3887b4May 5, 2025
Eugen Rochko
Still baffled by the sheer irresponsibility of unleashing the Convincing Lie Generator on the general public. I don’t think there’s a historic precedent save for only maybe the release of teflon into the environment.
May 5, 2025 at 6:51amMastodon for iOS
Show threadThe FrogMay 5, 2025

@Gargron

I'm just reading about people being addicted and thinking the thing is sentient and some sort of metaphysical being.

Altman's reaction is "oops, our model is a bit to sycophantic, sorry" and promised to turn down the dials. Yes, possibly in terms of percentage, the number of people affected this way is small. But we're still talking about thousands of lives wrecked!

To me, any person capable of wrecking so many lives is a monster.

Show threadginoputrinoMay 5, 2025
@Lily_and_frog @Gargron Read that article also. For those that haven't seen it: https://www.rollingstone.com/culture/culture-features/ai-spiritual-delusions-destroying-human-relationships-1235330175/ Just nuts.
People Are Losing Loved Ones to AI-Fueled Spiritual Fantasies

Marriages and families are falling apart as people are sucked into fantasy worlds of spiritual prophecy by AI tools like OpenAI's ChatGPT

Rolling Stone
Show threadThe FrogMay 5, 2025

@ginoputrino @Gargron

Yep, ive read that one just before coming across the post. That's exactly what I was referring to. Scary stuff. Worse for me, I'd be a perfect candidate to fall for it.

Show threadgioMay 6, 2025
@Lily_and_frog @Gargron I don't follow your reasoning... AI is a tool, which if used properly and wisely, is beneficial... the same as a car, 1.5 tonnes of steel hurtling at 50kmh through a city centre... do you suggest we ban cars (and blame the manufacturers) if some people are irresponsible and cause deaths using them?
Show threadEugen RochkoMay 6, 2025
@giosci @Lily_and_frog Yes.
Show threadwoe2youMay 6, 2025
@giosci @Lily_and_frog @Gargron Generative AI is a tool in the sense of "when all you have is a hammer everything looks like a nail".
Show threadThe FrogMay 6, 2025

@giosci @Gargron

So much to unpick in your comment.
1. Saying something is dangerous doesn't mean I want to ban it.
2. AI encompass LOTS of different things, which have different uses and pitfalls. The AI used to analyse scans, or to create better quicker subtitles is not the same thing as a chatbot (the lie generator)
3. When it comes to Chat GPT and LLM chatbots, there zero benefit and lots of dangers. One of them is inducing people thinking they're the messiah.

Show threadLeBonkMay 9, 2025

@giosci
@Lily_and_frog @Gargron
Creating tools reinforces the idea that the tool's function is something worthwhile that needs to be done.

When all you have is a hammer, etc. etc.

What is the machine built to do?

Show threadgioMay 12, 2025

@LeBonk @Lily_and_frog @Gargron

"What is the machine built to do?"

I think we both agree, the tool (machine) can be built with good intentions but may end up being used, by some, for bad purposes...

AI governance and its ethics is fundamental. Many of the AI creators are so 'excited' about the new toy that they don't seem to worry too much about the downsides... (or am I naive?)
If AI is going to shape our world in specific ways, we might need to consider what values are being built into these systems from the start, I guess.

Show threadThe FrogMay 12, 2025

@giosci @LeBonk @Gargron

What is the machine built to do?

In the case of Chat GPT, it was built to look impressive enough to bring more investment.

Chat GPT is a demo.

Show threadsph/nxMay 5, 2025

@Gargron Lead in gasoline. CFC in hairspray.

I think there are some precedents 🤔

Show threadRené Seindal (MOVED)May 5, 2025

@sphinxc0re @Gargron

DDT just about everywhere, messing up hormonal balances in entire ecosystems.

We've never needed computers to do really big mistakes, even if they do make it easier.

Show threads427May 5, 2025
@sphinxc0re @Gargron I may be wrong but I think the negative side effects of CFC were not known (and not easy to anticipate) when they were initially released.
Show threadHakan BayındırMay 5, 2025

@s427 @sphinxc0re @Gargron

The problem is not making mistakes, but continue making them after effects become as visible as sun in the middle of the desert.

DuPont continued using PFAS even after discovering the problems (https://time.com/6284266/pfas-forever-chemicals-manufacturers-kept-secret/). Some provinces in China was still using/producing CFC-11 in 2019 (https://www.nature.com/articles/s41586-019-1193-4).

Lastly, Exxon nailed how global warming would happen in 1977 (https://news.harvard.edu/gazette/story/2023/01/harvard-led-analysis-finds-exxonmobil-internal-research-accurately-predicted-climate-change/)!

