@CyberneticForests@assemblag.es
@CyberneticForests@assemblag.es avatar

CyberneticForests

@CyberneticForests@assemblag.es

Artist/Researcher looking at algorithmic bias, design research, Situationist Cybernetics and Mushroom Synthesizers. CyberneticForests.com

This profile is from a federated server and may be incomplete. Browse more on the original instance.

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

The issue with the existential risk from AI doomspiel is that the minute anyone actually begins to take the chain of causation seriously enough to think through, they realize how slim the chances of “all civilization ends” turns out to be

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

You could critique almost any “_____ AI” movement (“ethical AI,” “responsible AI,” “AI safety,”) as setting its own rules for others — because the critique is actually a critique of professionalization. An empowered, participatory AI governance model really ought to make room for more than experts and specialists.

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Lev Manovich bringing some “average Twitter user energy” to LinkedIn. I guess Lev is upset people are reading something he didn’t write. Nice to be hitting the stage of my career where Lev feels like he has to call my work boring instead of just ignoring it. :)

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Lots of new followers! Thank you! Just as a warning, one of my upcoming research projects may be a focus on how AI artists contribute to antisocial tech power by promoting the hype cycle and corporate mythologies of big tech. If that sounds like a thing that is going to make you angry you may not be following the right person! ;)

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Finally deleted Twitter. It wasn’t making me smarter, was often infuriating, its architecture was designed to erase nuance, and its owner is an asshole. Feel a bit antsy about it, but the trick is to find other ways to do the things I thought it did. If you get value out of my newsletter and would like to recommend this account to your networks, I’d appreciate it!

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Large Language Models were never designed to produce text that is true — they are designed to produce text that is statistically likely within a sample. Now, any company with a serious commitment to AI has to figure out how to constrain these errors of “truth” —descriptions of the world beyond small samples. They have to do this with a tool that was never designed for outcomes that align with the broader scope of the world in any way. https://cyberneticforests.substack.com/p/a-hallucinogenic-compendium

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Now that OpenAI has shut down it’s “Super Alignment” team, let’s clearly define “AGI Safety.” https://open.substack.com/pub/cyberneticforests/p/nobody-can-save-us-from-imagination

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Aurora Borealis over Lake Ontario tonight, as seen from from Rochester New York.

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Gaussian Pop: 14 Theses, a reluctant manifesto for the contemporary moment in visual art, sound and text.
https://cyberneticforests.substack.com/p/gaussian-pop-14-theses

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

The deadline for ARRG!’s “Models for Making Distance” Zine is in one week! Show us the art you make with broken algorithms as your medium; the manifestos that demystify AI and clarify its purpose, instructions for algorithmic resistance at home. https://www.cyberneticforests.com/zine

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

I've always assumed that "hallucinations" in LLMs were the result of Simpson's Paradox. But I've never heard anyone else say so explicitly. So I want to ask if my understanding is accurate. Here's what I mean, using sports and college admissions data. Long, mathy thread. [1/19]

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

My contrarian take on “things are moving so fast in the AI space” is that actually they aren’t. The details flicker but the structure stays the same. Slight variations on an LLM, or tweaks on a diffusion model, don’t change the fundamentals.

CyberneticForests,
@CyberneticForests@assemblag.es avatar

@luis_in_brief glacier takes

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

The Algorithmic Resistance Research Group (ARRG!) is calling for examples of the creative misuse of algorithms for human expression to be printed in “Models for Making Distance,” our print-only Zine. Deadline is just under a month away! Details: https://cyberneticforests.com/zine

tante, to random
@tante@tldr.nettime.org avatar

"It turns out that generative AI companies don’t like it when you steal, sorry, scrape, images from them. Cue the world’s smallest violin."

In Moment of Unbelievable Irony, Midjourney Accuses Stability AI of Image Theft

https://www.themarysue.com/midjourney-accuses-stability-ai-of-image-theft/

CyberneticForests,
@CyberneticForests@assemblag.es avatar

@tante @benh I have to admit I’m also confused about what SD would want from Midjourney data, but agree it is likely to be related to requests and corresponding outputs rather than using image-text pairs for generating images. If I was to guess, they’re building a shadow prompter to turn user prompts into more effective language for the model and want to see how MJ users prompt MJ. But it seems very weird. Hope they sue so we get discovery haha

CyberneticForests,
@CyberneticForests@assemblag.es avatar

@tante @benh That’s a result of training focused on weighing images for “aesthetics” though, I think, not prompt skill (whatever that is). But yeah, Midjourney is a massive dataset of user prompts just sitting out in public waiting to be scraped b/c they insist on making prompts public on Discord. So it’s a dataset someone might want “just in case” even if the immediate use case is totally vague.

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

I have a theory that children of narcissistic parents who have come to terms with it are particularly well suited for working with generative AI systems: models that pretend to understand and respond to you but only as a means of advancing their own conversations.

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

We’re no longer living in the information age — we’re living through its inevitable conclusion, where information is so bountiful that it becomes illegible. In a video piece and talk made for ACMI’s FACT symposium in Melbourne, I described “The Age of Noise” — and the strategies we need to make sense of it. You can watch (or read) it here: https://cyberneticforests.substack.com/p/the-age-of-noise

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

I’ve been boosting data rights because I knew where this was going: Docusign scraping your legal documents, Tumblr scraping private & deleted posts. Wordpress scraping your writing. Reddit scraping posts. They trapped us in their infrastructure and now they’re selling us out.

CyberneticForests,
@CyberneticForests@assemblag.es avatar

“The Biden Administration’s Blueprint for an AI Bill of Rights calls for users to be “protected from abusive data practices via built-in protections” and to “have agency over how data … is used.” Our images and words, too, are “data about us.” From 2023: https://www.techpolicy.press/a-new-contract-for-artists-in-the-age-of-generative-ai/

CyberneticForests,
@CyberneticForests@assemblag.es avatar

It is obvious to anyone who paid attention to the ideologies of tech companies: “democratize,” then entrap and extract. Wrote about that here: https://www.cyberneticforests.com/news/social-diffusion-amp-the-seance-of-the-digital-archive

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Do machines think? No. Can we redefine the word “think” in such a way that it can include machines? Seems we can. Should we? I don’t know, but I don’t see what it gets us.

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

In November we announced a zine, "Models for Making Distance," which would collect instructions and manifestos for creative acts of algorithmic resistance. The deadline is April 15, and then we print. We would love to have your contributions! More details: https://www.cyberneticforests.com/zine

CyberneticForests,
@CyberneticForests@assemblag.es avatar

@tante up to you, i think!

CyberneticForests, to random
@CyberneticForests@assemblag.es avatar

Another moire and noise work.

CyberneticForests,
@CyberneticForests@assemblag.es avatar

@SrRochardBunson i did not! Tell me more?!

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • megavids
  • mdbf
  • ngwrru68w68
  • tester
  • magazineikmin
  • thenastyranch
  • rosin
  • khanakhh
  • InstantRegret
  • Youngstown
  • slotface
  • Durango
  • kavyap
  • DreamBathrooms
  • JUstTest
  • tacticalgear
  • osvaldo12
  • normalnudes
  • cubers
  • cisconetworking
  • everett
  • GTA5RPClips
  • ethstaker
  • Leos
  • provamag3
  • anitta
  • modclub
  • lostlight
  • All magazines
    •