iiinova — where your digital self keeps evolving

Matthew Isaac Harvey @acornapocalypse.bsky.social · 08/06/2026

"Things That Probably Shouldn't be a Musical: The Musical" is a distinctive subgenre of musical theater—a strange and precious one. Only consider: en.wikipedia.org/wiki/Hands_o...

en.wikipedia.org

Hands on a Hardbody (musical) - Wikipedia

030

Matthew Isaac Harvey @acornapocalypse.bsky.social · 03/05/2026

Quite interested in what you think on that front :)

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 03/05/2026

Yep, I agree with this. And unfortunately my understanding of LLMs is limited to having messed around with basic neural nets in college—think 80s style connectionism. Got up to CNNs, but nothing within 3 orders of magnitude of modern machine learning.

100

Matthew Isaac Harvey @acornapocalypse.bsky.social · 02/05/2026

The question would be: What happens if the informationally constituted system were to cease acting? If that would kill it—the way a far-from-equillibrium thermodynamic system dissipates if it stops self-organizing—that's a pretty strong clue to interesting emergence.

100

Matthew Isaac Harvey @acornapocalypse.bsky.social · 02/05/2026

It depends who you ask—as originally formulated, yes, but if I'm remembering right, even Maturana was totally open to the idea that these organizational characteristics would appear in other substrates than the biological. I know there were digital models of autopoiesis in the 80s.

310

Matthew Isaac Harvey @acornapocalypse.bsky.social · 16/03/2026

Sorry if my replies seem terse, it's just hard to tell whether you're looking for a discussion or a fight—and microblogging doesn't really allow for serious discussion. Autopoiesis is Maturana's theory (googleable), refined by Varela and developed by Di Paolo, Froese, De Jaegher, and others.

200

Matthew Isaac Harvey @acornapocalypse.bsky.social · 16/03/2026

The similarity: Shoes and LLMS are both inert with respect to material self-creation. The background: If you actually want to dig into this subfield of cognitive science, start with either Evan Thompson's Mind in Life (2007) or Varela, Thompson & Rosch's The Embodied Mind (1992).

500

Matthew Isaac Harvey @acornapocalypse.bsky.social · 16/03/2026

Well, look, I hear what you're saying. I know that AI can feel REALLY impressive, the mystery of the black box combined with fancy math creating a feeling of wonder. I get that. It's also straightforwardly irrelevant to what I was talking about—a specific systems-theoretic model of agency.

101

Reposted by Matthew Isaac Harvey

Brett Karlan @oldjerryfodor.bsky.social · 09/03/2026

In an LLM world, this kind of pure, uncut human slop hits me harder than the goddamn Brandenburg Concertos. It's just so beautiful

1264

Matthew Isaac Harvey @acornapocalypse.bsky.social · 10/03/2026

Well, one way to think about it is that you're still prepped, you're just relying on prep from a few years ago. So it's a little hazier—but your attention is more on social cues and less on notes, maybe, and there's more airtime for questions. Classroom attention is super interesting

010

Reposted by Matthew Isaac Harvey

Katie Mack @astrokatie.com · 07/02/2026

GM: Charisma check. Mamdani: [rolls natural 20] GM: that’s a d6 how did you Mamdani: [direct to camera] Did you know you can check out board games at your local public library? 😊

222334636962

Matthew Isaac Harvey @acornapocalypse.bsky.social · 05/02/2026

Have you seen his Macbeth combo monologues from back when he was young? He had it then, too, and the sort of arrogant delight of the youthful virtuoso is fun to watch.

020

Matthew Isaac Harvey @acornapocalypse.bsky.social · 05/02/2026

New article up! AI skeptics and AI enthusiasts—even those who consider themselves reasonable, grounded members of both camps—can easily deter engagement through overstatement or overzealous claims. medium.com/@acornapocal... #ai #writingadvice #contentwriting #persuasino #rhetoric

medium.com

Arguing Past Each Other

AI Haters, AI Enjoyers, and the Humility of Truth

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/01/2026

This is the kind of thing that BlueSky is for

000

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/01/2026

I think the Oxford 4E Handbook is quite good! academic.oup.com/edited-volum... It's an anthology rather than a fully coherent discussion, but still great, especially for students. I also think Sune, Rasmus and I did a decent job on enactivism/interactivity: constructivist.info/11/2/234.har...

