👖Emancipants👖 @fgshprd.bsky.social · 16/09/2026The threat isn't AGI or "superintelligent" bots. We've created a world of text-driven systems and now systems that generate text that can exploit these existing systems. How will these systems interact? That's what we're about to find out. 000
👖Emancipants👖 @fgshprd.bsky.social · 16/09/2026Just scripts making scripts that find the holes in the vast ecosystem of scripts that support much of our day-to-day lives. These automated "bad actors" are a threat because they overcome the relative scarcity of humans. It's spam that targets infra systems instead of inboxes. And it's all text. 100
👖Emancipants👖 @fgshprd.bsky.social · 16/09/2026They try, modify, try again. Scripts are just text. LLMs automate text production. The execution of a script can be verified against a goal, so this automated textual production can be constrained and directed by objectives. No "intelligence", much less "superintelligence" needed. 100
👖Emancipants👖 @fgshprd.bsky.social · 16/09/2026For decades, most orgs building software have made the security of their systems a 2nd-or-lower rate priority. Now, the vulnerable surface area of these systems, taken globally, is huge. So far, this has been mitigated by a relative scarcity of "bad actors". "Bad actors" operate via scripts ... 100
👖Emancipants👖 @fgshprd.bsky.social · 22/07/2026The flood waters went rogue, finding their way through the window I left open. 020
👖Emancipants👖 @fgshprd.bsky.social · 09/07/2026"Here's what I found — and it changes the framing" 001
👖Emancipants👖 @fgshprd.bsky.social · 24/06/2026So I agree that these are important questions. And these are in fact live questions for at least some people who use the tools professionally. 010
👖Emancipants👖 @fgshprd.bsky.social · 24/06/2026Yes, I see. It's interesting that at my engineering day job, discussions about whether "agents" are "good" at coding revolve mostly around questions about how their use changes how we work and relate to one another. We don't really care about benchmarks and other marketing hype. 110
👖Emancipants👖 @fgshprd.bsky.social · 24/06/2026sorry, 300 chars makes it hard to be clear. I didn't mean to suggest that you equate affordances and uses. was just asking if it's really possible to distinguish the question of use/uselessness from a discussion of affordances and their impact (as you seem to suggest). 100
👖Emancipants👖 @fgshprd.bsky.social · 24/06/2026that said, new "affordances" represent real changes to the world. an LLM chatbot is a new interface to language and changes our relationship to language. given how integral language is to our being, this is undoubtedly a real concern. 000
👖Emancipants👖 @fgshprd.bsky.social · 24/06/2026I'm confused by these distinctions. use/uselessness, user experience seem to be on one side, and then affordances on the other? But isn't user experience all about affordances? Or maybe I put the boundary in the wrong place, but then isn't "use" part of "user experience"? 300
👖Emancipants👖 @fgshprd.bsky.social · 24/06/2026I worry that the shift from "intelligence" to "synthetic vs. natural" is itself problematic. Writing is both/neither synthetic/natural. We're still operating within the problematic of the pharmakon. 110
👖Emancipants👖 @fgshprd.bsky.social · 23/06/2026if the decision to make absolutist claims is merely a matter of tone (and so open to tone-policing), then I guess either 1) the real stakes are surprisingly low; or 2) tone-policing is all that's admitted; there can be no argument with the "substance" on this topic. 000
👖Emancipants👖 @fgshprd.bsky.social · 23/06/2026I was inspired by the BINA48 video at the New Humans exhibit. 010
👖Emancipants👖 @fgshprd.bsky.social · 23/06/2026My 4 week class is starting in a few weeks and filling up fast. Join and read Turing, Vinge, Kurzweil, Pseudo-Dionysius, Teilhard de Chardin, Gebru, Torres, Tirosh-Samuelson, Rothblatt, Land, Laboria Cuboniks, Kafer, Chiang, Wynter, Lyotard, Stiegler, Hui and more! 123
Reposted by 👖Emancipants👖Eryk Salvaggio @eryk.bsky.social · 22/06/2026Image models are not "a" model. Each component was built or integrated to solve a specific problem of computer vision — each engineered to automate a decision, each inscribing its bit of ideology into every generated image. My latest paper unpacks those decisions and what they inscribe.arxiv.orgThe Market in the Model: Latent Diffusion as Neural EconomyValuable critique of generative image models within visual culture and the humanities has emphasized the role of datasets in shaping the images they produce. Yet, close studies of the ideological posi... 05311
👖Emancipants👖 @fgshprd.bsky.social · 20/06/2026giftarticle.ft.com/giftarticle/... Alternative for Germany revives Nazi-era attacks on Bauhausgiftarticle.ft.comAlternative for Germany revives Nazi-era attacks on BauhausModernist art institution fears AfD ‘patriotic culture’ push nearly 100 years after closing under Hitler 021
