Reynier the Fox

Reynier the Fox

@lightrey.bsky.social

Generally, being a philosopher means being the most annoying person in the room. I am that room.

123 followers95 following346 posts

Reynier the Fox reposted

Wie vaak demonstreert, kan door de politie worden aangemerkt als extremist. Dat overkwam activist Frank van der Linde. Zijn naam belandde in politiesystemen als mogelijke terrorist. Sindsdien voert hij een eindeloze juridische strijd met de overheid, procedure na procedure.

ftm.nl‘Ik ben tering-irritant’: jarenlange strijd van beroeps­activist legt gebreken rechtsstaat blootSinds activist Frank van der Linde ten onrechte in verband werd gebracht met terrorisme, staat hij voortdurend in de rechtbank. Daar legt hij bloot hoe overheden burgers lichtvaardig verdacht maken en...

The weirdest thing about Hank Green's apology for using AI is that he spends much of it asserting his "nuanced takes" on AI while in the same breath admitting to having essentially an addiction to LLMs. It's like listening to a smoker asserting their "nuanced takes" on the dangers of smoking.

Reynier the Fox reposted

Reynier the Fox reposted

Reynier the Fox reposted

Listened to an NPR story this morning about mass child starvation in Senegal without a single mention of Elon Musk, the guy who gleefully created the conditions for the mass starvation

Reynier the Fox reposted

The Stop Destroying Videogames ECI has sent a powerful democratic signal. An invitation to OPENLY discuss the future relationship between games industry and its customers. The EU Commission must not fall for what seems to be a covert attempt to stop improvements. SKG open letter:

Reynier the Fox reposted

Replying to @zachweinersmith.bsky.social

I love universities. But the incentive structure is badly fucked up. Anyone who teaches has had the experience of teaching a class they don't want to teach to a room of students who don't want to learn. Like with a lot of things Ai is an accelerant, but it didn't create the problem.

Reynier the Fox reposted

New in Nature: LLMs give "the party line" in the languages of authoritarian regimes. This works when they control the media, which feeds pretraining data. We show more state control over the media means less critical LLMs. 6 studies spanning 38 languages & 13 models. Details ↓

This shows that the goal behind dogpiles is purely to hurt people.

Drunk On Mystery, but like....Spooky@drunkonmystery.bsky.social

The response was immediate, bizarre, toxic, and indescribably cruel. One woman just started flooding my feed with pictures of dead babies until I blocked her. Another informed me that heaven was filling up because I was killing so many babies by filling my wife's uterus with my toxic sperm.

Reynier the Fox reposted

Replying to @benjaminaengel.bsky.social

Points to another problem: academic journal publishing needs to die. Academic disciplines are of course different but as a historian, if I'm publishing a paper every two weeks, or even every two months, you know I'm not taking the research seriously. end

Reynier the Fox reposted

Trump & Co have been talking about US military action for weeks, if not months, and positioning the assets to do it - long before such intelligence ever could have become a factor. This is transparent nonsense.

Reynier the Fox reposted

Replying to @ealluia.bsky.social

Likewise, pointing out the involuntary nature of tics, in order to deny mens rea for their harm, does not assert that there is no steps a person can take that might, but also might not, reduce their frequency or severity.

Reynier the Fox reposted

Replying to @thecom.bsky.social

Asserting that is possible that symptoms will resolve is not the same thing as asserting that there is an intervention or collection of interventions that will reliably stop symptoms. There is nothing a person can do that will guarantee they will not have tics, which is the relevant point.

Reynier the Fox reposted

Replying to @ealluia.bsky.social

If someone says hiccups are irresistible and you reply, "Well, actually, have you tried holding your breath?" you are just being obtuse about what the person means.

Turns out that the "checks and balances" of the US government are just some guy standing next to the door leading to the authoritarianism room who might stop you if you try to open the door.

Plato (in the Sophist IIRC) actually addresses the fact that Socratic-type dialogues are not so much for the benefit of the interlocutor, who is often a sophist and too old to be open to the idea that he doesn't know what he thinks he knows, but for the benefit of the (often young) people watching.

This post is unavailable.

My biggest issue with LLMs and similar AI technology is the deception as an inherent quality. LLMs are designed to seem like people. Not *be* like people, just seem like them. They are not an attempt to model or reverse engineer (human) intelligence. They are an attempt to produce similar results.