010

Reposted by Matthew Isaac Harvey

Admirable Women @admirablewomen.bsky.social · 21/01/2026

DYK a language dies every ~14 days & more than half of the world's 7,000 languages could vanish by 2100. The Eyak language went extinct #OTD in 2008 when Chief Marie Smith Jones of Alaska's Eyak tribe & the last fluent speaker of Eyak died at the age of 89. #indigenous #language #linguistics (1/2)

An intimate indoor portrait of an elderly Indigenous Alaskan woman, Chief Marie Smith Jones, in her late 80s, seated and reclining slightly in a comfortable red upholstered armchair or recliner. She has short gray hair and wears large, round, gold-framed eyeglasses with thick lenses that reflect light. Her expression is thoughtful and somber, with a slight downward gaze and lips closed in quiet reflection or contemplation. She is dressed in a bright red long-sleeved top or shirt, evoking warmth and cultural pride. Directly behind her head, mounted high on the wood-paneled or warmly toned wall, hangs a striking traditional fur mask or ceremonial headdress resembling a bearded human-like face (possibly representing an ancestral or spiritual figure from Eyak or broader Alaskan Indigenous traditions). The mask has thick, shaggy brown fur framing a carved wooden face with prominent dark eyes, eyebrows, and a subtle smiling mouth, creating a powerful cultural backdrop that contrasts with her serene presence. The lighting is soft and natural, casting gentle shadows that highlight the textures of the fur mask, the red fabric of her clothing, and the rich tones of the room—possibly her Alaska home. The overall mood is one of quiet dignity, resilience, and profound cultural significance, capturing a moment that honors her role as the final guardian of the Eyak language and heritage before its passing with her in 2008.

13416

Reposted by Matthew Isaac Harvey

Naomi Kritzer @naomikritzer.bsky.social · 19/01/2026

TWIN CITIES FOLKS WHO WANT TO HELP AND DON'T KNOW WHERE TO START: I put together a big list of resources. Do you want to deliver groceries? do people's laundry? work in community defense? I want to help you get connected! naomikritzer.com/2026/01/19/h...

naomikritzer.com

How to Help: Twin Cities Residents

This is a post for people in Minnesota. I’m going to do a separate post for the non-Minnesotans when I’m done with this one. (If you’re in a hurry you can go to Stand With Minneso…

13682396

Reposted by Matthew Isaac Harvey

Complexity Cat 🐱 @amahury.bsky.social · 21/01/2026

We've extended the deadline for #NERCCS2026 ! Don’t miss it! ✅More affordable registration fee compared to other conferences ✅World-class researchers as keynote speakers ✅Conference proceedings published in a diamond open access journal 👉Submit here: shorturl.at/52o0q

033

Matthew Isaac Harvey @acornapocalypse.bsky.social · 26/12/2025

Would absolutely love to hear (a) a couple of your favorite ideas that you discuss in the book and (b) what you make of Olson's World on Paper, which I've always taken to be maybe the best discussion of print technology in relation to language, and (c) whether there's a preorder.

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 25/12/2025

This is why LLM-generated fake citations are a problem. AI as it currently exists just can't be used for knowledge work. The infrastructure we have doesn't safeguard against the problems it creates.

000

Matthew Isaac Harvey @acornapocalypse.bsky.social · 25/12/2025

Maybe consider school boards and school board associations? I don't know much about this, but I do know they maintain big mailing lists and active admin staff. There must be recommended reading and etc. Also school library associations.

020

Matthew Isaac Harvey @acornapocalypse.bsky.social · 25/12/2025

Like a lot of folks, I'm culturally Christian but not practicing. To me, that means embracing the message of love, acceptance, and universal human worthiness that we associate with Christ. Wise words from our malört pontifex. Merry Christmas 🎄❤️

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 25/12/2025

Attention Writers! Hot new tip. Next holiday season, ask for the gift of nunchuck

000

Matthew Isaac Harvey @acornapocalypse.bsky.social · 25/12/2025

I misplaced my chainmail veil and cowl last year 😔 been really struggling to write ever since.

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 23/12/2025

I understand that allocating grant money is complicated. Equity, access, prioritization, practical questions and fundamental research. I get that it's a hard, collective problem. But holy cow, we are NOT crushing it.

0132

Matthew Isaac Harvey @acornapocalypse.bsky.social · 23/12/2025

Surely the alternatives aren't "rely on market-based alternatives to provide healthy choices" vs "wail into the void." There's also "regulate this obviously harmful and unsafe technology." That isn't going to happen because our government does not serve the people, at present, but it is the answer.

000

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Cool! Thank you. I've written down a bunch of keywords, phrases, and concepts to look up and read about. Off to read your stochastic parrots piece — and thanks again for posting! I hope BlueSky cybernetics continues to grow :)

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Conclusion: Or I'm wrong due to lack of knowledge about LLM engineering. I'm very open to the idea that LLMs DO cohere over time in some emergent and important way, as you started to explain. Also FWIW Tom Froese thinks they might be cognitive agents, and he invented half these concepts, so...

120

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

E.g. yeah, LLMs are sensitive to the context of the conversation. But only for the duration of that one conversation, then the dependency disappears. The only enduring components are inert + don't self-organize (server racks, etc.). You can turn it off and on and have the same system. No so w/life.

