?
    • Does that perception still exist?
  3. Andrew Thompson on Jul 19th 2020
    • we considered using other texts but didn't for some practical reasons, partly that I already had the artforum data to begin with and partly because there are more artforum samples than pretty much any others that are accessible to use as a dataset
  4. Paige K. Bradley on Jul 19th 2020
    • AF has perhaps the most material to work with? Or for the GPT-2 to learn from?
    • It's been around for long enough
  5. Andrew Thompson on Jul 19th 2020
    • yeah its archives go back the furthest
    • and i think that in going back the furthest it has some kind of claim to be fairly representative of an established way of writing about, discussion and even conceiving of what art is
  6. Paige K. Bradley on Jul 19th 2020
    • Right
    • And so did you have particular questions in mind that you wanted to see if the generator would answer, in a sense? Or was it more open-ended, like, "what would that look like?"
    • Reading what the generator produces, did anything stand out to you, were you surprised, or do the results end up looking like something that would be in Artforum to you? Do they feel authentic?
  7. Andrew Thompson on Jul 19th 2020
    • i think that when you force a neural net to absorb a corpus and then spit out what it believes are the probabilities of language of that corpus, you can look at that corpus in a more essential way than directly looking at the text itself
    • they felt strikingly authentic to me
    • the quirks and occassional repetition/errors notwithstanding
  8. Paige K. Bradley on Jul 19th 2020
    • How does the model for generating AF texts get improved? How is it trained?
  9. Andrew Thompson on Jul 19th 2020
    • the short answer is that you basically show it a bunch of examples of an Artforum text, you tell it to produce a text that's like the originals, and the further away it is from the target, the more the model tweaks its own parameters to resemble the original more
    • again the very very short answer
  10. Paige K. Bradley on Jul 19th 2020
    • Interesting.
    • I wonder if a parallel could be drawn between this training, which seems to me like a kind of mimetic process, showing an example and then the manifestation of success/intelligence is reproduction or imitation, with an editorial process.
    • Or, we have idea of what art writing is. Or what Artforum produces, and the value of that. Which is lineage-based.
    • Reading the corpus has implicit instructions on how to write the corpus
    • An editor tweaks a text to hew towards a target
    • Which is their own amalgamated variation, in a sense, on what the AF text is
    • Or should be
  11. Andrew Thompson on Jul 19th 2020
    • so that's actually in a sense the thing i find most ultimately interesting about this whole model
  12. Paige K. Bradley on Jul 19th 2020
    • The writer perhaps is analgous to the learning language
    • The writer must learn what is desired from their text.
  13. Andrew Thompson on Jul 19th 2020
    • which is the way the GPT-2 mimesis of a style itself resembles the mimesis of style employed in the writing of the text itself
  14. Paige K. Bradley on Jul 19th 2020
    • By generating and regenerating.
  15. Andrew Thompson on Jul 19th 2020
    • exactly
    • and to me the GPT model can function as a prompt to ask that very question, which is what it means that the process of producing these texts as a human often follows the same mimetic route as the AI
  16. Paige K. Bradley on Jul 19th 2020
    • Yes.
    • I want to say more about this idea of formula/formats and mimesis/performing and how the subjective or "authentic" voice can operate along these mimetic routes. But also for context/the record/any future reader of this, I think it matters to be clear where I'm speaking from and my own degree of familiarity with all this. I've written for Artforum since 2013, and was an editor there from 2014-2018.
    • However, I came into this actually without "training" per se. In terms of routes, most people who write for AF have degrees in art history, master's level or PhD, and same for staff editors. Or they studied writing/journalism/a related humanities. Generally this means at minimum bachelor's degree. Ok probably a big 'no duh' from anyone on that.
    • I have a BFA in printmaking. I'm an artist. I never studied writing and I came into this essentially self-taught.
    • I learned how to write for Artforum by writing for Artforum. Or for some reason I already knew how without instruction. To the degree that I was sat in a chair and given texts to shape into AF texts, with some familiarity and a competent understanding of the material, but unlike an AI, which is fed everything, presumably, I can tell you I haven't personally read everything that AF has ever produced lol
    • So, here's a question that might emerge: how much information is needed to reproduce the style, ethics, or voice of what's considered a specific type of writing?
    • Or, if GPT-2 is fed everything from Artforum, it produces the texts it does now. Which includes a long list of writers, of varying quality, a house style, certainly.
    • What if GPT-2 is only fed Bruce Hainley's body of work for AF?
    • Or Rhonda Lieberman?
  17. Cade Diehm on Jul 19th 2020
    • it can only mimic what it is given.
    • But the more GPT-2 is given, the less absurd its outputs are
  18. Paige K. Bradley on Jul 19th 2020
    • Is what the generator produces now a kind of leveling out, and then it produces texts that are, indeed, a kind of humdrum low-level thrum of art writing
    • What if it was fed only the extraordinary?
