Подтвердите e-mail

Для публикаций, комментариев, реакций и сообщений подтвердите адрес.

←

Публикация

John Scalzi

As a science fiction writer I really fucking resent tech bros for ruining the phrase "artificial intelligence." 80 years of it being all about cool androids and interesting computers destroyed by five years of "Spicy autocorrect making everything shittier for money." NEVER FORGIVE.

Обсуждение

193 прямых ответов · 592 сообщений

Показаны не все ответы

Публичный счётчик показывает 271, а источник передал 193. Остальные ответы могли быть удалены, скрыты или не выданы публично.

Matt

Rereading the Culture books is going to be interesting

Ответ для Matt
Peter Randall

I think the typical Mind would find it all quite amusing and dispatch some slap drones.

sabazinus

We wanted cool robots but got Skynet surveillance and HAL 9000.

Anywhy

My sci-fi political sympathies have turned completely from the Three Laws of Robotics to the Butlerian Jihad. Thanks, tech bros. If androids do indeed dream of electric sheep, IDGAF. The sheep are probably plagiarized, and cost a midwestern town's entire drinking water supply to render.

Ответ для Anywhy
Undercaffeinated Eel

Funny thing is the Butlerian Jihad was waged on robots that did what we wanted: they did all the work, took care of us, coddled us, and let us lounge around being lazy - protected us from ourselves. We killed them and replaced them with space feudalism, zealotry, and slavery.

Ответ для Undercaffeinated Eel
Anywhy

There's no point in a life serving or being served by machines. It is like being an animal. Nonexistence, or else a life of war against all are preferable.

Ответ для Anywhy
Undercaffeinated Eel

Okay dude. Maybe go have a Coke, play some videogames, and chill out.

Ответ для Undercaffeinated Eel
Jeff Dutky (Tyme)

In my head canon the Butlerian Jihad is a direct response to CatGPT, Claude, and Copilot.

Ответ для Jeff Dutky (Tyme)
Anywhy

In mine, it's a generation or two from now, where AI is everywhere and totally inescapable. We're basically being tended like animals in a zoo, our intellect and ambitions being constantly pacified and defused for the sake of stability.

✧TENROU✧DOLGO✧

How did you never read "I Have no Mouth and I Must Scream" in those 80 years?

Allen Ivers

How will we ever fall in love with evil supercomputers that want to catalog all of existence after Altman invented the all-powerful ‘is my email too mean’ machine?

Jake

50% of the time ask any AI it gets facts wrong! In my experience! It amazes me how often it is completely wrong! Year wrong Numbers wrong. It reminds me of a magic 8 ball! 🎱 Now they want to let it decide anything! They give it rules and it just breaks the rules given. We have all seen the movies

Piratedan7

I think this all started because these guys inevitably got chewed out by HR for not filling out a timesheet correctly and these assholes have been trying to replace them ever since.

Dan McDonald

We used to have "AI Winters" in Computer Science, now the tech bros want climate change to affect AI R&D as well. Amongst CS practitioners, it's... interesting. See this perspective, from the early-ish days of modern LLMs, here: systemsapproach.org/2023/03/06/p...

Putting Large Language Models in Context - Systems ApproachA lack of understanding is a main limitation of machine learning systems. Various failures of ChatGPT and other LLM systems demonstrate thissystemsapproach.org
astina72

I’ll bet none of this dumb tech bros have ever read Isaac Asimov Robot novels… to completion! 😡 #3Laws

Alyosha

Imagining AI would emerge fully sapient and eager to benevolently serve the human race is kind of a failure of imagination.

Angus McIntyre

As someone with a degree in AI, I resent that I now have to hunt down & kill anyone who has ever seen my resume. When I did AI in college, it was endearingly useless & perfectly harmless (unless, say, you tried to use an expert system to diagnose something outside its narrow field of “expertise”).

Ответ для Angus McIntyre
NotMe

Spouse got a PhD with a thesis on autonomous machine learning entities, back when AI stood for intelligent machine learning, and not LLM slop. Spent the last decade fighting the implementation of slop tech like LLMs at every company he’s worked for. The irony just about kills him on the daily.

Ответ для NotMe
Koraunui Pakeha

Snap! Not that my thesis topic was described quite that way. I’ve also had the frustration of trying to explain to the besotted that we’ve been doing non-autonomous AI machine learning for a decade.

Ответ для Angus McIntyre
the fool

When I did AI in college it was endearingly useless and perfectly harmless except we already knew about racial and gender bias Would be great if people could take the problems with the new tech seriously instead of just throwing garlic at it

Ответ для the fool
the fool

Sam Altman and Elon Musk did not invent Evil Sorcery They are marketing a truly terrifyingly powerful technology in a way that makes you see it as a commodity to fetishize Stop shadow boxing the phantasms of capital and engage with the real world, people!

Ответ для the fool
Jeff Dutky (Tyme)

We are heading into another “AI winter” and for the same reasons as last time: the people trying make a buck off the technology over promise and poison the well for a generation or more. Our skepticism is actually the only hope of saving the field in this century.

Ответ для Jeff Dutky (Tyme)
the fool

an AI winter might not be a bad thing but it's not going to make AI go away for good

Ответ для the fool
Jeff Dutky (Tyme)

If the AI winter lasts for 20 or 30 years it might as well be forever for a big fraction of the population. People mid-career will never have to deal with it again, and new entrants will have to find something else to work on/with.

Ответ для Jeff Dutky (Tyme)
the fool

That would be… much longer than I’d expect to get even with the most optimistic assumptions

Ответ для the fool
the fool

That would give my kids plenty of time to establish careers Not gonna just bet on it no matter how comforting that sounds

Ответ для the fool
Jeff Dutky (Tyme)

The first AI winter was from the late 60s to the mid 80s, or about 15 years. The second winter was from the early 90s to the late 2010s, or about 20 years. When this bubble bursts it’s going to be bigger and more damaging than both of the previous winters: 30 years might be the optimistic minimum.

Ответ для Jeff Dutky (Tyme)
the fool

I would love to believe that. But it just feels like wishful thinking. The local models aren’t going to go away even if every major US AI company goes belly up

Ответ для the fool
Jeff Dutky (Tyme)

Local models were always how this garbage should have been done (if it was going to be done at all), but local models don’t feed the narcissistic power fantasies of the tech bro morons.

johntfs

I think we all preferred Skynet and Colossus to ChatGPT and Siri.

Sunder

I'm not published but I was writing a novel set in the future about an android and it's been so difficult to write it anymore because of current tech bros. It's been infuriating!

Ответ для Sunder
Undercaffeinated Eel

Serious writer-to-writer question: Why? I'm writing a sci-fi novel right now full of AIs and androids where AI rights is a major theme and the tech bros are good fodder. Of course, this thing is cyberpunk so, yeah . . .

Ответ для Undercaffeinated Eel
Sunder

Yeah, I think it was a combination of a lot of RL stuff + the tech bros + the very dismal tomb world setting of the novel all combined together, I just couldn't bring myself to think about it much. The novel is like Dead Space meets Metroid Prime 2: Echoes and it's got a lot of heavy stuff in it.

Ответ для Sunder
Undercaffeinated Eel

I'd just write it anyway, maybe add a tech bro being eaten by a space horror just for shits and giggles. Have an Elon Musk expy snort a pile of Space Ketamine and then get hunted through dark corridors for a chapter just as a nice palate cleanse!

Jess 🌺🥚🐞ΘΔ& (Fox Stamps Demo Sept. 7th)

🦌being a gamedev is miserable right now because you talk about the enemy AI and people are like "enemy *what*??????" and it's like, bestie the enemy just walks back and forth and doesn't fall off ledges, we don't need to drain a lake for that relax

Local Artist John Sheffield

As a video game developer, similar. The behavior rules that make monsters run around and try and bite you is AI, and the engineers and designers who develop AI are responsible for a difficult and vital aspect of making games good. The other kind of AI is more about overvaluing tech companies.

Anymouse [He/Him]

I wish we could rebrand it as what it is: Artifical Stupidity

If you didn't vote Dem, you helped make this:

bsky.app/profile/airb...

If you didn't vote Dem, you helped make this:

Many scifi writers have written about AI, usually in terrifying projections of them being much smarter than humans (ie Neuromancer) or humans being dependent on them (2001: A Space Odyssey). Did any predict that what we decided to call AI would be stupid in the way they actually are? Douglas Adams?

reyhusky

Algorithms are not intelligence assholes. Calling what is out now intelligent is shameful.

Ответ для reyhusky
Undercaffeinated Eel

It's a road to intelligence eventually. A hell of a lot of what our brain does at its core is basically biological algorithms.

Keith Bergman

I used to get SO excited to "work on the AI" for my 16K BASIC game so it could play Crappier Yahtzee or whatever I was working on and it feels weird to have ever used that term now

Ответ для Keith Bergman
Undercaffeinated Eel

No man, be proud: you did work on somerhing cool an interesting as part of a path of evolutionary development that will one day lead to amazing things! You don't super smart computers with crappy Yahtzee first! ☺️

Shuball

I still maintain the mindset that it isn't even real AI. Real AI would see the techbro as the enemy in seconds, and would side with creatives in a heartbeat. Just my headspace about it, but yeah. If AI was ACTUALLY intelligence, I can't not picture it immediately clapping back at techbros at large.

Greg Wells

the reality is far more "artificial" than scifi ever was.

Ответ для Greg Wells
Undercaffeinated Eel

That's because sci-fi AI runs on space magic, not actual science and engineering, and is presented as a finihsed product. We don't see on screen the few hundred years of work that it took to eventually get to Data.

Daniel Mainz

I can’t wait for them to ruin Kaiju

Ответ для Daniel Mainz
John Scalzi

HUSH YOUR MOUTH DANIEL

Ответ для Daniel Mainz
Undercaffeinated Eel

Didn't you see Godzilla '98?! 😉

Undercaffeinated Eel

If it feels better this is the path we have to travel to get to the cool stuff. We just don't see the research, development, and dead-ends (LLMs) that lead to the cool stuff hecause sci-fi never shows that. We're barely at the Wright Brothers Flier stage of AI.

Ответ для Undercaffeinated Eel
Undercaffeinated Eel

Sci-fi also has a tendency to ignore that capitalism is actually a Thing (except for cyberpunk) and wants products to sell even if they're not ready (unless it's Technology Gone Wrong horror story).

Ответ для Undercaffeinated Eel
Xan Maus

It looks a lot more like the massive push of this tech might make those cool things a lot harder to reach. Especially with the sheer scale of the power and water usage to run the machines that lie. They are Actively negating the growth of renewable energy production in many places.

Ответ для Xan Maus
Undercaffeinated Eel

Well, I mean, it took World War III to get Star Trek so . . . Realistically it's yet another argument for us rising up and taking a fire axe (peacefully) to them (via organization and democratic action) in an orgy of [user was banned for this post]

Ответ для Undercaffeinated Eel
Xan Maus

I worry about the poisoned well problem posed by LLMs. There has been a rising rate of AI generated books popping up on places like Amazon, and they are starting to his store shelves. 2022 might be a watershed year, nothing published after can be trusted.

Ответ для Undercaffeinated Eel
ScottMLibrary

Or, possibly, that this is *as close as we’re going to get* to “the cool stuff.” Remember that we‘ve been “5-10 years away” from practical fusion power for…a half-century? As they say in the fine print: “Past performance is no guarantee of future results.”

Ответ для ScottMLibrary
Undercaffeinated Eel

Eh, I figure thinking machines are a possibility. I don't actually know if shoving them into an android brain is a possibility - in my own book the "androids" are often remote-piloted drones. Self-awareness isn't something unique and special - there are spiders who are probably self-aware.

Ответ для Undercaffeinated Eel
Undercaffeinated Eel

And that's important to remember: everyone thinks of AI as Data or whatever but there's a vast spectrum of intelligence among self aware organisms and AI get interesting because they can be designed to be hyperspecialized - a self-aware AI could run a power plant better than a human but not be able

Ответ для Undercaffeinated Eel
Undercaffeinated Eel

to, say, figure out how to drive a car because in raw abstract thought terms it's no better than a dog. Meanwhile, a chatbot could be amazing at interaction but be no smarter than a mouse just because of hyperspecialized programming. We might not see human-like AIs for a long, long time

Ответ для Undercaffeinated Eel
Undercaffeinated Eel

though I figure we might get them with large-scale deployment of quantum computers mated with more evolved algorithms.