👖Emancipants👖 @fgshprd.bsky.social · 17/06/2026Things I’ve learned from claude pt 1: “IMPORTANT DISCOVERY” is an effective opening to a confession of incompetence 000
Reposted by 👖Emancipants👖WPRB 103.3 fm @wprb.com · 15/06/2026Hello, WPRB is now on Bluesky. Please tell your friends (and one enemy). 03515
Reposted by 👖Emancipants👖👖Emancipants👖 @fgshprd.bsky.social · 17/05/2026Please share with friends who might enjoy some good old-fashioned critical engagement with The Singularity! I’ve completely refashioned the syllabus for 2026! 001
👖Emancipants👖 @fgshprd.bsky.social · 17/05/2026Please share with friends who might enjoy some good old-fashioned critical engagement with The Singularity! I’ve completely refashioned the syllabus for 2026! 001
👖Emancipants👖 @fgshprd.bsky.social · 08/05/2026Best film adaptation of The Odyssey, wrong answers only 000
Reposted by 👖Emancipants👖Paul F. Tompkins @pftompkins.bsky.social · 05/05/2026Never seen a picture of him before, I always assumed he wore glasses 111557167
👖Emancipants👖 @fgshprd.bsky.social · 04/05/2026Bowie’s Fantastic Voyage is the song for our time youtu.be/FSCB_0SXFR4youtu.beFantastic Voyage (2017 Remaster)YouTube video by David Bowie - Topic 000
👖Emancipants👖 @fgshprd.bsky.social · 20/04/2026It's on! thebrooklyninstitute.com/items/course...thebrooklyninstitute.comThe Singularity: Artificial Intelligence and the Post-Human Future - Brooklyn Institute for Social ResearchAlan Turing, whose Turing Test set the initial standard for artificial intelligence, mused towards the end of his life on the prospects of a truly superintelligent machine: “it seems probable that onc... 010
👖Emancipants👖 @fgshprd.bsky.social · 13/04/2026New cognitive test dropped: - Dr, or - Lord&Savior 000
👖Emancipants👖 @fgshprd.bsky.social · 18/03/2026I grew up with one of these parked in the back yard. Over time a garage grew up around it. 010
👖Emancipants👖 @fgshprd.bsky.social · 15/03/2026This Nature opinion piece, like most opinions about “intelligence”, assumes that “intelligence” — whatever it is — presents itself in “levels” that express a natural hierarchy. Why? 001
👖Emancipants👖 @fgshprd.bsky.social · 14/03/2026This is the wrong distinction. The mafia also acts on principle. T doesn’t understand people whose principle for action isn’t always-be-richer. 000
👖Emancipants👖 @fgshprd.bsky.social · 14/03/2026This distinction is tricky because it undoes itself at every turn. There is a difference but it’s not the one or the kind you want. Maybe it’s like the difference between coloring and color-by-numbers. Wandering an open trail and a walk in the garden. (Ofc you can still walk forever in a garden) 000
Reposted by 👖Emancipants👖David Gunkel @davidgunkel.bsky.social · 06/03/2026The #AI critics are right about this much: #LLMs do not know what they are talking about, cannot understand the meaning of the linguistic tokens they stitch together, and have no access to real world embodied referents. BUT that is not a bug; it's a feature. link.springer.com/article/10.1... 452
👖Emancipants👖 @fgshprd.bsky.social · 07/03/2026It’s a trick question, as the post clearly demonstrates that consciousness doesn’t always work 000
👖Emancipants👖 @fgshprd.bsky.social · 06/03/2026Huh. blogs.newschool.edu/news/2026/03...blogs.newschool.eduThe Center for the American Experience Explores the Actions and Ideas that Shape U.S. History - The New School NewsWhat is “the American experience,” and who gets to define it? Whose actions, ideas, and aspirations have shaped American history? Which responsibilities does this nation hold to the broader world […] 000
👖Emancipants👖 @fgshprd.bsky.social · 05/03/2026The primary development is that the LLMs have been trained, using a variety of techniques, to reliably output strings of text that can be interpreted by the code managing the LLM as a signal to run a command or function. There's no magic here. 000
👖Emancipants👖 @fgshprd.bsky.social · 05/03/2026Given the marketing around "AI agents", there is some confusion about what this new offering actually represents. These are not AI models that suddenly have agency. They are not autonomous and are not approaching autonomy. The "agency" is old-fashioned software interpreting LLM outputs ... 100
👖Emancipants👖 @fgshprd.bsky.social · 05/03/2026Forgot @themountaingoats.bsky.social was on this comp!! 000
👖Emancipants👖 @fgshprd.bsky.social · 04/03/2026It's interesting, if not surprising given the training data, that image generation models struggle to produce images that aren't representational in some way. There is a strong bias toward 3d space, material texture (e.g., paint, fabric, concrete), landscape, etc. 000
👖Emancipants👖 @fgshprd.bsky.social · 03/03/2026“Our time is a time in which the mafia and the oligarchies remorselessly chase out the bourgeoisie-a bourgeoisie that, although it is philistine, is still too cultivated in their eyes." Stiegler 2017 010
👖Emancipants👖 @fgshprd.bsky.social · 02/03/2026hello nice people of the internet. please insert some random code into my systems. what could go wrong? 110