  19. Andrew Thompson on Jul 19th 2020
    • i'm actually curious what you would consider to be an example of the extraordinary in AF
  20. Paige K. Bradley on Jul 19th 2020
    • Rhonda Lieberman lol
    • The Loser Thing!
    • September 1992 man
  21. Andrew Thompson on Jul 19th 2020
    • well there are more of these critic picks than features in Artforum which has an effect on what the generator outputs
    • like the mag and website have way more small reviews than long features
    • and we also decided to omit straight news stories about a new curator talking the helm at the omaha museum of art etc
    • basically all the language model does is predict what the next word is going to be
    • so the first word in every text is nothing, there's no word
    • the generator then decides what's likely to come after nothing, and it does so stochastically so that it's not the same result every time (hence why it's non-deterministic)
    • so it decides that after nothing, the first word is "Since
    • then it decides what comes after nothing-Since and decides it's the
    • then it decides what comes after nothing-Since-the, and so on
  22. Paige K. Bradley on Jul 19th 2020
    • What comes after 'nothing'. Beautiful.
    • I felt that.
    • I, too, base my entire creative process on that.
    • Maybe I need to hang out with text generators more.
    • I could learn a lot.
    • I've gotten all I could from the writers!
  23. Cade Diehm on Jul 19th 2020
    • There's definitely a way to interpret GPT's generation technique as synthesizing meaning and direction from nothing
  24. Paige K. Bradley on Jul 19th 2020
    • Or, a whole lot of 'nothing'. Which is what art writing is, functionally.
    • It's pointless! I love reading about art but I also hold two thoughts in mind at once, that this is a whole lot of nothing. And it piles up.
    • Like a history.
    • Some people say something is pointless as a point of criticism though. I only like pointless things though, sadly.
    • And with joy.
  25. Andrew Thompson on Jul 19th 2020
    • i want to go back to the loser thing again and untease why you think it's exemplary
  26. Paige K. Bradley on Jul 19th 2020
    • Well a reader compendium of Ms. Lieberman's AF texts came out a few years ago, so I read that and old Scene & Herds and other essays altogether over the course of a month. She's a contributing editor and has written for the mag for a while, though not of late. I fell in love with her style, her firecracker prose and Jewish intellectual sass! With regular reminders of despair. I think her writing is fabulous. She has a manifest attraction to work, situations, dynamics that are borne out of desperation, which is perhaps big part of glamour. Glamour is maybe about being desperate, with ease, rather than disaster. But disaster always looms, the possibility that things could go wrong creates a kind of histrionic insistence on beauty. And in that text I think she's zeroing in on artists/works that are foregrounding the despair, and then that becomes glamorous, like the same ingredients are there but the top and bottom notes are inverted. So it's loser art, but glam, rather than trying to be glam because you're afraid of being a loser/admitting everyone's a loser trying to get the winner's clothes on.
    • And I'm obsessed with this line: "This is totally poetic, because poetry always says the stuff that’s extra, as if the universe appreciates it."
    • I would say that this is not what most art writing reads like. Or at least not art reviews.
    • So I was interested in thinking about how the generator flattens out something I would call unique, into the most typical, recurrent styles.
    • Because that's the majority.
    • And a writer who actually has their own way of writing about art, or a distinctive voice or style, is rare, I would say.
    • Most art writing is formulaic
    • Here's a question: does the generator possibly just reflect back what we already know? Does the automation reveal the automated, DOA style of most art writing? But is that a revelation, and if so, to who?
    • That's off the top of the noggin
  27. Andrew Thompson on Jul 19th 2020
    • I think the value of it is taking the very thing you said we intuit, taking a kind of writing that is figuratively formulaic, and making it literally formulaic. And I think the formulaic quality of what the generator produces not only mirrors the nature of the kind of non-Lieberman writing you're talking about but also much of the work the writing is describing
  28. Cade Diehm on Jul 19th 2020
    • What's intriguing is that it conceives of "plausible art" during its generative process.
  29. Paige K. Bradley on Jul 19th 2020
    • Yes
  30. Andrew Thompson on Jul 19th 2020
    • A lot of what you mentioned earlier about the typical path to Artforum I think is really important to note, which is that people typically pass through various institutional gauntlets before they finally arrive at this particular institution, and the same can be said of the artworld at large especially with the now dominant role MFA programs have come to play
    • the thing is that passing through those institutions (MFAs, PhDs, elite schools) requires one to actually become a formula to succeed in those institutions. This is something Barrett Brown and I talked about with the growing predominance of people at elite schools in journalism
  31. Paige K. Bradley on Jul 19th 2020
    • Absolutely.
    • As institutions of higher learning are essentially teaching class training, as much if not moreso than information. The latter of which you can get at the library.