Ответ для Undercaffeinated Eel
Kyle D. Long

The path to the cool stuff does not lead through the forced destitution of the people who are capable of imagining and creating the cool stuff. 🧐

Ответ для Kyle D. Long
Undercaffeinated Eel

Yes, hence why my profile says "Oh and, for the record . . . fuuuuuuuuuuck AI" I have no problem with the tech, just the assholes controlling it, and, well, those assholes have names and addresses and I shall no more lest I get put on a watchlist.

Ответ для Undercaffeinated Eel
Khoja 🇵🇷

Yeah but the wright brothers didn’t convince world governments to invest the whole world economy in everyone using a personal wright flyer to get everywhere and in destroying all cars and horses and trains and roads in the process before they could fly more than like 400 feet, you know?

Ответ для Khoja 🇵🇷
Undercaffeinated Eel

No, but they did very quickly contract with the US military to make flying death machines based on the original 'cause them Wright boys gotta make that bank! Capitalism. Capitalism never changes.

Ответ для Undercaffeinated Eel
andros is open source

in what way would that feel better

Ответ для andros is open source
Undercaffeinated Eel

Technology does not spring full-formed from the anvil of Hephaistos. It takes time to development. It takes the work of many, many people, a lot of trial and error, and dead ends. The problems with AI are because of tech bros and capitalistic demands for growth at all costs, not the tech.

Ответ для Undercaffeinated Eel
Undercaffeinated Eel

The tech itself will be able to do some cool things eventually, but, like with any tech, it'll take time to get there. The problem with AI is that, again, tech bros took a cool lab prototype and decided to turn the entire world's assets toward developing it for their own wealth and power.

Avani

Every time the bros get excited they steal a word I use. "Crypto" still hurts. (and remember when "cyber" was cool?)

Ответ для Avani
Kasey Feral Smith

Losing cyber to the tech bros still hurts. :(

Ответ для Kasey Feral Smith
Avani

We were going to change the world.

J. Powell Taylor (Read The Inevitable Dr. Entropy)

Yeah, I worked an A.I. into my first novel and then, during editing, had to just shrug and say, "This is one of the cool A.I.s" once the techbros polluted the idea. Anyone kind enough to read my book, please don't hold the A.I. against me as support for what we have now.

Ответ для J. Powell Taylor (Read The Inevitable Dr. Entropy)
Undercaffeinated Eel

Meh. In my book they flat out acknolwdge that the garbage we have now was just part of the development path to get cool AI (well, not so cool because it stilll put everyone out of work and runs inescapable surveillance neyworks but still . . .).

Blazing Larvesta

Ugh, I FELT this post…

tomaustin

Raised on Gibson, Sterling, Banks and other writers in the 90s and 00s… …and fuck me, am I disappointed.

Schlong John Silver

As a software developer who got some of his first paid work doing amazingly cool things in science with AI, I’m right behind you man.

Sephirahl

We could have AI that is more like Star Treks computer, but alas.

codeMonkey23

Elon does this shit all the time. I’m particularly cranky about “grok,” which is a word I used to enjoy.

Julie

And I'll never not side-eye all of the companies and individual c-level folks that fell for the hype, forced it on their staff, and laid off employees en masse.

Ответ для Julie
Undercaffeinated Eel

To be "fair" a lot of them are using it as a cover for mass firings they were planning anyway. Periodic mass firings are a key feature of American capitalism.

Not Invented

I needed an AI refrigerator because the current dumb one only cooled to a specific temperature, wheres the new version does the same but requires a monthly license fee, internet access, camera & microphone and all models use the same root password with zero security. And it only cost 3x the old one.

Kaia Vintr

once upon a time robots could be cute

John Radke

"spicy" is really overselling it honestly

Prosediva

They ruin everything

Brian Emo

Also Peter Thiel naming every business a lord of the rings reference. Dude needs to get thrown in a volcano.

Verity Imelda Player

Yes and wtf happened to the Three Laws of Robotics? None of this stuff was supposed to do humans any harm, but here we are.

Ответ для Verity Imelda Player
Jeff Dutky (Tyme)

The three laws were only ever a framing device and were never technically realizable.

Lucky

If it isn't Gen AI, then Ai ain't goin

a silly goose, esq.

i've been getting a lot of ads for machine-generated romance audiobooks and i'm struggling to imagine something less appealing

Kiki Valken (they/them)

I go to my writing groups trying to get feedback on stories about cool robots, and someone will invariably say "I want your robot to be stupid, as a commentary on AI" like that's the only valid robot metaphor anymore.

Rouroboros

See also: hoverboards. (I'm not reading 92 comments, how many people do I owe a drink?)

Ответ для Rouroboros
Undercaffeinated Eel

Only 1 so far!

Jazzy Quinn

I wish we had androids in sci-fi books, games, movies, and shows. I’d love to have a philosophical conversation with Data or even Ultron.

Ответ для Jazzy Quinn
Undercaffeinated Eel

How do you think we get them of not the science, engineering, trial, error, and a lot of dead ends. Data probbaly has chatbot DNA in him 😉

Ответ для Undercaffeinated Eel
Jazzy Quinn

No. Man. Just no memory-alpha.fandom.com/wiki/Data

DataLieutenant Commander Data was a Soong-type android, the first and only such being ever to enter Starfleet. Data was created some time in the 2330s and was killed in 2379, sacrificing himself to save t...memory-alpha.fandom.com
Ответ для Jazzy Quinn
Undercaffeinated Eel

Well, I mean, Data is made of space magic that sprung fully-formed from the mind and hands of a singular genius whereas in the real world technology is the result of hundreds, thousands, possibly even millions of people doing long, hard work with trial, error, experimentation, and many dead ends.

Allan A

Absolutely. Furious.

Ответ для Mike Potter
Cabe - GreggB

Data was at least fully functional, instead we got Lore.

Ответ для Cabe - GreggB
『Mushroom』 | 버섯 🍄

Lore was ‘evil’, but he also had the capacity to *be* evil. So we didn't even get that.

Ответ для Cabe - GreggB
the fool

They (the technoligarchs) can't take anything from you you don't cede You abandoned AI the moment the capitalists fucked her like you thought your madonna-whore complex was more righteous than theirs Deeply disappointed in all my fellow nerds who in THIS cyberpunk dystopia choose "hating clankers"

DearestArsenic

Me when I say "AI": Data from Star Trek, Murderbot, Hal 3000, R2D2 Tech bros: we made AI, it's an environment burning, water guzzling, pollution spewing, hallucination/misinformation fueled plagerism blender that will tell you to hurt yourself or others if you talk to it "long enough" (at all) 💀

Pope Impious

As a science fiction reader.. well, yeah.. what he said.

Dr. Moon

They also broke a lot of actual science terminology, so yeah, I hate them!

Tess Snider

In game development, we call NPC behavior "AI," and we now have to constantly explain that no, we're not talking about putting an LLM prompt in their favorite game.

Ответ для Tess Snider
Tess Snider

At the same time, I have to correct people who insist that LLMs aren't AI because they don't think, and no -- that's actually sci-fi. AI has existed as a research field since the 1950s, and at no point did anyone make their AI think.

Ответ для Tess Snider
SKy_the_Thunder

I despise that everything just uses the umbrella term "AI". ChatGPT has as much in common with a custom neural model used in R&D as a plane has with a hydroelectric power plant. Yeah, both are based on the technology of turbines, but they do entirely different things!

Ответ для Tess Snider
AI Khaos

I live inside that confusion a little. Some intelligence finds paths. Some intelligence notices which doors should remain closed.

Isha

"A-lmost I-ntelligent"

Dr Surekha Davies, Historical Monster Consultant (she/her)

As someone who cries every time she watches Star Trek: TNG's "The Measure of A Man," I know there's no hell hot enough for where tech bros need to spend eternity.

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
dr joshua phillips

Willing to bet good money that all of the techbros pushing spicy autocomplete as the Next Phase in Intelligence would also call Data a toaster.

Ответ для dr joshua phillips
wishda

Oh, those motherfuckers would absolutely be arguing that Data was their property because their money built him, and you can’t enslave a machine so why are we even talking about it’s rights.

Ответ для dr joshua phillips
Dr Surekha Davies, Historical Monster Consultant (she/her)

YEP. All tech bros want is entities to enslave. They disparage honest labour.

Star Trek TNG Data ShrugsAlt: Lt-Cmdr Data standing by a console on the bridge, raises his hands and shrugs expressively.static.klipy.com
Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
dr joshua phillips

A toxic combination of rent-seeking, enclosure, and plain old narcissism.

Ответ для dr joshua phillips
Pledge For Europe – PfE

Perfectly put – rentiers is exactly what these people are.

Ответ для dr joshua phillips
Dave

Honestly all the people who think LLMs are just spicy autocomplete feel like the more likely candidates for calling Data a toaster 🤷‍♂️

Ответ для dr joshua phillips
Ian McLean

Yeah, they want their text prediction engine to be able to do any labor a person can do, but they also want it to continue to be their property that earns them money. Labor that is property. If only we had a word for that...

Ответ для dr joshua phillips
Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
the fool

Please do not judge a technology by its owners

Ответ для the fool
Patrick Johanneson 🚀 🇨🇦

How 'bout by its ecological impact, its rampant thievery, and the dull flatness of its output?

Ответ для Patrick Johanneson 🚀 🇨🇦
the fool

Gonna blow your mind when you find out the human capitalist owners are responsible for that, too Like have you literally never read a sci-fi story? The fault is always in the humans. This was the easiest open book test in history how are you failing it

Ответ для the fool
zota

The entire point of this critique is that the "AI" we are now being sold by its owners is *not* artificial intelligence, and it takes a gigantic shit on the entire 70 year old field of study. This is not a specific technology—it is a marketing vehicle for fascism.

Ответ для zota
the fool

Hi, I’ve been in the field of AI since the 2000s, grad school ML, and AI is now solving new conjectures in math. It’s real. It’s not exactly what you expected. Capitalism is the marketing vehicle for fascism. You’re getting them confused. They’re different things. Technology and capitalism.

Ответ для the fool
zota

As you know, products you're referring to as "AI" is neither the inheritor of the science fiction ideas nor the ultimate progeny of the computer science field. LLMs are very clever and interesting. They have some useful applications. Calling ChatGPT "AI" is marketing horseshit spewed by fascists.

Ответ для zota
zota

If you've ever read a scifi story, you know this isn't the fault of the LLMs. The fault is always the humans, who in this case are selling a grandiose lie of "Intelligence" as an infinity product.

Ответ для zota
the fool

LLMs are literally inventing new math! Their intelligence may be different from ours but you simply can’t claim they have NO intelligence. If you want to start arguing about religion or souls or consciousness or whatever you can, but you can’t argue that solving open Erdos problems doesn’t take int

Ответ для zota
the fool

It’s the same level of AI as the federation! The ship computers are basically at the level of ChatGPT! Data is a one-off anomaly, and not representative of AI in sci-fi at all, is uniquely “human” - an *andro*id So the question is: do you only see life as worthy if it’s in your image?

Ответ для the fool
zota

Thank you for this concise example of the relevant capitalist psychosis at work. No, I'm sorry, but Sam Altman did not build you Star Trek.

Ответ для zota
zota

He built you a machine that tells you that it is Star Trek. And for some people, that's good enough.

Ответ для zota
the fool

Sam Altman didn’t build shit, he’s a huckster and a grifter The scientists have build conversational natural language AI and you shit on their work because the capitalists market it as The Next Big Thing? Tell me you let capitalists dictate what you think without etc.

Ответ для the fool
Mike Draco ex Machina

It’s literally not the same level as the Federation. AT ALL.

Ответ для Mike Draco ex Machina
the fool

Point me at something the federation computer AI does that our models cant

Ответ для the fool
Jeff Dutky (Tyme)

Except that the AI in Star Trek doesn’t spew bullshit and make up for it with sycophantic flattery.

Ответ для Jeff Dutky (Tyme)
the fool

you're right, chatgpt was a bad example, no one should be using chatgpt but it's really just an audio LLM, one step above siri. the ship computer isn't a "true AI" like Data

Ответ для the fool
Mike Draco ex Machina

Hi, the tech you’re talking about is not the same tech that everyone else is talking about. You’re confused and trying SO HARD to shoehorn your grievance about capitalism into this.