View full thread

And any idea that being better at feigning intelligence will eventually lead to actual intelligence is just as misguided as the idea that achieving enough skill at cheating on tests will eventually amount to possessing the actual skills the tests are trying to measure.

LLMs are at best a deeply misguided attempt at effectively trying to reach infinity by trying to count high enough and at worst a deliberate attempt to falsely make you believe that is possible. It is a scam technology meant to scam people because it is, at its core, designed to scam.

My biggest issue with LLMs and similar AI technology is the deception as an inherent quality. LLMs are designed to seem like people. Not *be* like people, just seem like them. They are not an attempt to model or reverse engineer (human) intelligence. They are an attempt to produce similar results.

View full thread

You're telling me that the best form of complex algorithms that can specialize to a truly impressive (and within those contexts superhuman) degree is something that happens to function at its core in exactly the kind of way that makes people mistakenly project humanity onto it? Come on.

The purpose of a system is what it does, and what this one does is deceive. No level of added sophistication or improved design is going to change that. It'll just get better at being deceptive. A better deception machine doesn't make an intelligent one, just like a better chess bot didn't.

My biggest issue with LLMs and similar AI technology is the deception as an inherent quality. LLMs are designed to seem like people. Not *be* like people, just seem like them. They are not an attempt to model or reverse engineer (human) intelligence. They are an attempt to produce similar results.

View full thread

It's designed not to be (super)human intelligence in machine form but to mislead people into thinking it is (or will be). It's designed to fool you, in the same way that the effort of copying your friend's homework and changing a few words to pass it off as your own is designed to fool your teacher.

For all its potential actual uses (even ignoring all the downsides), this core quality defeats basically all of them. We are already psychologically prone to anthropomorphize. And now tech bros have conveniently created a machine distinctly not anthropomorphic, designed to be anthropomorphized.

My biggest issue with LLMs and similar AI technology is the deception as an inherent quality. LLMs are designed to seem like people. Not *be* like people, just seem like them. They are not an attempt to model or reverse engineer (human) intelligence. They are an attempt to produce similar results.

View full thread

LLMs are designed to reproduce that look (again, not the method). It creates the appearance of originality in fundamentally the same way as someone who copied their friend's homework and changed a few words. It is just designed to do so much more convincingly (superhumanly, in a way).

Just like computers have surpassed us at chess, they have surpassed us at this. But this thing they have surpassed us at is not intelligence, but the ability to feign intelligence. But it is then, by definition, designed to seem to have surpassed us at intelligence, or at least be on the way there.

My biggest issue with LLMs and similar AI technology is the deception as an inherent quality. LLMs are designed to seem like people. Not *be* like people, just seem like them. They are not an attempt to model or reverse engineer (human) intelligence. They are an attempt to produce similar results.

And they do so by very much supervening on actual such results, innumerably many in fact (way more than any real human conceivably could, or would need to in order to produce such results), so it's not even a spontaneous product. All it does is add the appearance of originality.

Why? Because that makes it look more like it does something on its own. In humans, originality means independent (re)production. An original answer to a question on a test means something that was produced through understanding of the subject; it *looks* original because it's "in your own words."

Reynier the Fox reposted

By the way, for those of you not in MN: Homan’s strategy was to withdraw some agents, so it felt like we were in the denouement of the crisis, and then escalate hard against observers. They’re being incredibly aggressive with observers now. National media may not want to check out just yet

Reynier the Fox reposted

Replying to @jcollins.bsky.social

“If I am killed doing this, throw my body at the White House, martyr the shit out of me & raise hell. Do not be sad. Do not think I would do anything differently. I would do it over & over again — this is too important to sit down & shut up and not do anything.” www.mprnews.org/story/2026/0...

mprnews.orgPursued by federal agents, suburban ICE observers remain resolvedAfter a frightening pursuit by federal agents, suburban ICE observers say fear hasn’t stopped them from protecting their communities.

The problem with the concept of an external world is that it implies an internal one. While it's perfectly valid to say that the thing we might call our internal world is real, any boundary between it and the "external" is completely imagined. These aren't separate worlds at all, but a continuum.

Reynier the Fox reposted

Alex Pretti "used to tell people off when they made sexist comments to female physicians. He bought me coffee when I had a really bad day as an intern...I laughed alongside him daily. He made a point to teach medical residents without judgement, but with a smile on his face and a joke."