    • If your libraries haven't been defunded yet that is...
    • But so if the institution is required to teach writing, to teach about how to write about art, then on a logical level of inputs to outputs it doesn't even compute that I was an editor! Would an algorithm have predicted it? Yes I went through art institutions (CalArts, RISD), but I wasn't in any training for writing/discourse per se. My education was distinctly marked by a kind of "we just want to make cool/sexy art", perhaps due to the age of my peer group in undergrad. So how was I rolling up for four years to edit texts from people with masters degrees/PhDs? It doesn't add up. Routing people through paths of learning, as if that will turn them into generators of the thing needed, doesn't work in reality, I don't think. But actually I think what I'm trying to think through here is something about autodidactic doing. The generator just does something, based on what's given, but it can't produce something of interest or value automatically. It lacks something, but so do people.
    • What is lacking in the failure of the generator to produce something unique, and writers' failure as well. Is it a common flaw, or are these different flaws?
    • It can reproduce some of the worst qualities of art writing, like a kind of insider-y, assuming voice, e.g. here
    • "as we all know"
    • This is a generator text, but I still felt for a moment JUDGED and DUMB for not knowing what I should have known, according to the authorial voice suddenly before me
    • That's funny!
  32. Andrew Thompson on Jul 19th 2020
    • lol yes
    • what i find really fascinating over the years is how this is something that seemingly everyone acknowledges about art writing and yet it just continues as a convention
    • so your remark "The generator just does something, based on what's given, but it can't produce something of interest or value automatically. It lacks something, but so do people."
    • seems to be because what is taught is to literally follow a set of steps
    • not to cite our own shit too much but it's how we wind up with bodies and spaces
  33. Paige K. Bradley on Jul 19th 2020
    • Right. I think teaching can be damaging, in a sense.
  34. Andrew Thompson on Jul 19th 2020
    • there's this line by Bordieu somewhere about cultural capital as like "knowing without feeling"
  35. Paige K. Bradley on Jul 19th 2020
    • Or its damages are distributed in the traditional ways. In terms of art education, Judy Chicago wrote a whole book about the pitfalls of college levels arts education in the U.S., based on her various teaching appointments. Female students being stranded inside their own educations by pedagogy that did not serve them, etc.
    • Right, you learn through mimesis how to perform expression.
    • Which is fundamental, but that brings up masking, which is a protective technique when actually no one is helping one learn how to be, or how they could be
    • Just to be like whatever is necessary
  36. Andrew Thompson on Jul 19th 2020
    • so we not only a predominant form of art writing that's algorithmic but also art itself that is algorithmic in the same way, that mirrors the writing. there are a few consequences to this:
      1. you end up with art that is the opposite of visceral
      2. you end up with art that is utterly irrelevant except to a cultural overclass
      3. you end up with participants throughout all aspects of the discipline, except the upper overclass, who seems to actually acknowledge all of this but feel trapped within this system of art that feels empty even as they continue operating within it
    • my personal experience of a lot of the artworld's participants is as straight up sociopaths the way AI is, a system that is optimized solely to win a game
    • like the way AlphaGo can do nothing but win at Go
  37. Paige K. Bradley on Jul 19th 2020
    • Yes. Though I also think it's valuable to think through intellectual/visceral as a binary and on what connotes either.
  38. Andrew Thompson on Jul 19th 2020
    • sure and i don't think everything in our world needs to produce a visceral response, but i think a lot of the emphasis on intellectualism in the discipline is designed precisely to project a sense that actually there's something visceral involved when there isn't
    • i don't know if you're familiar with that piece An Oak Tree
    • i find it important because it was one of the early examples of translocating artistic value in textual argumentation instead of the object, and that all the actual artistry would henceforth be located in placards.
    • there's a great detail about this in the wiki
    • because yeah, there isn't a fucking tree in that jar
    • and it feels like the entire discipline has had this weird self-loathing for the past like 20 or 30 years now where even they don't believe there's a tree in the jar anymore, the reigning sentiment among even participants in this very space is that all of its totems are predatory institutions designed to create value for pedo art collectors and little more
  39. Geoffrey Mak on Jul 20th 2020
    • Hi guys! Jumping in here a little bit late. A brief intro: I started writing for Artforum I think a year or two ago, and mostly do Critic's Picks. I've written for a lot of art magazines, though I think I enjoy writing for Artforum the most because I feel like I get an education from my editor every time. I, like Paige, did not study writing, I'm an autodidact, which was fortunate in the sense that I only read and learned from what gave me pleasure, stuff that I picked because I liked it, and that oriented my education.