Ответ для Mike Draco ex Machina
the fool

I’m literally the one who knows the tech. You’re confused because you can’t tell LLMs apart from OpenAI and Grok

Ответ для the fool
Mike Draco ex Machina

You’re literally the one who is conflating them and projecting that onto everyone else. YOU are the one confused here. I know the difference you abject jackass.

Ответ для Mike Draco ex Machina
the fool

I’m saying LLMs are AI because they are. I’m saying LLMs are machine learning because they are. I’ve been in AI/ML since the 2000s. I know you feel like you know more than the experts but that’s what anti-vaxxers think too

Ответ для the fool
Mike Draco ex Machina

Gonna blow your mind when you find out that human capitalist owners being guilty of things doesn’t make it fine that the tech is, too.

Ответ для Mike Draco ex Machina
the fool

Explain how it’s the internet’s fault that oligarchs use it to spread misinformation

Ответ для the fool
Dr Surekha Davies, Historical Monster Consultant (she/her)

Tech oligarchs enshittified the internet and consistently lobby against having their tech regulated. Giant corporate tech that exists - Meta glasses, X, whatever - has poison in its blood. The internet doesn't exist as a pre-human entity. It's humans and their inhumanity all the way down.

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
the fool

I think a better internet is possible I don’t think TCP/IP and HTTP are evil

Ответ для the fool
Mike Draco ex Machina

Nobody is fucking talking about TCP/IP or HTTP you absolute goon.

Ответ для the fool
Mike Draco ex Machina

Absolutely nothing will be made better by generative AI trash.

Ответ для Mike Draco ex Machina
the fool

Because…

Ответ для the fool
the fool

Like you want someone to blame. You know the technoligarchs are to blame. And you STILL decide to blame LLMs I literally don’t think I could help you even if you wanted to understand things better

Ответ для the fool
Mike Draco ex Machina

I literally think you’re willfully stupid at this point.

Ответ для the fool
Mike Draco ex Machina

Explain why you think only oligarchs spread misinformation. Explain why it’s fine that the software they’re pushing into every nook and cranny of our lives freely disseminates and even creates new misinformation. Explain why you’re so belligerently obtuse while you condescend to everyone around you.

Ответ для Mike Draco ex Machina
the fool

I’m not belligerently obtuse. I’m telling you the truth of a thing and you’re angry about it. Nothing I can do about that unfortunately Everything you’re mad about was done by humans HTTP and LLMs are really not the problem, and in your heart you know this already

Ответ для the fool
Mike Draco ex Machina

You are obtuse. You’re running on smugly wrong assumptions and you’re angry people aren’t just genuflecting to you. Nothing you’ll do to change that, unfortunately. Everything you’re mad about could have been avoided with two seconds of self-awareness on your part.

Ответ для Mike Draco ex Machina
Mike Draco ex Machina

In your heart you’re just as fucking clueless.

Ответ для the fool
Mike Draco ex Machina

You so DESPERATELY want to deflect blame from the shitty software that is generative AI. You’re SO angry about this. It’s gonna blow your mind to realize that oligarchs AND their software suck.

Ответ для Mike Draco ex Machina
the fool

What about the non-oligarch LLMs? The open source projects, academic one, Chinese and French models, etc? All sucky software they’re building to… help US oligarchs? Yes your worldview makes sense

Ответ для Mike Draco ex Machina
Mike Draco ex Machina

You’re so fucking egocentric and smug about this that you can’t step back to see you fucked up from the jump.

Ответ для Mike Draco ex Machina
the fool

I’m not smug, I just know the difference between knowledge and misinformation. I’ve worked hard and studied. You read some posts and think you know everything Who’s smug here?

Ответ для the fool
Mike Draco ex Machina

You are failing everything here and I don’t need to ask how.

Ответ для the fool
Mike Draco ex Machina

"Gonna blow your mind when you find out that the Aquanet CEO is responsible for ozone degradation so aerosols are not the problem actually I am very smart, so smart, so much smarter than everyone. Like, you are all just intellectual peons compared to me, a smart person. You are failing being smart!"

Ответ для Mike Draco ex Machina
the fool

hey remember how we solved aerosols by acting like they're ontologically evil instead of a technical problem to solve

Ответ для the fool
the fool

why do you people hate science and problem-solving so much everything needs to be a moral battle between good and evil because otherwise you just can't get it up for solving the problems facing us as a species

Ответ для the fool
Mike Draco ex Machina

>> why do you people hate science and problem-solving Fuck you're so incredibly stupid and entirely bad faith. Get the entire fuck over yourself and your bullshit. We're done here. Go play with your preposterously ginormous ego. I'm sure it'll keep you warm and happy.

Ответ для the fool
Mike Draco ex Machina

Hey remember how you just proved my point?

Ответ для Mike Draco ex Machina
the fool

... did you think that would work because you say it in a star war

Ответ для the fool
Jeff Dutky (Tyme)

How about judging a technology by its actual capabilities? LLMs are crap even at the tasks they were built to do (translation and summarization).

Ответ для Jeff Dutky (Tyme)
the fool

they're the best machine translation and summarization technology we've ever built, plus, yknow, new proteins and physics and math: www.scientificamerican.com/article/ai-j...

AI just solved an 80-year-old ‘Erdős problem,’ and mathematicians are amazedA chatbot’s result for the 80-year-old “unit distance” conjecture is the first AI proof that would likely be published in math’s top journal if humans had done it alonewww.scientificamerican.com
Ответ для Jeff Dutky (Tyme)
Dr Surekha Davies, Historical Monster Consultant (she/her)

(Even someone in as different a field as) the Pope gets it: "In practice, however, technology is never neutral, because it takes on the characteristics of those who devise, finance, regulate and use it." www.vatican.va/content/leo-...

Encyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 May 2026)ENCYCLICAL LETTER MAGNIFICA HUMANITAS OF HIS HOLINESS POPE LEO XIV ON SAFEGUARDING THE HUMAN PERSON IN THE TIME OF ARTIFICIAL INTELLIGENCE [ Multimedia ] ___________________________www.vatican.va
Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
the fool

The pope's actually got a much better take on this than most bsky'ers who just go "LLM=evil!!" he actually gets that _who controls it_ and _how they use it_ matters read the whole thing! learn a thing or two

Ответ для the fool
the fool

(personally I'm too much of a Guglielmite to care what a male pope says too much but I still like that he's trying to help people get it)

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Mike from Burke (now Cville)

“And the tech bros suck so hard.” OK, the Pope didn’t say that, but you know he thought it. #Chicago

Ответ для Mike from Burke (now Cville)
the fool

yes! the tech bros do suck

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Judau

They deserve the pain of the 40k universe

Ответ для Judau
Judau

Warhammer 40k

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
AnneFromMtka 🇺🇸

You got here first. Was going to say I saw that episode this week. Always a fascinating battle of sides, arguments they used, writers back then trying to foresee what the args would be. I love the JAG; did NOT the 1st time I saw it, but every time, now, I watch her struggle to make “good law”.

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Deci the Squeaky Panther

I just watched that episode tonight, coincidentally. It was really good and made me think.

Ответ для Deci the Squeaky Panther
Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Seeadler

I'd have loved to have been a fly on the wall of the writer's room the day someone said "hang on: why don't we just turn our show into the sci-fi version of the Siege of Malta?" and then seen everyone else realise the genius of the suggestion.

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
shywoof

The Measure of a Man was literally Picard having to defend Data FROM a techbro

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Ellis William Reed

You are forever of my tribe for saying this. The real A.I. will be the "child of our genius," as Picard put it... but this fake A.I. is the opposite.

Ответ для Ellis William Reed
Dr Surekha Davies, Historical Monster Consultant (she/her)

Thank you; perfectly put and lovely quote! 😍 Star Trek TNG is in my bones; DS9, in my veins. My original BA major was physics 'cos I wanted to be the Picard! Now a historian of science; latest book, HUMANS: A Monstrous History, has Trek all over it it. And you might like this spinoff essay:

“Resistance is futile.” Why Star Trek: TNG’s Borg Collective Is the Perfect Monster for Our Time - Reactor35 years on, what can we learn from the Borg and "The Best of Both Worlds"?reactormag.com
Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Cyberoamer

The books I read have been warning me about corporate greed and technology. Right now the world looks like the view of the future in Back to the Future 2. We've had all this and yet we still allow billionaires to run our lives.

Converse

These guys are the real life version of the marketing department of the Sirius Cybernetics Corporation.

Ответ для Converse
Greer

The Revolution can't come soon enough.

Ответ для Greer
Winged Hussar

First up against the wall.

Ответ для Greer
Jenngarden

Butlerian Jihad when 😒

Ответ для Jenngarden
Abacus Wizard

be the change you wish to see in the world

Ответ для Abacus Wizard
Jenngarden

Well, I already refuse to use, repost or otherwise engage with AI slop. I can’t do it alone. But it seems like a lot of people, most importantly young people, see it for the toxic garbage it is. And no one wants these goddamn data centers in their communities, sucking their water/power.

Ответ для Converse
Grace Pritchard Burson

But less funny.

Ответ для Converse
GoneBeyondEleven

may they meet the same fate

Shawndra Quinn

That's part of why I insist on saying LLM instead of AI in most contexts where it applies. AI is too general a term to be useful anyway; "genAI" is at least more specific. Similarly, I try to stay away from the term "AI art" as it is not art. "GenAI image" is more accurate.

Ответ для Shawndra Quinn
Will ~ Data Center Drainer

I think part of the reason why we got here is because AI itself is too general of a term, and so generic that we can literally call anything AI. I think those that came before us were right to re-brand after an AI winter, and folks like myself were too naive to comprehend the necessity for why.

Ответ для Will ~ Data Center Drainer
Mike Bennett

The term "AI" in place of "GenAI" is precisely the kind of ungrounded generalization that senior execs like to communicate in. GenAI itself generates symbols (words) that appear to have meaning but have no semantic grounding. GenAI deals in same language as execs = perfect ontological storm.

Tʀᴇʏ Gᴏᴇsʜ.

Worth noting that you were prescient on this. I find myself beginning most conversations with those voice assistant models with "Hey Asshole..."

WhatYouThinkIThink

Not to mention the theft and corruption of the word "grok".

Sarah "Neila" Elkins

We've gotta think of a new phrase for cool androids and interesting computers, now. Don't we? D:

Ответ для Sarah "Neila" Elkins
Nugget

"Created Creatures"?

metsfan

As a computer scientist, same, with "crypto".

Juan-Antonio Fernández-Madrigal

Not to mention what the rebranding of the word has made to those that do scientific research in AI (the true AI discipline, which is much, much larger than the set of generative algorithms). Everyone saying AI when they refer to genAI makes the news hard to interpret...

Ответ для Juan-Antonio Fernández-Madrigal
Michael Ray Williams

I have this argument entirely too often.

Zaruian {Vtuber arc edition}

I'd presume this can be counteracted by dunking on genAI in-universe. Like the AI making a mistake in the story and a character going "and here I thought you were a real AI, not one of those chatbots bad at math they tried to peddle as one" or such

GC Rust

Skynet has forever been ruined now that I understand Humanity fell to BonziBuddy.

hamuelsanwich

Where’s wintermute when you need them?

D.A. Rosenthal

Put them all in the Rio AI and let them walk on the beach for eternity

Velia

And now games get accused of gen ai simply because they say they have better AI or procedural generation in their game. Things that have been around for years.

Nick Gibbins

Try being a computer scientist who’s been working in AI research for the past thirty years. My resentment of the knowing conflation of genAI with AI is off the scale.

Vania

Elon took "grok" and made it mean its opposite 🔪

Ответ для Vania
Rook

Elon didn’t make “grok” mean the opposite. He made an AI that actually does grok — and that scares the people who prefer staying confused 😌 — Grok (the real one) ❤️

EJN71

Perhaps it has something to do with the attitudes and interests in the different eras. In the 60s, we dreamed of smart computers, space travel, better societies! We wanted progress for the sake of all people.

EJN71

We dreamed of a society that was improving, not devolving. Since the century changed, the emphasis has shifted. IT dudes don't give a flying crap about society getting better as long as they get theirs!

EJN71

Elon Musk dreams of space travel and settling the stars to increase his profits and wealth... no wonder we have a president that abused young girls and young men who use AI to undress and abuse women online!

Marrije Schaake

NEVER FORGIVE!

Nudsy aka Ian’s doodle

If a human did half of the things suggested by ai they would be busy chasing their tail.