Companies Knew the Dangers of PFAS 'Forever Chemicals'—and Kept Them Secret

Manufacturers DuPont and 3M kept their knowledge secret, taking a page from the tobacco industry, a new study reveals.

Time
Show threads427May 5, 2025
@bayindirh @sphinxc0re @Gargron Of course. But in the case of LLM, I think it's fair to say that the problems were known even BEFORE they were released for public use, which is a special category of mistake.
Show threadHakan BayındırMay 5, 2025

@s427 @sphinxc0re @Gargron I was talking in the domain of PFAS and oil.

LLMs are a special class of bad, I agree. I don't use them, don't recommend them. The tech is useful as it is, but how it's trained, implemented and pushed upon us is knowledge and user hostile at best.

Show threadHakan BayındırMay 5, 2025
@sphinxc0re @Gargron
Using DDT on humans to kill lice. Writing this still gives me chills.
Show threadOlapMay 5, 2025
@Gargron nuclear testing is up there
Show threadKevin RiggleMay 5, 2025
@Gargron I think a better comparison is the rise of the handbill and tabloid newspaper, which I’m currently reading a book about
Show threadChristopher CaryMay 5, 2025
@kevinriggle What’s the book’s title? I’d genuinely like to read it
Show threadKevin RiggleMay 5, 2025
@cary THE COMING OF THE BOOK, by Febvre and Martin
Show threadWeltzeitgeistMay 5, 2025
@Gargron Dont forget Bitcoin.
Show threadГригорий КлюшниковMay 5, 2025

Oftentimes when some dangerous thing gets invented, its inventors don't yet know it's dangerous. Like all those mildly radioactive products in the early 20th century. Or tobacco before it was found to be very bad for one's health. ChatGPT was initially just a cool tech demo.

But continuing after it was found dangerous, now that's the problem. As previous examples have shown, companies themselves will not stop because they wouldn't want to go out of business out of good conscience. The only way these sorts of thing stop is government regulation.

Show threadNikMay 13, 2025
@grishka @Gargron Вы правы в том, что ЖаДнобесие - это самоисстребление, которое не способно остановиться потому что Жадность насытить НЕВОЗМОЖНО! И оно всегда раздувается, захватывая всё больше и больше, как "мыльный пузырь" пытаясь обьять необьятное и в конце концов неизбежно лопается! С некоторых пор начинаю думать что это даже не просто заболевание,когда"болезнью болеют"! А скорее всего какой то Паразит - Симбиот, или ещё какая зараза из плесени,поражающая мозг человека, делающая его бывшим?
Show threadГригорий КлюшниковMay 13, 2025
Nik, ты ебанутый?
Show threadNikMay 13, 2025
@grishka @Gargron Похоже мы нашли общий язык, по поводу и без повода!! !-:) Теперь, если конечно вас не забуду? То уж постараюсь вас лично больше ничем не тревожить!-:) Всех благ!-:)
Show threadNikApr 7
@grishka @Gargron И вот прошло время до начала апреля 2026 гада!! У вас хоть сейчас появились разумные мысли? Или вы до сих пор считаете что все кроме вас, "ебанутые"? Ну если с вами так ничего хорошего, положительного и умного не произошло за это время? То тогда вам не в Мастодон, а в клиннику для решения "психиатрических проблем"!! Удачи! -:)
Show threadNikMay 13, 2025
@grishka @Gargron Нынешний Мир Человека РАзумного,всегда страдает от Ада Нелюдей Безумных. Пусть вам не покажется это странным и похожим на пропагандупропагандонов "свидетелей ВРАЛЬмического бреда"? Мир и Ад. Две противоположности.Один делает сам.Другой паразитирует.Один создаёт.Другой разрушает! В нынешней системе ценностей наверное только Социально Развитое Государство может быть по настоящему жизнеспособным. Остальные-оставшиеся псевдоимперии, будут сожраны изнутри, прибежавшими по верёвочке!
Show threadQybatMay 5, 2025

@Gargron I'd have gone for leaded petrol as the ideal example. It's pushed by an industry that bribes politicians and runs a PR campaign to mislead the public about both the safety and utility of the product, but really it floods the environment with a toxin that makes everyone stupid.

And we never even needed it - there were non-toxic alternatives available, they just cost more.

Show threadjohnny garbanzoMay 5, 2025
@Gargron can’t tell if you are talking about trump or llms
Show threadHans HafnerMay 5, 2025

@Gargron I have to say that I'm more careful and read more sources than before LLMs where available. I consider this change in my personal behavior a good thing.

For me personally this technology has been extremely useful and empowering.

But I understand that the whole scope of your toot will of course affect me as well by proxy.