    • I'm thinking more about the project, though. And it seems that the AI can detect certain patterns in the Artforum house style. But the AI (to my limited experience) can't mimic the experience of being moved by aesthetic experience and representing it in prose. I'm thinking of one of my editors Chloe Wyma's review in the current issue. Paragraphs like these are really exceptional:

      "If Morris’s evacuated landscapes and foreclosed views, as Searle wrote forty-two years ago, led us to “infer a somehow sacrosanct ‘humanist’ or ‘Existential’ content,” today they seem almost too resonant with the vacant street I see from my apartment window, the blanketing stillness perforated by ambulance sirens, and the atmosphere of a world hollowed out. This reading—sentimental, ahistorical, unavoidable—is compounded by the flat interface of my computer screen (another window), which so efficiently exchanges substance for appearance, objecthood for representation. Morris labored over his work for years and rarely let anyone, with the exception of his wife, into his studio. It feels obscene, maybe even fraudulent, to write about this private, intense art at such a remove. But this impoverished mode of spectatorship also heightens the contradictions that give Morris’s work its tensile charge, poised between repellent surface and seductive illusion, dense matter and mythic image—between the “desire to create a world in one’s work in which one could live” (as the artist wrote in 1981) and the knowledge that living there, or anywhere else, means living with distance, mediation, and displacement. As the artist once put it, “We necessarily experience the world through our bodies, we have as it were a view from a house of blood.” "
  40. Paige K. Bradley on Jul 20th 2020
    • I would 100% agree many if not most people working in the arts, who I’ve met, loathe much about it.
    • Or loathe themselves for participating, sure.
    • I don’t hate myself! For the record.
    • But textual argumentation, yes, I’ve argued often that art is mostly about writing now. The recent Whitney Biennials have been about letters and critiques and counter critiques. There’s so much to read who has time to look, much less see.
    • The context in which art takes place has become overbearingly dominant, yes, but that’s also due to structural systemic issues around funding, access, purpose.
    • But in terms of writing, I would say quite a lot of the art I hold most dear, that has had and continues to impact my life, is intimately tied to writing. And that work isn’t new or recent.
    • Thinking of Mike Kelley’s performance writing, his video scripts, or even just statements and interviews he gave about his own work.
    • David Wojnarowicz, an untrained artist and writer, his writing just flys off the page and throttles your skull.
    • I can’t separate Hannah Wilke’s work from the instance she wrote “I want to overwhelm you”
    • Museums putting up placards to tell you what the work is doing is I think moreso about trying to control the narrative of works and art.
    • Because authoring narrative and owning those stories is in a sense the main control valve for them, maybe even moreso than owning the object.
    • I’m not sure that’s the same thing as saying art or artists abandoned aesthetics for text or argument. That was part of the visual appearance in conceptualism at a certain time and maybe even place, but has text on a wall ever been the only game in town, as it were?
    • Here’s a point I want to make though that loops this back around to the text generator. Art writing is full of humdrum garbage, with what I would consider singular voices/notable intellects sprinkled amongst, and the generator by feeding on vast quantity seems to equalize or mediate between all, to turn out a digestion that reflects back the most rote qualities of the source texts. That’s useful in a sense, but I don’t see it as a condemnation of the ongoing effort by writers to articulate what is happening in a work or body of work by an artist. This response is essential to art, I think, because it’s modeling the idea that when an artwork rolls up we have any kind of responsibility to it besides not physically damaging it.
    • The artist makes the thing, someone else ideally pitches in to place it in the vicinity of people who could care, and now there is work to be done.
    • There is an ideal, and maybe it’s just my own, that at least half of the work, of the total work that actually constitutes the artwork, is done by the viewer.
    • I feel a responsibility to a work, even if I don’t have an actual work assignment per se to represent it. Someone made this, it must be engaged with in best faith, and even if eventually one concludes that it is Not Good, you should be patient about arriving at that dismissal.
    • There are these kinds of broad, generational critiques that go like “well in this era the art was this but now it’s like that and we don’t like that” and I’m always suspicious of such.
    • And so when someone says “art writing,” ok some people are thinking about magazine reviews. I’m thinking about how text is literally projected onto Lee Kit’s paintings.
  41. Cade Diehm on Jul 20th 2020
    • I keep thinking about your observation about the context of art.
    • and the different functions of art writing, both as a conceptual or execution element within a piece itself, but also the various forms it manifests
  42. Paige K. Bradley on Jul 20th 2020
  43. Cade Diehm on Jul 20th 2020
    • Placards, or concept notes, or descriptive reviews.
    • I've sort of existed on the outskirts of the art world for a long time, and I sort of feel this deep numbness to much of what I've viewed or read up on. As I browse this algorithm's reviews, I feel the same kind of numbness. Not just an alienation to the review, but an alienation to the work that's detailed in the review
    • When Andrew was showing me the early tests of the generator, I couldn't really reconcile the uncanny valley nature of what it generates. OpenAI has a new version of GPT out –– GPT-3, which is all over the internet right now for its supposed accuracy, but when I view outputs from both GPT 2 and 3, I'm reading into an emptiness.