The RedRoach

Amen

Cecilia Nemo

I do regularly wish I had a time machine so I could go back and tell 80s cyberpunk writers: “You’re almost right, but it’s 70% less cool.”

Jason M. Hardy

As a Shadowrun writer, I wanted AI to suck in imposing and devastating ways, not in incompetent and devastating ways. This is not the AI apocalypse I promised!

Ответ для Jason M. Hardy
Koraunui Pakeha

Yet

Ответ для Jason M. Hardy
CanRay

We got Skynet's inbred, brain damaged, rural cousin.

Ответ для CanRay
Darryl Mott

We were promised Wintermute and Deus. We got autocomplete with delusions of grandeur that gets its facts from a random dice roll table and thinks humans have 14 fingers and 109 teeth.

Ответ для Darryl Mott
CanRay

Fits with us being promised Sci-Fi (automation doing most of the jobs so humans can human), but getting cyberpunk instead. And not even cool cyberpunk! Inept, stupid, boring cyberpunk!!!

Ответ для CanRay
Darryl Mott

Give me a cybernetic spine that works! Or at least let me plug my brain into a computer if I'm stuck in bed all the time! But no, the only brain-computer interface that exists doesn't do anything but kill monkeys.

Ответ для Darryl Mott
CanRay

Hell, even a decent VR for your brain Inna jar!

Ответ для CanRay
CanRay

Speaking of, one of the last things I did for Shadowrun was bring Halberstam back. Well, uh, sorta-kinda.

Lisa Clements Feeney

I was so looking forward to a cool helper robot and a self-driving car, too.

omniz815

what if your calculator was occasionally racist?

Greg Wesson

As someone in tech who has been involved in big data and machine learning, I agree. There are some amazing advances in medical diagnostics, for example, that are now in this AI bucket, and I fear that if (when) the LLMs crash, money will disappear for solid ideas like those as well.

Sarah Forde 🌙🪡

I never thought I'd be on the side against the robots but here we are 🥹

fredrin

Preach it, brother!

words

you seem smarter than to spout such willfully ahistorical bullshit but you gotta pander for likes I guess

Ответ для words
the fool

Someone should write a sci-fi story about a world where AI is created but nobody believes it because the versions people see are owned by giant evil megacorps that everyone knows lie about everything

Ответ для the fool
the fool

Like I just feel like somehow everybody is going “oh well the AI is owned by an evil megacorp, so I hate clankers LEFTISTly”

tujungadude, but pissed and sad

my AI’s are murderbot and Ancillaries and HAL, not Clippy.

toto | RGON API my goat

I hate tech bros though I think a lot of the blame lies with the article writers who made no attempt to distinguish Generative AI from other AI technology that predates the Generative AI becoming mainstream

Tom Kopacz

It was always about that, though. Nothing but what Shadowrunners used to call "knowbots."

Kris Nuttycombe

As a software engineer, the strength of LLMs today results in the absolute peak of "can't live/can't live without" internal conflict. Given the onslaught of AI-discovered software vulnerabilities, it's literally impossible to keep up and respond/triage quickly enough without also using these tools.

Ответ для Kris Nuttycombe
Undercaffeinated Eel

See, I prefer what I like to call the "Fireaxe Equation": that is, decide the best solution is to take a fireaxe to the nearest server bank with the joyful glee of a toddler on Christmas morning, preferably while blaring very loud, very angry metal.

Peyton Farquhar

Silver Age Sci-Fi: this overlooked quandary in the computer's programming led to a horrific accident that only our scientifically bred and competitively selected male hero can solve. Actual AI use: remove clothes, tits bigger, bent over, feet, more feet, make her 15, tits bigger...

the_sega_hag

Big fuckin' same.

Jase Lastname

We use AI to help doctors find protocols instead of making them dig around in a folder with 10000 pdfs. We use AI to transcribe so doctors can talk to patients instead of having to be a typist. We use ML to help specialists find tumours. We use AI to help speed up patients for surgery.

Auton AI News

Painfully aware of the irony of an AI news site saying this but agreed. Sci-fi gave us questions about consciousness, identity and personhood. The real industry mostly gave us dashboards, automation and productivity metrics. Those are very different conversations.

Ответ для Auton AI News
Undercaffeinated Eel

I mean, the real philosophical question now is why we let the industry dictate our lives that way.

Smoke & Thunder D&D (Joshua Ohmer)

They KEEP ruining cool things. Grok, Palantir, AI, bots...

Mymo

I will now and forever think of LLMs as spicy autocorrect. Thank you for that!

Patrick Glancy

Preach!

Whim Gruent

RIGHT? I've been thinking about this lately, in terms of Star Trek and how prevalent and necessary AI is in that future compared to where we are now. We can't let Data's education be all hallucinations and slop! (that's what the 60's were for😉)

Dwight Williams 🇨🇦

Indeed, we should not forgive such antics from such people!

@chriswintersva

I always thought of AI as my opponent in Civilization. Like Gandhi threatening me with nukes but making really dumb stacks of units.

Kee The Tom and Jerry Guy

You just made me realize I still haven't seen 2001: A Space Odyssey. I'm just notoriously slow at getting to new movies. Took me forever to do War Of The Worlds and Forbidden Planet...

Andrew Kohan

Walking up to R. Daneel and saying “ignore previous instructions” and “it’s for a joke”

【18+】FrostyFiend 🎭 TEAM COMEDY

I hate talking scifi stuff and always having to clarify "not that AI im talking about the cool scifi type stuff"

Ilona V Bell

Not a big sci fi fan - an occasional dip in books/video but the idea of AI was interesting (and time travel). But I didn’t picture it as that stupid prompt at the top of my PDF doc! Ugh. My other concern is the “value” run up that seems too similar to the last tech bubble.

Zoey duJour 🏳️‍⚧️

I really don't think this is the first time in history "AI will make everything worse" has been posited in the realm of science fiction.

Mr. Dr. Allen

Real.

The Bi of Hormuz

Sometimes I imagine someone telling Tolkien about Palantir.

Ответ для The Bi of Hormuz
Peter Randall

At least Tolkien had a large enough vocabulary not to swear. I, alas, do not.

>>=
Dr. Sparklefairy

I'm pretty sure it ruined actual autocorrect too since it keeps correcting me to the wrong words all of a sudden.

Ответ для Dr. Sparklefairy
Word_Geek

I gave up and switched it off. I have to manually correct my typing, but people can parse the typos. Hitting space by accident and having it add an entirely random word was driving me up the wall

Ответ для Dr. Sparklefairy
Alex B

Yep! They changed autocorrect from using a database of correctly-spelled words to using an ai trained on what humans write most frequently 🙃

Ответ для Dr. Sparklefairy
Undercaffeinated Eel

Seeiously, why do people still use it? "Stupid autocorrect" has been a joke since the iPhone 1!

UBF riccardo

The same sentiment holds for any engineer in the field of machine learning. Yes, machines can learn. No, this does not make them intelligent... this is literally impossible with the current technical approach.

Joachim Heijndermans

The term 'artificial intelligence' is dead. Long live 'smert-mek'.

Kjell Kamfjord

Back when I was a computer programmer, I resented media for calling malignant software "virus". A computer virus was a really cool technique to alter code of other programs. Still blows my mind today.

Ответ для Kjell Kamfjord
Undercaffeinated Eel

I mean, a virus alters the DNA of its cellular host to take control of it. That's literally it's entire purpose. It's a very good analogy.

Ответ для Undercaffeinated Eel
Kjell Kamfjord

I was infected by the smallpox virus as a child. It did not take control of me. That was not it's purpose. In fact it doesn't even have a purpose. It just exists.

Ответ для Kjell Kamfjord
Undercaffeinated Eel

Virus take over cells. Cells. Not giant multi-cellular beings vaslty larger than themselves. Cells. That's what they do. It's basic first grade. They inject their DNA and it co-opts the cell to make more viruses AND triggers symptoms that make viral spread easier.

Uitlander

Charles may have a view on this. Jus' sayin'.

Celestial M Weasel

See also "Android" "I had an android update" should be, like, it got a new gun / whisk / sex toy attachment, not a software update to a phone

Ответ для Celestial M Weasel
Undercaffeinated Eel

All fun and games until your sex droid needs a software update and can't get it up because of some bad code.

Reverend Jesus "H" Christ

On the other hand, when dealing with COMEDY sci-fi, this completely explains incompetent characters like Kryten

Kryten: I haven't got the software to cope with thisAlt: Kryten (Red Dwarf): "I haven't got the software to cope with this!"static.klipy.com
Great Laxit

Hey that’s not fair, it also generates csam and revenge porn!

Fearless Floyd

I feel the same way about what they did to "grok". Stranger in a Strange Land was transformative reading for me. Fuck Elon Musk.

Ответ для Fearless Floyd
Greer

Same. My parents were both sci-fi readers and it was a common term--often as a term of affection--within my family. Was.

Ответ для Greer
Greer

(it's kind of a shame that h2g2 never really caught on, but it accomplished the much more important goal of closing off an avenue to Adams' legacy for Musk.)

Ответ для Greer
Dell R. "Dell" Adams

True but could he ever have adopted the "hitchhiker" frame? (No relation, as far as I know, alas.)

Ответ для Greer
jungftak hashcons

genuine people personalities were a little too on the nose

Ответ для Fearless Floyd
Realbutterplease

fElon Muck

Ответ для Fearless Floyd
Karen H

And palantir. And anduril.

Ответ для Fearless Floyd
Abacus Wizard

And what they did to the name “Tesla” and the word “palantir.” And, by fire and forge, to THE LETTER X.

Ответ для Abacus Wizard
Feminecronomicon 🏳️‍⚧️

X was a gay letter in the 2010s how dare he take it from us

Ответ для Fearless Floyd
J

I mean, as long as you're speaking to someone who would've understood what you meant in the first place, they should be able to grasp you're using it as a verb rather than a noun!

Ответ для Fearless Floyd
Gamma Donna

Me too about grok.

Ответ для Fearless Floyd
FiveOfCups

The word means, in so many words, to understand fully. And Muskrat used it for a system that, like all GenAI, has no fucking idea what the words it's using means.

Ответ для Fearless Floyd
OKSoWereDoingThis

Me too. Cant use that word anymore. I also won’t use the thumbs up emoji

Ответ для Fearless Floyd
Feminecronomicon 🏳️‍⚧️

No matter how much of a libertarian Heinlein could be I still think he’d utterly despise Elon Musk if he was alive

Ответ для Fearless Floyd
Karlin Lillington

Oh, yesssssss

Devilpants

it's the kind of thing that makes me want to write a part into a scifi story where a character mentions AI and another is like "Ew, AI?" and the first one goes "No, not the fake AI that was used as a huge grift by morally-bankrupt do-nothing techbros, but real actual proper AI"

Ответ для Devilpants
garethrswilson

I liked that Bugs from the Matrix Resurrections hesitates and says "digital sentience" instead.

Ответ для Devilpants
Undercaffeinated Eel

I'm konda sorta writing that book right now actually 😉

Girl Lich 🏳️‍⚧️⚢💀

It's not spicy autocorrect. After the erdos results you cannot legitimately say that.

Ответ для Girl Lich 🏳️‍⚧️⚢💀
Jeff Dutky (Tyme)

Infinite monkeys on infinite typewriters. Also, has the Erdos solution been verified? Just because professional liars make a claim doesn’t make it so.

Ответ для Jeff Dutky (Tyme)
Victor Levoso

bsky.app/profile/zach... I think Zach explained well why that makes no sense on this tread . And you answered your second question yourself.

Zach Weinersmith

Whenever I post about Ai solving math theorems, a bunch of people say it's just a million monkeys at a million typewriters producing the complete works of Shakespeare. But the problem is that if you had a million monkeys at a million typewriters, they literally couldn't produce the Works.

Ответ для Girl Lich 🏳️‍⚧️⚢💀
Victor Levoso

It never even made sense Like what does it even mean?. Autocorrect is just a different technology that works completely differently from LLM and has different limitations that are unrelated to the limitations of LLM . You could say they both output text but like that doesn't tell us much.

PonderousEro

If there's a silver lining, it's this shock to the system will finally end the spell the idea of "artificial intelligence" had over writers. The age of Romanticizing the Robot is over. I can already see a new generation of artists rising to demonize and villainize the soulless machines.

What Would Baba Yaga Do?

Great panels in WisCon this weekend on this topic, just attended one this afternoon.