210

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

connection and do not rely on one another for their existence. So where biological systems have observable emergent properties, LLMs do not (as far as I'm aware, and maybe excepting scaling laws?). That's why I don't think they're precarious.

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Could LLMs work the same way? Maybe. The dependencies you point out are meaningful. But whether they cohere into agency depends on whether they have some kind of stable operational (f-r) grounding. For humans that's provided by metabolism; LLMs' constituent servers and code bases no inherent 1/2

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Humans definitely aren't operationally closed, but we manage our precariousness using systems that ARE (cells, circulatory systems, some tissues, etc.). And then we use other sorts of temporary coupling (conversations, enskillment, entrainment) to exercise new forms of agency.

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Yes, fantastic points. I agree about spatial boundedness (sufficient but not necessary for autonomy), but not operational boundedness (which IMO is a functional-relational criterion rather than a dualistic claim about substrates): constructivist.info/11/2/234.har...

constructivist.info

Interactivity and Enaction in Human Cognition - Harvey M. I., Gahrn-Andersen R. & Steffensen S. V., 2016

Context: Distributed language and interactivity are central members of a set of concepts that are rapidly developing into rigorous, exciting additions to 4E cognitive science. Because they share â¦

120

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

/thread lmao sorry for this 47-bleat chain. I loved the piece, I think you've identified a real contradiction in the examples you give, and I'd love to hear your thoughts on the LLM-as-agent issue, which IMO is a huge, unresolved question in cog sci.

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Like, cells define themselves in space and time through their own activity — that's autopoiesis. LLM's don't exist as coherent spatial, mechanical, information-theoretic, or any other type of coherent unity. What actual entity are we even attributing precariousness to?

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

sorry, *f-r. In the blackmail scenarios, testers wrote a fake office drama and the LLM performed its assigned role. There's no f-r analysis at all, and the narration just assumes that LLMs function as autonomous sense-makers with rich, humanlike internal worlds. Are LLM's even Maturanian "unities"?

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

I'm also super curious about the AI-as-precarious idea! Like extremely curious. What're the "functional-relational expressions of distress"? I only see surface-level printout-style evidence. An f-m argument would need to be about system architecture, resting states, energy flow, etc., no?

100

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

IMO the core enactive claim is that metabolism is the only mechanism that reliably meets the criteria of autopiesis, but if an artificial system meets the conditions for precarity we SHOULD take that seriously as evidence of autonomy. E.g. www.sciencedirect.com/science/arti...

sciencedirect.com

Enactive artificial intelligence: Investigating the systemic organization of life and mind

The embodied and situated approach to artificial intelligence (AI) has matured and become a viable alternative to traditional computationalist approac…

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

Great post! Am I right that the crux here is the necessary/sufficient problem? Basically: While metabolism is sufficient for cognition it isn't necessary, b/c you can get precariousness without it? I think most enactivists actually agree, and the problem here is more sloppy phrasing than error.

110

Matthew Isaac Harvey @acornapocalypse.bsky.social · 21/12/2025

It absolutely is not. Information needs context and provenance to be meaningful. AI removes both — and anything it prints can always be arbitrarily wrong.

1401

Matthew Isaac Harvey @acornapocalypse.bsky.social · 20/12/2025

Seven insurmountable problems with AI: 1. Cost/economic waste 2. Water and land 3. Energy 4. IP theft and training data shortages 5. Bias and social violence 6. Hallucination 7. Damage to learning (esp higher ed) and loss of professional expertise I suppose citations fall under 6, really.

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 20/12/2025

The damage from this is tremendous. Citation chains are fragile and flawed, but they're the connective tissue of our epistemic system. Every fake citation is a broken thread in a network that only functions when it's widely, richly connected. We have an 8th intractable problem with AI.

120

Matthew Isaac Harvey @acornapocalypse.bsky.social · 18/12/2025

My own little guy came from Where the Love Is! I was in Guilford, CT at the time. I'd recommend them strongly to anyone nearby — they rely heavily on volunteers and, at least in Kiwi's case, rescued a litter of stray puppies from a ranch that couldn't keep 'em. Where's your pet from?

A quizzical dog leans over the back of a wide green armchair, his nose black and his chest badged white his floppy ears are framed by the bright blue sky out a bay window.

111

Matthew Isaac Harvey @acornapocalypse.bsky.social · 18/12/2025

I thought about this for a minute and concluded that any answer I gave would just be a guess. I do know that literal lists of words with high collocation frequency have been called representations of a given word's meaning. I figured we were talking about something similar here but could be wrong!

010

Matthew Isaac Harvey @acornapocalypse.bsky.social · 18/12/2025

That's definitely true, good point. And tbh I should just admit that my beef here is with the idea that semantics can be captured in text — I think this fundamentally mischaracterizes language, which is temporal, interactive in real time, organic, experiential, and of course multimodal.

120