    • Part of this is definitely from the nature of this project's execution, and my technical understanding of how it works. But I still can't quite reconcile the reaction I have to these synthesised alternative universe works.
    • If we step away slightly from the deeper exploration of writing and authors and so on, what feels so raw to me here is the uncertainty of my own reaction to it. On one hand, it's clearly somewhat absurd in its vague nonsense, but I remain deeply drawn to these 'alternate universe' art pieces, and my own reactions to what they might look like. They resonate strongly with me, but what I feel is a similar kind of greyness.
    • What I feel is a sort of synthesized hyper-reality. The same sort of thing that I feel when I see, for example, the amount of work that's been taken to create the uncanny reality of the NaturalVision Evolved Grand Theft Auto mod:
    • Increasingly disillusioned with cities and their aesthetic landscapes, yet somehow enthralled by these weird mimickries.
    • Its hard to unpack this response to hyper real Grand Theft Audio and GPT Artforum. Are they drawn from a similar place as Geoff's Edgelords of participatory alienation, where my discordant anxiety is the product of a social and political void? I don't think so, even if I am someone who has resoundingly jettisoned my own career in design for similar reasons. Is this a defence mechanism to the increasing misalignment I feel towards built environments? It's cliche to veer off towards back to natures meme, whether it's simplistic fantasy of life as a trad-girl or being deeply thrilled by an under-developed solarpunk aesthetic.
    • Regardless, I feel the same when I browse Deepfake ArtForum as I do in within these spaces themselves. And I don't know why.
  44. Andrew Thompson on Jul 20th 2020
    • Geoff can you briefly describe the Edgelords essay Cade just mentioned, because it's actually I think relevant to a lot of what we're talking about
  45. Geoffrey Mak on Jul 20th 2020
    • Hi guys! Yes, "Edgelords" is an essay of mine that examines my experiences during the 2010s—the decade of platform capitalism—particularly the "edgelord" aesthetic that emerged from the way social media platforms privilege shocking and transgressive content. I talk about the rise and fall of a media industry shaped by Gawker, The Huffington Post, and Breitbart, and various "shocking" cultural phenomena that grew alongside it: Milo Yiannapoulos, Gamergate, Herrensauna in Berlin, Kanye West's psychosis-cum-celebrity praxis, the Art Basel Miami stabbing, Balenciaga crocs, and K-Hole.

      The essay can be found here:
    • It took me four years to write the essay, and since publishing the essay at the end of last year, my thoughts on it are continuing to evolve beyond my original intentions. Like, for instance, I think now the essay senses (rather than articulates) a new hermeneutics created by social media and platform capitalism. How art is valued has been irretrievably changed in this recent platcap decade we've emerged from.
    • But to be clear, it's not a work of criticism. And—bringing it back to the topic of the text generator—my essay senses/intuits a sort of rift from traditional art criticism (emblematized by Artforum or eflux, which is quoted in the piece) and the art that is being made right now. I find comfort in the prose of traditional art criticism (like I just bought the new Hal Foster book and a ridiculously excited to read it) partly because it feels nostalgic and comforting, as if an old world were still relevant. But it's not. I'm not sure we have found a new prose to describe this new hermeneutics conditioned by platcap/social media.
  46. Cade Diehm on Jul 20th 2020
    • I wasn't framing Edgelords as a criticism, rather as a discrete disorientation that resonated with me.
  47. Geoffrey Mak on Jul 20th 2020
    • Yes, I know you weren't. Though maybe there's something to be said about what kind of prose resonates with readers who have sustained art practices, and whether or not it's a problem that art theory "mumbo jumbo" (as another art writer friend of mine calls it) leaves many cold
    • I personally really enjoy reading art criticism, and found that with a lot of my free time during the pandemic, I did return to a lot of art texts that excited me. Even reading Paige's description of Rhonda Lieberman excited me, as a lot of her prose has from the magazine over the years. But there is a huge range of accessibility in Artforum's archive. Even an writer like Hannah Black will slide along the accessibility scale—some of her stuff is extremely cogent, intended for a popular audience, and some of it is completely impenetrable and leaves me cold.
  48. Andrew Thompson on Jul 20th 2020
    • the apparatus that embraced K-HOLE (who were written about in artforum and put on panels with hans ulrich obrist and so on) is the same on is one that seems fundamentally premised on embracing falseness, either cynically or otherwise---the falseness of brands or the falseness of a tree being in a jar, a belief in something that is literally demonstrably absent
    • to me, the void-like reaction Cade has to the generated texts is reacting to one step of abstraction beyond something that's already hyper real
    • so GTA 5 is a simulation of los angeles, which itself is already a simulation. the simulation of a simulation is kind of a new order of simulacrum. and i think the GPT reviews trigger a similar reaction for the same reason---not simply because they're false, but because the data they're trained on is false to begin with
  49. Geoffrey Mak on Jul 20th 2020
    • Yes. A quote from a video of one of my collaborators, Christopher Kulendran Thomas, says "Everything is a simulation of simulated behavior"
  50. Paige K. Bradley on Jul 20th 2020
    • Yes, but this gets to the assumed, false premise that there is a format to writing about art. If the imitations proliferate, they are both good and bad faith mimicries.