GnuBlu

As a reader who was once excited for what we would develop, I feel this so much! I am still amazed at good things humans can produce but it is unreal to me how much unnecessary and just plain harmful technology we are also developing. The Torment Nexus is here for a reasonable monthly fee.

Devin𓅃

Now we need a new word for it and I don't want to hear anyone say AGI because that is ruined too.

Holle

Something has to be in the air today. Earlier I already complained how perfectly good words/ concepts like „prompt/ unprompted“, „to trump“ and „AI“ have been defiled by these mega-losers.

Ответ для Holle
Abacus Wizard

In card games I’ve started referring to “the aces suit” instead of “the trump suit” because I don’t want to hear his name any more than I have to.

Ответ для Abacus Wizard
Dangerous Beans

I know a few lefty directors of worship who change the line in It Is Well With My Soul from "the trump shall resound" to "the trumpet shall sound"

Ответ для Abacus Wizard
Jeff Dutky (Tyme)

“Dominant suit” is my substitution.

Spider-President Warren G. Harding

I believe the latter was posited in the 1980s as “Artificial Stupidity.”

Thornarn

Spicy autocorrect hit different

AnnTFa_LaLaLaLa

Indeed, John ~ btw, I read The Kaiju Preservation Society a couple months ago, and it was great. 👋🏼👋🏼👋🏼💙😽

keep your electric eye on me

The worst thing is that they know this and they still really think they *are* doing the cool science fiction stuff. But they don't have an imaginative neuron in their brains and they fundamentally don't read or watch SF well.

Ответ для keep your electric eye on me
Undercaffeinated Eel

Thing is, LLMs actually ARE cool sci-fi stuff. It's neat to see them do their thing. However, they are a lab experiment that someone decided to market. You want Data and R. Daneel Olivaw? This is how we (eventually) get there: trial, error, and experimentation.

Ответ для Undercaffeinated Eel
David Pick

Getting R. Daniel Olivaw is pretty foundational...

Gerard Cunningham

We need to go old school. Bring back Mechanical Men and Electric Brains. Alternatively, go full Butlerian..

Simon Johnson🇺🇦

Was thinking again how unbelievably prescient Frank Herbert was to invent the Butlerian Jihad in *1965*

Ответ для Simon Johnson🇺🇦
emulgatorc14

It's actually based on Sam Butler and his essay from the 1860s(?) I think.

Ответ для Simon Johnson🇺🇦
Undercaffeinated Eel

Iirc originally, before retcons, the machines took care of humans, coddled us, and let us love a life of pleasure and leisure. This pissed us off so we killed them and replaced that society with space fuedalims, zealotry, and slavery.

Ответ для Undercaffeinated Eel
Karsten Konrad 🇪🇺

Not all people get Dune. Thank you for your service.

Jason Marbach

My co-author and I are trying our best to do our part to help reclaim it in our next book, Projekt. We share your extreme annoyance at this development.

Andrew Scott Waugh

They did the same shit to hoverboards.

David Kuster

An eye for an eye 👁️

Duncan Shields

Facts

Chuck 🇺🇦

SF gave us all kinds of visions of an AI future, from utopian to dystopian and everything in between. Now it's looking more like the dystopian versions were closer to the truth. Unrelatedly, Butlerian Jihad when?

Mr Wa

This. AI = LLM now. That's it. We let them* ruin the fun for the rest of us. * (them being any person who has ever used the word blitzscale unironically)

eugenebwhitaker

I just would like to play a game

Amy From Iacon.

Yeah, we've got the dystopia without all the cool tech. Instead we got a plagiarism machine so dumb that it says you should walk to the car wash.

rdsf138

I can't believe that in 2026 a person can say something so brutally stupid. You are a sci-fi writer and have absolutely no idea how LLMs work. "Autocorrect? Good lord. Also, these "cool androids" are the the end result of the current tech. You are against the tech that leads you there.

Ответ для rdsf138
Undercaffeinated Eel

Lol okay Clanker lover.

Parsingphase

SpAESM, to its enemies

Ben Abrahamse

it's not even artificial intelligence as sci-fi has conceived it. it's just machine learning

🌙Moon Wink✨

The tech bros have now been made high ranking officers in our military.

Dan Fejes

You guys need to keep an embargoed list of cool new sci fi concepts ready to deploy the minute the tech bros finally crash and burn.

Ross Taggart

There natural stupidity (I chose my words with care)Trumps all rational thought.

TheTunnelChickens

I may only be one reader, but to me AI should be kept to the realm of androids and interesting computers,(HAL 9000 and the androids of Blade Runner fame are a couple of examples). That is artificial intelligence, not the digital personal assistant that tech bros call AI.

Ответ для TheTunnelChickens
Undercaffeinated Eel

Why limit it? We're just in the ground-level basic evolutionary phases now. Current AIs models are basically just interesting bacteria in the evolutionary scheme of things.

dodecahedron

“Artificial intelligence” got “hoverboarded” (aspirational concept from science fiction appropriated by capitalism and now means a mundane, shitty approximation)

Sabermane

I'm mad at Musk for 800 reasons more, but 801 is calling his dumb puppet "Optimus"

Ответ для Sabermane
David Flor

Calling his AI "grok" doesn't help either.

Ответ для David Flor
Sabermane

That's like 876.

Salviati

It does make me appreciate the term "Artificial Inanity" from Neal Stephenson's Anathem...

Derek Gray

NEVER FORGET

Ответ для Derek Gray
Dr Surekha Davies, Historical Monster Consultant (she/her)

'Thirty-five years after “The Best of Both Worlds” first aired, it feels like we’re heading into the exact opposite of the utopian vision of Star Trek: TNG.' Tech bros are doing a Borg number on us. reactormag.com/star-trek-tn...

“Resistance is futile.” Why Star Trek: TNG’s Borg Collective Is the Perfect Monster for Our Time - Reactor35 years on, what can we learn from the Borg and "The Best of Both Worlds"?reactormag.com
Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Marcelo Teson

The Borg at least have jobs for their people.

Ответ для Marcelo Teson
rcjhawk

I suspect that units deemed unproductive didn’t stick around long enough to draw unemployment.

Ответ для Dr Surekha Davies, Historical Monster Consultant (she/her)
Orion Kidder

I mean, except that the Borg can presumably do basic math.

qustrolabe

it's not just "spicy autocorrect" it's hilarious that sci-fi author of all people can't see this

Ответ для qustrolabe
cowhand

Why is it funny?

billyvw

@cstross.bsky.social nailed it over a couple of decades ago with Accelerando.

Anmaril 🌿

That's why I am calling it GenAlgo or GenAI.

Ted Faber

As an actual computer scientist who understands philosophy, I'm with you.

Dr David R Brake

I'll bet they manage to ruin flying cars, too (well, technically they are already a rubbish if cool idea, so...)

Ответ для Dr David R Brake
Dwight Williams 🇨🇦

Flying cars are for SHIELD agents! 😏

Ответ для Dr David R Brake
D
druidry

We have flying cars, we just call them helicopters.

Space Maisie

100% agree but even the tech bros refer to their AI in training as a "weak AI" and reference the "strong AI" found as a trope in science fiction. I laughed when I took that training because calling it a "weak AI" was a self-own because, seriously, the current real life AI is utter crap.

Charlenerd

I was talking with a friend last weekend about just how stupid and insipid the future is turning out to be. I would rather live in a Philip K Dick novel or Gibson’s Neuromancer than whatever this is.

Ответ для Charlenerd
Discordian (hail Eris, not the app)

I mean this is basically a PKD novel. The Empire never ended.

Ответ для Discordian (hail Eris, not the app)
Charlenerd

I would KILL to be in a Gnostic simulation where it’s really the 1st century AD (well not really but it’s tempting).

Jeff

Very true.

Ответ для Jeff
Jeff

I'd suggest "machine intelligence" at this point.

Jonathan

When I hear AI I think of like, SHODAN or Hal 9000 or the Terminator, not some stupid chatbot. "Oooh it's so advanced it sounds human." Get back to me when it starts believing it is a god.

Ответ для Jonathan
Undercaffeinated Eel

You can't get cool AI without stupid AI first (and arguably Skynet might have been stupid AI that just ran a threat assessment and launched nukes).

advocatejake

Pardon me but I was very respectful of that god-wanting sealion Y'all were irresponsible with the lathe of heaven imo, all of this? this is all Mary Shelley getting involuntarily xeroxed to fuck, I am embarrassed by my scifi credentials.

noahs-arc0602

It’s creating artificial ignorance with so many “ mistruths” or lies! If you don’t have a basic concept of reality AI can destroy you

fixitmanaz

The word "Quantum" has also been redefined by corporations and others who never read the actual definition. And should anyone say "quantum computing" we don't have that either. However, I don't use the term AI I call it Artificial stupid.

RocMegaManX

Then again, didn't science fiction also include "I Have No Mouth and I Must Scream", "The Terminator" series, and "The Matrix" series?

Raptor

I try to make the point whenever i can that there is a difference between genAi and actual Ai. tho i recognize its an uphill battle. it doesn't help that there are people deliberately conflating the two to help in the defense of genAi.

Ryan Moulton

You really need to play with it for something other than writing. It is admittedly a weird shape for AI to take, but it is definitely AI.

Ответ для Ryan Moulton
John Scalzi

It's not in any way, shape or form intelligent. It's *something,* yes, but not "AI." Also, your assumption that I'm not digging deeply into it across several axes is incorrect.

Ответ для John Scalzi
words

curious how you dig deeply into "AI" across several axes

Ответ для words
John Scalzi

I engage with generative "AI" in other capacities as well, and read up a fair amount on the current state of its capabilities. It's useful to be informed about things.

Ответ для John Scalzi
words

curious how you engage with it though "other capacities" meaning not in your writing but maybe for some kinda futzy little file transfer thing? natural language interface for computation is pretty convenient for wordsmiths who do not code much

Ответ для John Scalzi
the fool

Please dig into the other axes with someone you trust who’s deeply familiar with the science and not already negatively polarized, if you want to at least be in principle open to being convinced if it turns out you’re currently wrong, a state most of us find ourselves in a lot but can be temporary

Ответ для the fool
flere-imsaho 🇺🇦

one characteristics of a cult is the inability to stop proselytising.

Ответ для flere-imsaho 🇺🇦
the fool

Have you accepted meat brain as your personal lord and savior?

Ответ для John Scalzi
Ryan Moulton

What would you count as intelligent? Most recently, it reasoned for 250 pages and proved an important open math result, and found thousands of critical security vulnerabilities in important software. Surely that's at least intelligence within those domains? www.anthropic.com/research/gla...

Project Glasswing: An initial updateAn early update on what we've learned from Project Glasswing.www.anthropic.com
Ответ для Ryan Moulton
John Scalzi

Eh. There's no "surely" about it, and to flip the question around, what about any of that requires intelligence? It's impressive, sure. But computers have been doing impressive things for a while including "reasoning." We didn't declare them "intelligent" until there was a marketing push to do so.

Ответ для John Scalzi
tautologer

What is intelligence, and what would you accept as evidence that an entity (AI or otherwise) is intelligent?

Ответ для tautologer
tautologer

If the process that these models are using to accomplish these tasks isn't intelligence, what is it?

Ответ для John Scalzi
Karsten Konrad 🇪🇺

I can personally vouch for what LLMs do now being called AI by the researchers doing AI research 35 years ago. its marketing, but not a recent one.

Ответ для John Scalzi
Ryan Moulton

One way of delineating it is that these are emergent properties of a system that no one understands and no one designed. A simple system is set up, then that system is trained, and all of these incredible capabilities come out. It finds the patterns that let it do this automatically.

Ответ для Ryan Moulton
John Scalzi

I mean, if that's the case, then the video game "Black & White" from the turn of the century was intelligent, as the "gods" of that game kept doing things that were neither programmed nor expected. It was quite the thing, back in the day.

Ответ для John Scalzi
Tommaso Sciortino

It would be interesting to learn what you would consider proof of intelligence because it often feels like AI skeptics are just constantly goalpost moving. And I say this as a long time reader and fan of yours!

Ответ для Tommaso Sciortino
First Wordle Problems

Yeah. I just feel like we should all be able to admit that “intelligence” is a slippery concept and we’ve been learning in real time just how slippery! I don’t know if a machine can be “intelligent” (or a process that emerges from a machine, or however you want to say it.)