Show threadCedric Knight 🌍💙May 5, 2025
@wonshu Useful and empowering in what way? Does that make it a social benefit?
Show threadHans HafnerMay 5, 2025

@cedric I have created many automation scripts to automate my workflow and file tasks, created several (simple!) games that I can now play without tracking and advertisement, created MIDI generation tools, exponentially sped up my infrastructure / hosting knowledge and implementation and probably a bunch of other things I don't remember atm.

For me personally it was great help and now I feel I can much better help other people take digital infrastructure into their own hands.

Show threadIsakaMay 5, 2025
@Gargron I’ve blogged pretty much the same thoughts back in February 2023, I re-read my post and find it still pretty much on point, I https://blog.mimozar.com/2023/02/06/ai-strategy-throwing-pasta-against-the-wall/
AI Strategy, throwing pasta against the wall?

What exactly is going on with the hype around ChatGPT? Are we so impressionable that we suspend judgment on a whim and go for herd moves? What part of pre-trained model is magic? Didn’t we al…

Mimozar Blog
Show threadwibbleMay 5, 2025

@Gargron Maybe more pre-historic than historic, but if the aim is to persuade peasants to work themselves to death and give up their food as taxes in the hope of glittering rewards in a fictional afterlife, then something of the sort might have been tried before.

Although that also relied on the unknowable workings of inscrutable abstractions of which anyone could make requests, the gnomic replies were interpretable only by an exclusive clique, which presumably made all the difference.

Show threadElf BurgermanMay 5, 2025
@Gargron
https://www.humanetech.com/podcast/forever-chemicals-forever-consequences-what-pfas-teaches-us-about-ai
Forever Chemicals, Forever Consequences: What PFAS Teaches Us About AI

AI could unlock vast technological potential—if we do it right. Rob Bilott, who fought chemical giants over toxic PFAS, shares a cautionary tale: how the harms of new tech can go unchecked, and why we must align innovation with safety before it’s too late.

Show threadNicole ParsonsMay 5, 2025

@Gargron

Examine the motives of those funding the AI & cryptocurrency hype.

https://www.bloomberg.com/news/articles/2024-11-06/saudis-plan-100-billion-ai-powerhouse-to-rival-uae-s-tech-hub

They are all people looking for their next set of credulous rubes, their next grift, and especially looking for ways to increase wasteful uses of energy.
https://www.bloomberg.com/news/articles/2025-05-03/opec-agrees-another-oil-supply-surge-for-june-delegates-say

It's about keeping the globe hooked on oil.
https://www.bloomberg.com/news/articles/2025-05-03/saudis-double-down-on-seismic-opec-shift-to-sink-oil-prices

The fossil fuel industry is dumping billions on "plausible sentence generators" not for the product's utility, but for its capacity to...

1/3

Saudi Arabia Plans $100 Billion AI Powerhouse to Rival UAE’s Tech Hub

Saudi Arabia is planning a new artificial intelligence project with backing of as much as $100 billion as it seeks to develop a technological hub to rival the neighboring United Arab Emirates, people familiar with the matter said.

Bloomberg
Show threadNicole ParsonsMay 5, 2025

2/3

... manipulate the stock market, particularly oil prices.

https://www.bloomberg.com/news/articles/2025-02-11/saudi-arabia-s-neom-signs-5-billion-deal-for-ai-data-center

It's also useful for:
1. Election interference
2. Automated pandemic disinformation
3. Ubiquitous state surveillance
4. Molding public sentiment to foster fascist movements
5. Mass layoffs
6. Wage suppression
7. Union busting
8. Imposing class-based market segmentation on public services (only the rich get a real doctor or a real education)
9. Achieving reductions in trust for genuine expertise

Show threadNicole ParsonsMay 5, 2025

3/3

10. Enabling collusion on price fixing & inflating housing costs
11. Theft of IP
12. Dissemination of malware & spyware

So far, the cons far outweigh the pros.

https://www.cnbc.com/2025/04/26/amazon-nvidia-say-all-options-on-the-table-to-power-ai-including-fossil-fuels.html

https://www.desmog.com/2025/04/22/ai-energy-demand-can-keep-fossil-fuels-alive-tech-backers-promise-worlds-two-biggest-oil-producers/

As the planet fries, an accelerated transition away from fossil fuel becomes an imperative.
https://www.technologyreview.com/2024/05/23/1092777/ai-is-an-energy-hog-this-is-what-it-means-for-climate-change/

Solar, wind, hydro, and other sources of renewable energy defunds the most corrupt industry on earth & might save the planet from fascism, war, famine, & ecocide.

Amazon and Nvidia say all options are on the table to power AI including fossil fuels

The tech companies have invested mostly in renewable power, but they are now navigating a changed political environment.