    • I don’t think there actually is per se. There is a history of writing about it, and so perhaps you can go to Stanford and Harvard and they will demonstrate what John Ruskin was doing.
    • That could be interesting and meaningful.
    • But I think that’s one way, and trying to imitate October theorists from the 1970s is an abstraction.
  51. Geoffrey Mak on Jul 20th 2020
    • I feel like there is a format to writing about art, no? It has a tradition, and it's institutionally acknowledged/protected as a discipline.
  52. Cade Diehm on Jul 20th 2020
    • I'd agree with that I think
  53. Andrew Thompson on Jul 20th 2020
    • i'd also say there's a format and we're seeing much of that format materialize in the generator. it doesn't mean that format is the only way people can let alone should write about it, but it has become a de facto format nonetheless
  54. Geoffrey Mak on Jul 20th 2020
    • Hal Foster's recent Artforum piece (haha) about Rachel Harrison was about how her format of sculpture is the organizing principle that includes other mediums, such as painting. And I think art criticism is a format that can include many other things that aren't recognizable as writing.
  55. Paige K. Bradley on Jul 20th 2020
    • There is a format that’s been codified, but that doesn’t mean it is the only way. Writers who understand the format enough and who can then fuck with it, break it up, or treat it as so much pliable material, aren’t codified into or by the format.
  56. Geoffrey Mak on Jul 20th 2020
    • Yes—I think Bruce Hainley is a good example of that
  57. Paige K. Bradley on Jul 20th 2020
    • Totally.
  58. Geoffrey Mak on Jul 20th 2020
    • Another interesting question is whether Scene & Heard/Diary pieces are considered criticism (I think they are)
  59. Paige K. Bradley on Jul 20th 2020
    • Oh I think so, in the hands of the right writer.
    • I mentioned Lieberman’s Diaries yesterday as well.
    • Though that specific column’s energy, and critiques of it, both feel like they have subsided at this point, perhaps appropriately.
  60. Andrew Thompson on Jul 20th 2020
    • let me go back for a second
    • because what I read in your work Geoff was among a lot of things, a kind of howl from the corruption of the artworld at large, which is not an unusual reaction one would has toward it but which you seemed to share
  61. Paige K. Bradley on Jul 20th 2020
    • But something I did want to interject here with is that with the ever-present refrains of “art writing sucks/leaves me cold” (surely there were people saying this in the 80s? The 70s? Perhaps the 1890s? Is this a new criticism? I wonder) I have to ask if this is unique to the field of art, that its discourse is considered tired, rote, exhausted, meaningless.
    • Are the people over in physics complaining that a lot of their colleagues are just greasing their field’s wheels as well, with pointless papers?
    • Wouldn’t any industry have its babbling chatter?
    • Its needless overproduction of hot air?
    • And does all that discourse served cold become easily ignorable then, like a whole lot of silence, if nothing is being said?
  62. Geoffrey Mak on Jul 20th 2020
    • because what I read in your work Geoff was among a lot of things, a kind of howl towards the corruption of the artworld at large, which is not an unusual reaction one would has toward it but which you seemed to share
      Yes, this is certainly true, particularly the Basel Miami section, although the essay as a whole wasn't all negative—I did prop up Christopher Wylie as the hero figure in the piece, even if outside of my mythologized version of him I have mixed feelings about him.
    • But yes, Paige I agree that ppl feeling cold about the mumbo jumbo is not a new condition and is certainly not unique to the art world
  63. Andrew Thompson on Jul 20th 2020
    • i would also say it's not unique to the artworld but i would say the artworld depends on it for its very existence in a way that is particular to both it and portions of academia
    • which is why i ask about geoff's reaction towards a corrupted artworld, wondering A) what encourages continued participation in it and B) what a non-corrupted version of it would look like
  64. Paige K. Bradley on Jul 20th 2020
    • And so in this Deleuzean way we can think about where in art writing, between all the chasms of silence where nothing is being said, the thing that is worth saying? He said oppressive regimes don’t really prevent people from expressing themselves, rather they force them to. The art writer has to keep getting up on the platform to chatter to get a byline to get a grant to make up for not being paid properly in the first place. And so rarer and rarer becomes the writer and the text that says something worth saying. But isn’t what’s actually great, or good, always rare?