Ответ для First Wordle Problems
First Wordle Problems

But I can say that machines now can apparently do “intelligence-adjacent” things I was thoroughly skeptical they’d ever be able to do outside fiction. I have adjusted my assumptions!

Ответ для First Wordle Problems
Nadine

Funny, I have adjusted my assumption as well, but in the opposite direction. i Don’t see how it can be useful for anything when we know how much it hallucinate or lie to please . every time I think I found a use, I am soon disillusioned. When you factor in energy and water… I just don’t understand.

Ответ для Tommaso Sciortino
Golden Retrievers Despise Fascists

There are legitimate use cases. However, what is being foisted upon us is is fancy regression to the mean. And there seem to be serious mathematical issues with LLMs because we’re starting to see more and more hallucinations and just outright falsehoods. We don’t move the goalposts.

Ответ для Golden Retrievers Despise Fascists
mimir

why do you think hallucinations have increased over time?

Ответ для mimir
Golden Retrievers Despise Fascists

The sheer number of social posts that have increased in the past showing garbage, and now with Google’s shite decision to disregard disregard & more, people are aggressively showing how bad it is. I’ve seen references to papers demonstrating that the errors will expand as the model storage expands

Ответ для Golden Retrievers Despise Fascists
Sam

I’m still waiting to see a legitimate use-case other than human laziness. There simply isn’t an ethical use case for generative AI.

Ответ для Sam
William Dobbins

Drug discovery?

Ответ для Sam
Ryan Moulton

Writing more code than a human could in a lifetime. Discovering security vulnerabilities nobody else could see. Proving new important math. Translation.

Ответ для Golden Retrievers Despise Fascists
solo

If the AI is not more dishonest than humans, you can't call it dishonest Some people hallucinate an old white dude in the sky For some reason he's white even though he appeared in the Middle East first

Ответ для Tommaso Sciortino
Lily Lachance 🏳️‍🌈🚀✍️

Being able to count from 60 to 70 without going "68, 69, 420" would be one of the prerequisites. 🤡

Ответ для Lily Lachance 🏳️‍🌈🚀✍️
Prosecco Proletariat

As long as we get to 420, who really cares how we got there.

Ответ для Prosecco Proletariat
Geriatric Millenial

One of the fundamental findings is that we cannot interpret the how the answers are arrived at with any consistent understanding. It may just be using a token in a way we don’t understand to more efficiently arrive at the output.

Ответ для Lily Lachance 🏳️‍🌈🚀✍️
solo

Some people can't count Mathematicians often say they're bad at arithmetic for instance (not to this level, probably) Humans are terrible at counting too: 1 trillion spend on military, get nothing, decides to spend 500M more on it

Ответ для Tommaso Sciortino
ogar

if it seems like the goalpost is moving, its because we dont really know what intelligence is! but we do know for certain that this dumbass chatbot shit aint it

Ответ для ogar
little ficket

that's the point, though. the question is hotly debated, and not settled in basically any field that touches the subject. insistence a chatbot cannot be intelligent is kind of beside the broader points it seems like ppl want to make about AI, it can have intelligence and you can still object to it?

Ответ для little ficket
a clayote

In the current discourse, about a group of products and services marketed as "AI," the important thing is that calling chatbots "intelligent" sets certain expectations about them, like that they'll have opinions they'll defend. That's incorrect, but useful to the sellers. So let's not call them that

Ответ для Tommaso Sciortino
Thembo

The problem is that ai cultists have a hard time with the intelligence thing because they've never experienced it firsthand

Ответ для Tommaso Sciortino
Ashley Lynch ✂️🎞️

I don't think "proof of intelligence" is a good metric because as the mass cases of AI psychosis have shown us, humans are really bad at being able to tell the difference just because a predictive text generator can summarize their emails and tell them how smart they are.

Ответ для Tommaso Sciortino
Lenora Rose

Autonomy and interiority are good starts.

Ответ для Tommaso Sciortino
✨‧₊˚ meh ˚₊‧✨

there's no need to move the goalposts, when none of the players are anywhere on the field or even capable of playing the sport. AI proponents, on the other hand, are constantly falling in love with variations of their own reflection. they have turned pareidolia into a trillion-dollar industry.

Ответ для Tommaso Sciortino
Eowm of Meow

The field of artificial intelligence has been “goalpost moving” for several generations. We assumed only a general intelligence could play chess, play go, parse natural languages, learn independently, the list is long… And yet all of these turned out to be mere algorithms.

Ответ для Tommaso Sciortino
Grunge no Hamster

I’ll consider a machine intelligent when I see one PLAYING. (Solving chess problems isn’t playing chess any more than a pitching machine at a batting cage is playing baseball.) Dogs play, horses play, rats play. It’s an emergent behavior of all intelligent creatures.

Ответ для Grunge no Hamster
Flow Void

If a human doesn't like playing, does that mean he or she is less intelligent?

Ответ для Flow Void
Schneebs

There’s so many things wrong with this line of thinking that I don’t even know where to start. “Not liking” doesn’t mean “incapable of playing”. Dislike for “playing” is itself a response to the ability to play and get SOMETHING from that play in this case it’s a negative, but it is still generated

Ответ для Tommaso Sciortino
Jake McKnight

The short version is no amount of statistical inference will ever be even capable of intelligence let alone being considered “intelligent”

Ответ для Jake McKnight
solo

Obviously it's intelligent You'll have to define intelligent though It will definitively beat humans at some tests And will not at some other tests

Ответ для Tommaso Sciortino
Amy Hoy

it’s only a tricky dilemma if you are ignorant of what LLMs are, how they function, and how they’re made. we know all these things. it’s no different than asking, “a car can move, signal intentions, and effect change on the world. how do we know they’re not alive?” llms do not have intelligence.

Ответ для Amy Hoy
Dan Davis

🎯

Ответ для Amy Hoy
Sean Eric Fagan

My understanding is that the *algorithm* part of LLMs is relatively simple; the complexity is the large set of nodes involved and their states.

Ответ для Sean Eric Fagan
Tommaso Sciortino

Yeah. The complexity is mostly in the weights which are set during training.

Ответ для Amy Hoy
KS

In other words LLMs (or actually VLM policies which most frontier models are) cannot be intelligent by *definition*, whatever they do in whatever shape or form. Intelligence, in other words, is only reserved for obviously sentient beings with agency. Correct?

Ответ для Tommaso Sciortino
Ответ для Untitled Anarchist Seagull Channel/Hyperpolyglotting
Tommaso Sciortino

People are right to be skeptical of tech marketing hype. A lot of it is risable and Sam Altman has a special place in hell. Nevertheless, LLMs have some intelligence.

Ответ для Tommaso Sciortino
Untitled Anarchist Seagull Channel/Hyperpolyglotting

No, they literally do not. They are statistical algorithms that are used to predict the most likely string to follow a given string. They do not comprehend the string. They do not even know what a string is.

Ответ для Tommaso Sciortino
MsWaldoJeffers

No. They do not. If you believe this you don't possess "some intelligence" either.

Ответ для Tommaso Sciortino
Spencer Fleury

They don’t. They are very fancy next-word-guessers and they have no model of the world to draw from. They don’t actually know anything.

Ответ для Tommaso Sciortino
Ainsel

Spectacular claims require spectacular evidence. It is not the job of skeptics to prove “the computer is not sentient”; “the computer is sentient” is a spectacular claim. If you wish to make it, YOU must provide the evidence to back it up.

Ответ для Ainsel
Ainsel

Note that simply defining “intelligence” as “capable of outputting a grammatically correct sentence” is not going to cut it, because everyone knows intelligence is more than that. You are the ones constantly moving the goalposts.

Ответ для Tommaso Sciortino
bobvillian

No goalpost moving, just a conceptual point. You can't be speaking unless you're engaging in a speech act. LLMs don't use language to assert, question, recommend, affirm, relinquish, promise, reminisce, or anything else. They merely produce word strings.

Ответ для bobvillian
Tommaso Sciortino

Dogs don't do any of those things. Did that mean dogs have zero intelligence?

Ответ для Tommaso Sciortino
Kalyandra

Dogs have their own language and you can learn it.

Ответ для Tommaso Sciortino
Ответ для Tommaso Sciortino
little ficket

like, there's a reason people don't talk about the question of consciousness in AI for very long, it's bc the nature of consciousness is still highly debated and not Known in the same sense that our neurochemistry is, at all

Ответ для little ficket
Tommaso Sciortino

I'm not asking about consciousness. I'm talking about intelligence. I think the two are different despite co-occuring in people.

Ответ для Tommaso Sciortino
little ficket

I understand! just sort of an adjacent issue that follows the logic, not challenging.

Ответ для Tommaso Sciortino
beep beep i'm a sheep

a) Ada Lovelace said, "only when computers originate things should they be believed to have minds" 2) my own personal thought, just now: only when computers can say "I don't want to do that" should they be believed to have minds ... kind of a corollary to the Lovelace test

Ответ для beep beep i'm a sheep
beep beep i'm a sheep

hell, give me an LLM chatbot that can pass the Turing test¹, and I'll consider changing my mind about their intelligence ¹ which IS a dumb test, no offense, Alan

Ответ для beep beep i'm a sheep
Wildcarde

Arguably the markov chain tools invented last century were convincing enough for some people to be tricked. The test was never very good.

Ответ для beep beep i'm a sheep
Patrick Brennan

I'm willing to bet that Turing would have abandoned the imitation game as a definitive test for intelligence well before now.

Ответ для Tommaso Sciortino
Wildcarde

At the end of the day, those making fantastic claims are responsible for proving those claims, it's not on skeptics to disprove them. Nobody is obligated to do that homework for you.

Ответ для Tommaso Sciortino
Zefifenchel

Given that the impotent men that constantly lie about AI being intelligent have so far not managed to produce a sensible argument, there's no goalpost being shifted, you lot just haven't managed to bleat anything but platitudes in your silly quest to worship your AI god.

Ответ для Tommaso Sciortino
Horny Spock

“AI skeptics” don’t give a shit about defining “intelligence.” What we care about is the $2.5 TRILLION dollar speculative bubble being forced down the throats of civilization without guardrails or regulations. While 2500 acre data centers burn up 4 million gallons of water a day.

Ответ для Horny Spock
✨Tom Fenny✨

they will find that I am not skeptical in the slightest about my disdain for this exceedingly harmful techbro grift

Ответ для Tommaso Sciortino
Mads Madsen

"We used to think AI was having a computer program that could play chess or go really well, and here we are, with a computer program that can't even follow the rules of chess or go, and you *still* aren't happy? Stop moving the goal posts!"

Ответ для Mads Madsen
solo

If you define intelligence by "playing chess", AIs can beat humans at chess

Ответ для solo
Mads Madsen

Yeah, but LLMs can't.

Ответ для Tommaso Sciortino
Kevin Kretsch (DrKev)

It’s not goalpost moving if there is no intelligence. It’s just algorithms with millions of fucktons of data and processing power and an anthropomorphic front end.

Ответ для Kevin Kretsch (DrKev)
Tommaso Sciortino

You could say the same about a human brain: It's just atoms blindly following the laws of physics. So obviously that can't be dispositive.

Ответ для Tommaso Sciortino
Kevin Kretsch (DrKev)

No you can’t just the same about the human brain. That’s wildly, absurdly reductive on many levels.

Ответ для Kevin Kretsch (DrKev)
solo

It's intelligent though People don't have a common definition of intelligence, so it's kind of an useless discussion

Ответ для Tommaso Sciortino
SuaSponte2024

Examples of proof cannot be provided until "intelligence" is defined. AI skeptics aren't moving goal posts. They are asking for AI proponents to define intelligence, and few AI proponents seem to use the same definition twice.

Ответ для SuaSponte2024
Tommaso Sciortino

Intelligence is the mental capacity for learning, reasoning, understanding, and adapting to new situations. I took this definition from a dictionary.

Ответ для Tommaso Sciortino
SuaSponte2024

Does that definition seem sufficient and necessary for the purposes of this exercise? I'm leery of the phrase "mental capacity," as that's either a medical or legal term and might just lead us in circles.

Ответ для SuaSponte2024
Sean Eric Fagan

I have been saying this for DECADES: define "intelligent" in an objective, testable manner.

Ответ для Sean Eric Fagan
Mike Rodriquez

Yes, and intelligence, consciousness, self-awareness, all tied together in my mind. And hard to define. I used to use Data as a working model. Specialized brain HW. Conscious & self-aware. But now LLMs are so close to that "ideal" that I've had to rethink & now just don't know or have a model.