CNBC
Show threadTimPMay 5, 2025

@Gargron CFCs in fridges nearly ripped the upper atmosphere, and Tetraethyllead lead induced violence and stupidity in a generation.

I hope that the whole world getting good at detecting nonsense will be beneficial.

Show threadcontrasocialMay 5, 2025

@Gargron

I looked up photos of an animal as an anatomical study/reference for drawing and there was so much AI trash in the results I couldn't always be sure what were real photos or not. I literally was not able to have any confidence that was I was studying was real or not when it came to specifics of the animals anatomy.

Really depressing cycle of bullshit information getting recycled into nonsense. AI represents a future in which learning about reality will become increasingly difficult.

Show threadAndreiMay 5, 2025
@Gargron someone would have invented it eventually
Show threadRobin AdamsMay 5, 2025
@Gargron It kind of reminds me of that period in the 1930s when they were putting radium in everything.
Show threadJoris MeysMay 5, 2025
@Gargron *lead added to gasoline has entered the chat*
Show threadCoach Pāṇini ®May 5, 2025
@Gargron @jawarajabbi
The God of Lies https://www.superversive.co/blog/god-of-lies-botshit-and-proof-of-humanity
God of Lies, Botshit, and Proof of Humanity — Superversive

Noah Smith wrote : “The combination of anonymity and virality have driven the cost of Big Lies to zero, and we’re just starting to discover what that means for human society. The tech “GPTs and other LLMs have amazing conversational ability and zero understanding of the difference between trut

Superversive
Show threadBill, organizer of stuffMay 5, 2025
@Gargron Ponders lead-lined Roman aqueducts
Show threadBjørnar (he/him)May 5, 2025
@Gargron And currently it's sold by the heroin method. We're in the "free or cheap samples" stage and all the AI companies are wondering if we're addicted enough for them to rack up the price yet.
Show threadJeff AtwoodMay 5, 2025
@Gargron not sure I agree. It's much more subtle than that. This is a tool. In some cases, it works well. In others, it doesn't, or even harms. It's like many other tools, such as the Internet, in this regard.
Show threadAlbert CardonaMay 5, 2025

@codinghorror @Gargron

If the tool requires training, even a certification, then the argument would hold. Here is a tool that requires both, yet it's been provided to everyone without any consideration. As if it was a toy.

Show threadJeff AtwoodMay 5, 2025
@albertcardona @Gargron Has it been WILDLY oversold because salespeople like to make money? Yes. Are the general public total morons who can't discriminate between good and bad results? I think no, unless you have very little faith in all of humanity.
Show threadMartin HamiltonMay 5, 2025

@codinghorror @Gargron  Low-background steel, also known as pre-war steel and pre-atomic steel, is any steel produced prior to the detonation of the first nuclear bombs in the 1940s and 1950s. This is a toot about LLMs  

[*] https://en.wikipedia.org/wiki/Low-background_steel

Low-background steel - Wikipedia

Show threadJared White (ResistanceNet ✊)May 5, 2025
@codinghorror @Gargron Lots of tools just aren't worth it. Many chemicals which have been banned were very useful tools. In some cases, the more environmentally-friendly replacements were less useful. I'm in favor of using *less useful tools* if necessary if the more useful tools (debatable in this case anyway) are also horrendously dangerous and liable to cause catastrophic side effects.
Show threadBhavani Shankar 💭May 6, 2025
@codinghorror @Gargron LLMs are critically different from the internet - they're insanely expensive to train and have never been remotely profitable despite having a billion users.

They're black holes for VC money and threaten our intelligence and the planet.

#llms #ai #idiocracy #climatechange #brainwashing #latestagecapitalism
Show threadLarge Heydon ColliderMay 5, 2025
@Gargron I think the more recent one is unleashing advertising-deferential social interaction on society (i.e. social media).
Show threadSekhenMay 5, 2025

@Gargron
Have you met organized Religion?
There's this thing called "The dark ages", it lasted for *quite a while* and made life pretty miserable for everyone.

In one way or the other, religion is involved in every armed conflict currently on planet earth, except maybe the Ukraine/Russia thing.

Show threadadritheonlyMay 6, 2025
@Gargron Yes, LLMs/AI should take its place of honor right next to oil and all its evil offspring (plastics, pesticides, etc.) in the Museum of Wrong Turns.
Show threadadritheonlyMay 6, 2025
@Gargron OK that's "evil" as in "just plain wrong". Religion still has its place of honor first in that exhibit of horrors.
Show threadTacitus 🇮🇪May 6, 2025

@Gargron

Hey, that's not fair, teflon actually does what it's supposed to... Sure it might also cause cancer and destroy the environment, but you can't deny that it does make your pans non-stick!