    • Has it been otherwise? When?
    • There’s always an oldster to tell you things used to be better.
  65. Geoffrey Mak on Jul 20th 2020
    • Hmm... Well Andrew's questions are deeper questions with long answers. But the short version is that I believe in the perhaps old-fashioned notion of aesthetic experience. I feel moved by it, my spine shivers when I encounter good art (and I don't have qualms about using the word "good" because i believe in craftsmanship and expertise) which is the closest approximation or simulation I have to love.
  66. Paige K. Bradley on Jul 20th 2020
    • I believe in this too.
  67. Geoffrey Mak on Jul 20th 2020
    • And secondly, I don't think there is a non-corrupted version of the art world, that's impossible
  68. Andrew Thompson on Jul 20th 2020
    • i think all four of us here fundamentally believe in the first part
  69. Geoffrey Mak on Jul 20th 2020
    • Also—I don't think craftsmanship and expertise can be simulated. But that may be conservative of me (I'm a little conservative in my art criticism tbh)
  70. Paige K. Bradley on Jul 20th 2020
    • I continue to participate because thinking deeply about art and trying to articulate something worthwhile about it and making it or understanding what needs to be remade and why is endlessly engaging to me . I don’t really do drugs so it’s the high I know!!
  71. Geoffrey Mak on Jul 20th 2020
    • It really is a high
  72. Paige K. Bradley on Jul 20th 2020
    • Some people get their high from the market and the strange ululations and precipitations of financial value that it puts on a show of.
    • There are many shows happening at once.
    • I’m frustrated that anyone who does not feel that way about aesthetic experience is in any way involved in art.
  73. Andrew Thompson on Jul 20th 2020
    • right i would agree with all of this too
    • so like not to be a dick and quote him but it kind of all goes back to the nietzschean idea of the universe only being justified as an aesthetic phenomenon
    • and i think your standard artist and art writer and anyone within the artworld would at least agree with that on its face
  74. Geoffrey Mak on Jul 20th 2020
    • The high from the market is a real and distracting thing, and the critic experiences this high when she is constantly courted by galleries, collectors, and artists for her power, which deceives her into believing that her proximity to the upper classes can be mistaken for the real thing. (For instance, AF's Diaries articulate this well as some form of essentially institutional critique)
    • Oh interesting. Could you explain more about "the nietzschean idea of the universe only being justified as an aesthetic phenomenon" bc I'm not familiar with it
  75. Andrew Thompson on Jul 20th 2020
    • yeah if anyone outside gets this far in the conversation they might take me to task for bungling this idea but
    • a lot of his philosophy was based on the thought experiment that you'd have to live your same life over and over for all eternity
    • and the only way to justify that phenomenon of living over and over again was to basically see the universe as art
    • and for nietzsche/schopenhauer the form that aesthetics took was fundamentally musical. so music was the thing that both kind of is the universe and also makes the universe worth it
    • i can listen to music and intuit that idea, and it's why music has a vitality for me and probably for you that feels unique to it
    • i don't feel that with the fucking jar of water with an absent oak tree
  76. Paige K. Bradley on Jul 20th 2020
    • The eternal return
  77. Andrew Thompson on Jul 20th 2020
    • right
  78. Paige K. Bradley on Jul 20th 2020
    • Ok sure. But I’m glad you brought this up because this is a common thing, this idea that music can give us, or we’re more receptively built to feel, a feeling that stationary art objects cannot.
    • But music does more easily, to more people, faster.
    • Ergo it’s superior.
    • I’m not convinced. I love music, it has been important in my life etc there’s nothing I would say on that which would be any different than what another might.
  79. Andrew Thompson on Jul 20th 2020
    • But music does more easily, to more people, faster.
      are you saying it can't do this or that it can do this but that doesn't make it superior
  80. Paige K. Bradley on Jul 20th 2020
    • If you can find a way to feel about the absent oak tree and the glass of water, and I do mean find a way, precisely because the path is not clear, it has not been prepared for you, then I wonder if reaching a point where that becomes beautiful to you? That enlargement of capacity to be moved? To recognize and to see, rather than look, that is surely part of the whole point.
    • This is what I meant about the viewer doing half the work
    • At least
    • I wrote about this specifically re: Lee Kit’s work for AF:
    • It can but that doesnt make it superior.
    • If five hundred people can love something that doesn’t mean the thing is better than the thing five people love
    • That’s a capitalistic kind of notion. That things are better if they can be scaled
    • No?
  81. Andrew Thompson on Jul 20th 2020
    • i would take the total opposite tack and argue that the sense that something can only be appreciated be a tiny subset is what imbues a lot of art with a rarefied value that appeals to collectors and is the very basis of the idea of cultural capital
  82. Cade Diehm on Jul 20th 2020
    • Consensus can transcend capitalism
    • Both are leveraged by capitalism, really.