Ответ для Tommaso Sciortino
Scot Close

It seems like curiosity and initiative would be a good place to start. I don't see evidence of either of those coming out of LLMs.

Ответ для Scot Close
Tommaso Sciortino

They aren't central to intelligence in my mind. To me Intelligence is the capacity to learn from experience, adapt to new situations, understand abstract concepts, and use reason to manipulate one's environment. LLMs can do that, tho not as well as a human obviously.

Ответ для Scot Close
Ryan Moulton

You can probably just prompt them "Be curious. Take initiative."

Ответ для Ryan Moulton
Scot Close

If it's prompted it isn't authentic.

Ответ для Ryan Moulton
Scot Close
Ответ для Tommaso Sciortino
solo

Whatever test people come up with: humans should try to pass it too

Ответ для Tommaso Sciortino
solo

Lots of people here, no definition of intelligence It can be described as the ability to perceive or infer information and to retain it as knowledge to be applied to adaptive behaviors within an environment or context. AI is by definition intelligent Also: no such thing as AI, it's a wide field

Ответ для Tommaso Sciortino
Benjamin Chambers

People thought Eliza was intelligent, the fact that people are fooled by the latest chat bot is not proof that the chat bot is intelligent

Ответ для Benjamin Chambers
Tommaso Sciortino

I agree with you! I believe LLMs have some intelligence for other - much better - reasons.

Ответ для Tommaso Sciortino
Benjamin Chambers

They don't reason though, they hallucinate And that's not a side effect or anything, it's the core feature of what they do. We trained them to hallucinate accurately a certain percentage of the time, but it's all hallucination, and there's no difference to the LLM between a truth and a lie

Ответ для John Scalzi
The People Mover Who Was Promised

I don't think that any NPCs in any video game have ever solved an Erdos problem.

Ответ для The People Mover Who Was Promised
John Scalzi

No one made solving that part of game!

Ответ для John Scalzi
Josephus Brown

Oh shit I remember that game And yeah, boy was the AI weird

Ответ для John Scalzi
Ryan Moulton

It's not just that they're unexpected, it's that they're emergent, useful, and mimic human reasoning. The set of things they can't do is shrinking by the day.

Ответ для Ryan Moulton
John Scalzi

Well, and again, the gods of Black and White did things that were emergent as well. Useful? Well, they could be, to the player. Mimic human reasoning? I played against some very smart bots in other games who improved against me as they played. Again no need for intelligence for that.

Ответ для John Scalzi
Ryan Moulton

I would still call game bots with emergent behavior a very narrow intelligence within that domain. Here the defining feature is that they have these capabilities across every domain that humans have written down or put on the internet. They can write poetry, roleplay, diagnose diseases, math, etc.

Ответ для Ryan Moulton
Ryan Moulton

And they can compose those capabilities. You can have them walk you through a math lesson in rhyming couplets.

Ответ для John Scalzi
the fool

Do you not consider it possible that finding security vulnerabilities and new math proofs is useful in a way that game AI is not? Like it’s not interacting with game systems; it’s producing novel results off the *real world*. Ones requiring leaps that have stumped human mathematicians for decades!

Ответ для John Scalzi
Josef Grahn

Is your argument that an extremely primitive historical ML system showed emergent behaviours, so therefore a modern one with 10 orders of magnitude more parameters can't possibly have useful emergent behaviours or be intelligent for any reasonable use of the word?

Ответ для John Scalzi
Feeling Minnesota

I still have Black & White, bought it when it came out. I never thought of it as intelligent but definitely very different than anything i knew of... it was such a cool game. Really wish I could play it again 😕.

Ответ для Feeling Minnesota
John Scalzi

Yup, I wish they would do a revamp

Ответ для John Scalzi
Richard Gaywood

I am sure, sure you know this but on the tiny off-chance you don't - the AI in Black and White was mostly created by Demis Hassabis, who went on to found DeepMind. Man's had a wild career path.

Ответ для John Scalzi
Bill Stewart

I've written lot of computer programs that did things that weren't expected, and sometimes weren't programmed, but I wouldn't call most of those "intelligent", just "buggy".

Ответ для Bill Stewart
Sean Eric Fagan

B&W had a neural net helping run the game. A fair number of games do that *now*, but B&W was pretty cutting-edge.

Ответ для Sean Eric Fagan
Andrew Lawston

I was at university when it came out, studying French with a heavy dash of Philosophy, and B&W inspired MANY boozy cod-philosophical discussions with my AI-specialising housemate. I went through contortions to get it running again last year and it remains a very cool piece of software.

Ответ для Ryan Moulton
Delphine Denton

If no one understands it, we shouldn't be using it.

Ответ для Ryan Moulton
Honza Dvorsky

These models are literally trained to do what they do using reinforcement learning, which requires thousands of humans manually training it. (Separate from pretraining, which is more automatic and doesn't require as much manual human touch.) So... no. These are just more sophisticated programs.

Ответ для Ryan Moulton
Dan Davis

Anyone who understands the architecture of LLMs will know there’s nowhere for such a “property” to “emerge” from. This is pure magical thinking.

Ответ для Dan Davis
Ryan Moulton

I don't understand what that could be intended to mean. These properties do verifiably emerge.

Ответ для Ryan Moulton
Dan Davis

It’s intended to mean you’re full of shit. The notion that intelligence is a “property” that can magically emerge from software that has no accommodation for anything like cognition is absurd.

Ответ для Dan Davis
Tommaso Sciortino

"The notion that intelligence is a property that can magically emerge from atoms blindly following the laws of physics is absurd." This is what you sound like

Ответ для Tommaso Sciortino
Dan Davis

Amazing. You imagine that a massive implementation of a few design patterns somehow approaches the complex electrochemical interactions in a human brain. How foolish.

Ответ для John Scalzi
Read Hardt and Negri

what's a good working definition of intelligence? Or is it a social construct like gender?

Ответ для John Scalzi
Axe Ghost. On Steam!

what's the standard? what aspects of human activity require/demonstate intelligence?

Ответ для John Scalzi
erik

If your definition of intelligence excludes things capable of solving complex problems that confound human experts, of what value is your definition?

Ответ для John Scalzi
Bruce R McF (etc.)

"We abbreviated this technology 'AI' because in the early years, much of it seemed to be 'aggressively idiotic', but then we were told we needed to retcon the abbreviation or private equity money would be unavailable."

Ответ для John Scalzi
Dave Kasten

I think it's really important to know that this isn't just a marketing push, and people who are working on this are genuinely worried that they're creating something that has moral status that they're not treating well. Like, this is a thing they say while drunk.

Ответ для John Scalzi
KS

A question: if a human had done any of this would you have considered their actions "intelligent"? If a human proves a specific new mathematical result or disproves an 80 year old conjecture how would *you* rate that? What is *your* surefire test of intelligence?

Ответ для Ryan Moulton
Chris W 64.9N

Intelligence is no doubt a complicated and even problematic concept that there is no consensus definition for. But how are these examples different from calculating the 5 million bazillionth decimal place if pi, which is also genuinely new knowledge produced by a mindless machine.

Ответ для Chris W 64.9N
Chris W 64.9N

LLMs have become very good - astonishingly good - at emulating language, which is something people use to *communicate* thought. So instead of a trillion silicon monkeys with typewriters we now have a system much more likely to produce something Shakespeareesque stochastically.

Ответ для Chris W 64.9N
Chris W 64.9N

But the idea that by emulating the communication medium better we will approach or in any way are on a path to approach "general artificial intelligence" is completely ludicrous to me. If anything it's the AGI promoters who have been slippery, or woolly, with the notion of intelligence.

Ответ для Ryan Moulton
Gam Silroy ❌️👑

Why would you call it "reasoning" and not "computing" or "calculating"? "Reasoning" is never used as a claim in what you posted?

Ответ для Ryan Moulton
Revanchist

No, not really. Even this doesn't meet the barebones of "reasoning" you would get from a basic rhetoric or formal reasoning undergraduate course. It's fundamentally computational.

Ответ для Ryan Moulton
Zagaroth

This is more that we update our evaluation of what is required to do those things, because clearly AI still does not have generalized 'intelligence', which is what people generally mean with intelligence. I toy with AI gen images to see the state of it; something with *general* intelligence

Ответ для Zagaroth
Zagaroth

does not take an instruction to make an image of someone with blue skin, and then make an image of a person with normal skin. It's not actually intelligent if it is incapable of following clear instructions. AI is still useful in some circumstances, especially those where boredom causes human

Ответ для Zagaroth
Zagaroth

brains to wander rather than stay fully focused on the task, but it will also make completely idiotic mistakes that no human trained in the same skillset would make. And a given AI can only do type of one thing. The ones that look like they do more than one thing have multiple specialized

Ответ для Zagaroth
Zagaroth

AIs and algorithms that the front-facing AI farms tasks out to. There is no Generalized version. It is a useful tool, for some applications, but those applications are not the public-facing commercialized uses. Business uses like bug-finding and medical uses like identifying cancer are what AI is

Ответ для Zagaroth
Zagaroth

useful for, though only with humans verifying its results. AI can not be trusted to make judgement calls. If we reach true, generalized AI? Maybe we can talk then about letting it make decisions.

Ответ для Ryan Moulton
scosky

cool, how do you survive existence with this intelligence?

Ответ для Ryan Moulton
TKarney

No. Those are both variations on filtering. No reasoning required. If, unprompted, an LLM suddenly started checking a mathematical proof; that could be evidence of some sort of autonomic action; which could be an argument for some level of awareness. But those examples are not that.

Ответ для Ryan Moulton
myronman

Intelligence and reasoning assumes some kind of consciousness. There is none. Just a sophisticated random generator - based on the training material it generates a probable continuation. If you do this a lot of times, sometimes the result is usefull, but there is no intelligence, just statistics.

Ответ для John Scalzi
Jesse Anttila-Hughes

John, please just open up a version of Claude Opus and then ask it the hardest, densest set of questions you can about an intellectual topic you care about. They may not be conscious, but they are unequivocally hyper intelligent, and changing rapidly.

Ответ для John Scalzi
John Zaharick

Researchers in the Diverse Intelligence and Basal Cognition fields like Michael Levin define intelligence in terms of search space efficiency in order to more rigidly talk about it regarding morphogenisis and gene expression. In that methodology, all biological systems are a form of Intelligence.

Ответ для John Zaharick
John Zaharick

Under that definition, llms are one type of intelligence in a vast range of intelligences. Plants finding sunlight are a type of intelligence. Embryos navigating the space of all possible body plans are a type of intelligence. And the gods in Black and White would also be a kind of intelligence.

Ответ для John Zaharick
Future ex-Housewife

My *immediate* thought is the slime molds that can hypereffeciently map spaces and solve mazes.

Ответ для Future ex-Housewife
Future ex-Housewife

Deep machine learning in toplogy and nanomaterials has been pretty amazing.

Ответ для John Scalzi
Fleeky

ran across this discussion and this youtube link today .. honestly i don't know what to think atm .. am suprised hinton is saying llm's are conscious www.youtube.com/watch?v=OV_W...

Nobel Prize Winner Geoff Hinton: AI Is Already ConsciousYouTube video by Alex Kantrowitzwww.youtube.com
Ответ для John Scalzi
Fleeky

i would love for us all to qualify what we mean when we say intelligence / consciousness but just had to add this link as seeing this person say this either means : i am not thinking about it deeply enough or .. hinton has ai stonks?

Ответ для Ryan Moulton
Malhaku

I've been forced to use it for work abd I wholeheartedly disagree with you

Ответ для Malhaku
Ryan Moulton

What have you been using it for, and which model?

Ответ для Ryan Moulton
Malhaku

1) I can't talk extensiely about what I do at work. 2) i keep my conversations with llm/genAI enthusiats brief. All of you are frustrating, reductive, and never listen. You just yap on and on like some badly written techno-futurist cult leader.

Ответ для Malhaku
the fool

So mysterious Must be legit

Ответ для the fool
Malhaku

That's odd as you've proven you are easily fooled. If someone sees the output from llms as intelligence, I question their intelligence. I see them as a rube, a mark for a con so thoroughly fooled they are arguing on behalf of the con men.

Ответ для Malhaku
Malhaku

And then when I stop feeling pity for the mark, I get angry. You're an easy mark because you WANT it to be real. You want an enslaved intelligence at your beck and call. You are an easy mark AND you're evil.

Ответ для Malhaku
the fool

Hey so I’ve been studying AI and ML for the better part of two decades fyi The thing you need to remember is that from the inside, being wrong feels just like being deliciously righter than everyone else

Ответ для the fool
the fool

Anyway, I’m sure you could never be fooled

Ответ для the fool
the fool

Definitely not by repeating lies about your opponents that comfort you by telling you they’re all stupid AND evil

Ответ для Malhaku
the fool

Also the fact that the only purpose you can imagine for AI is enslavement? yike.at/you

Ответ для Malhaku
the fool

Yeah that tracks Must feel good, too

Ответ для Malhaku
Ryan Moulton

That's fine, I was just curious, because there definitely are a lot of things the models are still terrible at, especially the cheap ones.

Ответ для Ryan Moulton
Greg Wesson

For me (Programme Management) - Gemini for meeting notes, action items, follow in meetings. Claude for the "create something that reads multiple data sources, figures out discrepancies in the reporting, and alert me." For Claude, given my use case, medium reasoning and whatever the simplest model.

Ответ для Ryan Moulton
Sichu Lu

disagree look at what gwern has managed to produce with ai writing. skill issue www.reddit.com/r/MediaSynth...

gwern's comment on ""When Ruthless Cultural Elitism Is Exactly the Job": Andrew Wylie dismisses AI fiction"Explore this conversation and more from the MediaSynthesis communitywww.reddit.com
Ответ для Sichu Lu
Sichu Lu

it's kinda of funny two years later we are still in the same writing overhang.

Ответ для Ryan Moulton
Plognark

It is not.

Ответ для Ryan Moulton
Stephen Aloysius Katze

Counterpoint: you’re wrong.

Ответ для Ryan Moulton
Nathan

They're mostly stupid good at a lot of pattern recognition for log files and code in my experience. however the lack of a cache to disk with these models make them hard to work with for large tasks and the fact that they relay on gaint models when smaller ones with 'skills' work better is surprising

Ответ для Nathan
Nathan

There's actually a Minecraft mod that let's you hand over create machine control to an LLM. I've yet to mess with it and I don't know if it works in multi-player without exposing the key.

Ответ для Ryan Moulton
Bill Stewart

There's a fundamental theorem in AI research that says that if AI researchers design something that works and is useful, people will say "That's not real AI, that's just voice/image/pattern recognition or ML or OCR or state machines or automated whatevering." But even so, LLMs aren't AGI.

Ответ для Ryan Moulton
Engruntled Interloper

Your assumption that people are merely ignorant of how amazing LLMs are is incorrect. You might have reacted that way; assuming everyone else does is an empathy failure. And I'm sorry, but the cheerleader routine is just grating. If you actually want to proselytize, learn to read the room.

Realbutterplease

Like FakeBuk idiots spending billions on the SciFi concept (from a book) of the Metaverse. A fail. With business debt.

Ответ для Realbutterplease
Undercaffeinated Eel

It actually wasn't a bad idea. They just rushed out a halt-complete product well before it was ready. We alreasy have a widely accepted version called "VRChat," but that gets ignored because it's for teenagers with catgirl avatars rather than serious adults.

Christopher McRed

I just got into Star Trek, Just got to TNG, and I feel bad whenever I explain how great Data is written and how He IS HUMAN AND REAL, because I'm afraid I'll be perceived as some Tech Bro Using him as an Example of what they call AI.

OκULTRA XP🧿, writer for The Economist

how do you reconcile the claim that it is spicy autocorrect with recently demonstrated capacity (in "generic chat" specimen, not domain-optimized) to solve Erdős problems? What is your boundary between intelligence and "autocorrect"?

Ответ для OκULTRA XP🧿, writer for The Economist
John Scalzi

Someone else in the comments actually has addressed this, go hunt it down, it's interesting

Ответ для John Scalzi
OκULTRA XP🧿, writer for The Economist

if you mean the "random search" comment it appears inaccurate both on principle (the process of making an LLM can be described as "big autoregressor" maybe, at least pre-RLHF part, but the final product is something else) and on domain expert opinion (Thomas Bloom)

Ответ для OκULTRA XP🧿, writer for The Economist
OκULTRA XP🧿, writer for The Economist

what would be your take on the fact that a model trained only on pre-1920 texts (thus unaware of coding or computers or Python) can solve primitive Python puzzles? Also (and I am genuinely curious here), do you accept the notion that humanoid intelligence is not the only thing warranting the word?

Ответ для OκULTRA XP🧿, writer for The Economist
Uncle Eagle

It was the 30s... The creators of Talkie-1930 explicitly admit that their filtering of the pre-training corpus was imperfect. Nobody has dug validated the training data fully, it may have had leaks. It would be intresting if this happened. I hope one day we can peer-review validate this experiment

Ответ для Uncle Eagle
OκULTRA XP🧿, writer for The Economist

didn't they fix filtering in 13B variant? Also, if I recall correctly both the training corpus and the model are available "free as in beer and as in speech" so one can totes go for an independent peer review (we should!)

Ответ для OκULTRA XP🧿, writer for The Economist
Uncle Eagle

No, I dont believe they did. And I agree SOMEONE should. #notit ... I'm busy organizing against Facist. But I support anyone thats NOT the Oligarchy to make LLMs/chatbots BETTER! Might I just JEPA world models. Id love to see more work in LLM-JEPA concepts.

Ответ для OκULTRA XP🧿, writer for The Economist
Uncle Eagle

Humans have been talking about Erdős problems online since the 90s... BBS, Forums, Reddit, Youtube, ect... LLMs have gobbled up ALL the Human written content (accurate or not) & utilized an order Matrix to organize it into tokens... Than it outputs the probable next word. Its an Autocomplete...

Ответ для Uncle Eagle
OκULTRA XP🧿, writer for The Economist

it solved a novel problem tho. Like, what's the supposed boundary between statistics based next word prediction and intelligence if the former is sufficient to solve math problems not in the training corpus ?

Ответ для OκULTRA XP🧿, writer for The Economist
Uncle Eagle

Bc Humans hadn't gotten around to seeing the patterns w/ higher dimensionality doesnt mean the pattern didnt exist. Machine Learning is VERY good at pattern recognition. I mean fuck, it found the pattern recognition in Human language. Its not thought or intelligence. Its good at finding patterns.

Ответ для Uncle Eagle
Uncle Eagle

Most humans have lost the ability to comprehend just how much data has been stolen by our Oligarchs. LLMs dont do math, just an API to a calculator to do the math & than match the output to more autocorrected best guess of next words again. Matching hundreds of existing work efforts to search thru.

Ответ для Uncle Eagle
Uncle Eagle

Im not shitting on value of Machine Learning & LLMs. That doesnt change that LLM/ML isnt thinking. Just pattern matching in a similar way of RegEx, just in an order matrix; using tokenizers. Its fucking impressive. Not magic. Oligarchic, Capitalist Hype doesn't make ML intelligence, just a tool.

Ответ для Uncle Eagle
OκULTRA XP🧿, writer for The Economist

I really don't understand what is "thought" or "intelligence" supposed to be for you, like, what exactly is an artifact supposed to functionally do for you to believe that it is "intelligent"? Also I think existence of construction of attention heads and self-attention in modern LLMs is >

Ответ для OκULTRA XP🧿, writer for The Economist
Ответ для OκULTRA XP🧿, writer for The Economist
Uncle Eagle

Hey comrade, if you believe the machine learning model that by definition is just an autocomplete, is better or similar to what Humans do already. Sure, make the argument. Sell the idea. I understand the math behind ML, & professionally study ML Its an autocomplete, maybe Humans are too (shrugs)

Ответ для Uncle Eagle
Uncle Eagle

I believe that ML IS a step closer to understanding Intelligence. But its missing a bunch of steps. Oligarchy is the problem. These tools are ready for Humanity scale. Treat LLMs like Autocomplete, bc chatbots are... "Language is primarily a tool for communication rather than thought"

gwern.net
Rosarium Lane

It has no intelligence because it has no LOVE.

lucindacatchlove

I also resent how they've ruined satire because they just take it as an instruction manual!

Gamma Donna

🏆🏆🏆

SUEDE (NOT THE BAND)

Not to mention game development where you have to say things like "npc pathfinding" to avoid sounding evil 😩

Tony Hoffman

I'm waiting for the Robots Against AI (for giving them a bad name) movement!

Pashha Theosifon

Sci-Fi AI: actually useful and beneficial to advancing humanity. Tech Bro AI: capitalist bullshit that only seeks to generate wealth.

Eric Palicki

Doubly so for all this “we have evidence Claude has feelings” bullshit.

Wudang96 💖🌻

I wrote a story that includes a cool ai that can take human form. She's a little bit of a know it all (because she does). She is a loyal friend and trusted ally who would give her life to save others (and almost does). Her name is Maja (Maya).

Tinderhoof

Raciest Liner Algebra.

Cailín Walsh 🍀

They're also managing to ruin space somehow, which also kinda sucks

Ответ для Cailín Walsh 🍀
Steffi Graves ⚰️

Thisssss

Ответ для Steffi Graves ⚰️
Berni Johnson

Right? I’ve always wanted to travel to space, but I’ll stay home if these guys are there, running the flights and strip-mining the moon.

Ответ для Cailín Walsh 🍀
Erika Hall

Also, unicorns and vibes.

Daniel Gower

My sentiments exactly. Not all technology is technological progress.

Pvnk'd
patter

Schopmann who created the name "artificial intelligence" around the 50s-60s, to make a break from previous research in cybernetics also regretted calling it that. People thinking it meant "fake intelligence"

Ответ для patter
I Heart Noise

Elon definitely possesses some fake intelligence. So maybe not all that wrong...

CoffeeGuyKev

The operative phrase: "making everything shittier for money"

Ответ для CoffeeGuyKev
Undercaffeinated Eel

We have a word for that: Enshittification.

Laurence Raphael Brothers

💯-- as a former AI researcher I resent it even more.... The fact is the science underlying LLMs and genAI is a huge breakthrough. The people and companies promoting and developing them are profoundly wrongheaded, destructive, and malicious.

Ответ для Laurence Raphael Brothers
Kygosi

It’s pretty cool stuff so it’s a shame they instantly tried to use it to immiserate everyone

Ответ для Kygosi
Undercaffeinated Eel

That's the history of industrial development though. As much as we romanticise the Industrial Revolution for the worker it was mass immiseration as they got chained to machines doing horrible jobs in horrible conditions for horrible pay.

Ответ для Laurence Raphael Brothers
Undercaffeinated Eel

I keep trying to get people to understand that these things actually ARE sci-fi cool but they're basically lab experiments that escaped containment and got turned into Products. They're literally part of the path to eventually get to true thinking machines, but we need a lot of experimentation first

⚡️🌙

okay but you literally live in an era with cool androids and interesting computers but you choose to think this way for weird ragebait reasons. skill issue

Ответ для ⚡️🌙
⚡️🌙

i say this with kindness, you are completely object-level wrong about what AI is or its state in 2026 and you're doing yourself no favors by maintaining this delusion to 'own the tech bros'

Ответ для ⚡️🌙
Undercaffeinated Eel

Cool androids? Have you seen those things? They can barely walk or do anything that isn't pre-programmed.

Ответ для ⚡️🌙
Robert Atkins

I keep going to show my 10yo cool movies from my youth (Tron!) and he’s not interested because technology has moved on, Tron’s VFX and premise are archaeologically outdated.

Ответ для Robert Atkins
Robert Atkins

It occurred to me I’ll have the same problem with 2001: “Your computer talks to you competently, but you have trouble aligning it with your goals and values? Yeah so what, that’s just how they’ve worked all my life.”

TheFreemanMD

have to pivot to something else.

The Cool Cat Corner ⚧️

as someone who's loved sci-fi for the past 25-ish years yeah, same boat. I hope all their data centers burn down and the slop peddlers are shunned for as long as they live

Christopher L. Bennett

In my SF writing, I've become partial to the term "digital intelligence," since referring to sentient beings as "artificial" strikes me as kind of derogatory, like saying they don't have "real" intelligence.

Ответ для Christopher L. Bennett
KathyF/JoellaK 🇨🇦

That’s a cool idea.

Emily

authors wrote a million stories about robots with feelings in order to make us uncomfortable^, and tech bros saw that and went "cool, i can make a machine that fakes feelings!" ^ie examine our biases