  83. Andrew Thompson on Jul 20th 2020
    • i mean you can go either way really, but in the artworld i see an intentional emphasis on the rarefication of appreciation
  84. Paige K. Bradley on Jul 20th 2020
    • But does everything have to be plentiful? Besides food, water, shelter? Is it inherently wrong for something to be rare? I understand rarefied art, are there not other thing that would be rarefied to me then? Things I can’t understand, see no use for, are closed to me? Does everything need to be accessible at all times?
    • Does the flower that only grows in one field in one part of the world get criticized for its elite, rarefied nature
    • I’m just never quite convinced that calling something rare is automatically a condemnation.
  85. Cade Diehm on Jul 20th 2020
    • Rarity implies a value appeal
  86. Paige K. Bradley on Jul 20th 2020
    • Yes supposedly the artworld does insist on rarefied appreciation. But then, pre covid-19, how do we square this with the demonstrable phenomenon of “blockbuster” exhibitions, at least in American, even specifically nyc museums, which are also supposedly enforcing this elite, enclosed atmosphere? If you love art you can’t go to MoMA on a weekend, you’ll have to dodge around people taking photos of art. It’s a mall. I found this reality incompatible with what critics of the art world tell me about how art exists in the world today.
    • Sure, but again having 500 of something instead of one doesn’t make it more democratic or accessible per se. I’m suspicious of this.
  87. Geoffrey Mak on Jul 20th 2020
    • Rarity and exclusion is the sine qua non of luxury
    • But rarity and exclusion doesn't necessitate luxury
  88. Paige K. Bradley on Jul 20th 2020
    • I agree.
    • If you’re the exclusive owner of a Brancusi or some such, wouldn’t that actually comes with tasks, responsibilities, expenses. It seems like a burden! Is that luxurious? Maybe on the day you get to consign it to the auction house or write it off on taxes after donating to a museum.
    • It’s not like taking a bubble bath while someone gives you a manicure
  89. Geoffrey Mak on Jul 20th 2020
    • Oh the tasks, responsibilities, and expenses are totally luxurious
    • Constantly checking your investments in the stock market is luxurious
    • Complaining about property taxes is luxuriuos
  90. Paige K. Bradley on Jul 20th 2020
    • Haha right
    • Dunking on art writers seems too easy, I’m not sure if the context and economy around what produces this distressing amount of bad writing we all decry is examined enough.
  91. Andrew Thompson on Jul 20th 2020
    • it's not simply dunking on art writers per se as much as it is raising the possibility that if everything is affect---a tendency that as we discussed at the beginning goes back to elitism in some way, either people passing through elite institutions or those imitating who have (and i would say this accounts for the droves of people who show up to MoMA, to fulfill some obligatory cultural requirement)---then we can just do all that with a machine now and it won't make any discernible difference. and now we have that machine. we can produce the very works of art the machine's texts are talking about, and we of course have the review to go with it which can be adapted into either a statement or a placard to accompany the piece, and the entire apparatus can be taken over by our AI
  92. Paige K. Bradley on Jul 20th 2020
    • Yes that’s true. In a sense. But what it cannot do is tell you accurate information, not interpretations or poetic conjectures, about what Robert Smithson made, when he made it, where it was, what materials involved, etc.
    • There is no art criticism or art writing without that basis. You cannot make a claim about Spiral Jetty without saying where it is, that there is water, there are rocks.
  93. Andrew Thompson on Jul 20th 2020
  94. Paige K. Bradley on Jul 20th 2020
    • The generator works with everything that springboards from that. The flight and takeoff.
    • The crash and the shipwreck!
    • With no disrespect intended to Smithson’s own death circumstance lol
    • Smithson was mentioned in the generator text Cade posted earlier
  95. Cade Diehm on Jul 20th 2020
    • So above, the two of you are discussing the accuracy and intent of generative art writing, and how much of the apparatus can be automated in this way.
  96. Paige K. Bradley on Jul 20th 2020
    • American artist from the 1970s who did land installations
    • Coincidentally Smithson writing about his process, what he was doing and thinking about, are actually some of the earliest major essays in Artforum
    • A generated text might say something about the convolutions of a history of interpreters of a work. Of Smithson, to stay with an example.
    • An interesting expression about the contortions or inspirations of statements around a work.
    • It can sit next to Smithson’s essays from Artforum, if you like.
    • It does not replace them.
    • And isn’t that an interesting thing, a generative premise, when 1 and 2 are next to each other, and third meanings can proliferate? They complicate one another, rather than competing per se.
    • If one isn’t meant to supersede the other, but talk back. And the feedback draws in the addressed and other addresses.


GPT-2 generator: Jules Becker, Becky Brown

Discord app: Becky Brown, Khalil Moriello