Salta al contenuto principale





Supereremo con l’IA le correnti gravitazionali. “Universo senza segreti”
https://www.repubblica.it/cronaca/2025/09/05/news/onde_gravitazionali_intelligenza_artificiale_gran_sasso-424827230/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Cronache italiane - La Repubblica @cronache-italiane-la-repubblica-repubblica



Processo a Ciro Grillo, corsa contro il tempo per udienza e sentenza il 22 settembre
https://genova.repubblica.it/cronaca/2025/09/05/news/processo_ciro_grillo_sentenza_22_settembre_csm-424826904/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Cronache italiane - La Repubblica @cronache-italiane-la-repubblica-repubblica






NSFW 18+ Nudity
  • Sensitive content
  • Parola filtrata: nsfw



playing no mans sky again and every time i get out of a vehicle and end up on the wrong side i am reminded that br*tish people made the game


Anyone else having controller troubles with #silksong on #steamdeck ? I'm using an xbox controller but the game seems to have trouble with inputs from the analog stick. I can't jump or attack when moving
in reply to Steffo

@steffo
Try setting your left joystick up like this in the controller configurator, that solved my issue



6 novos personagens que estrearão em Demon Slayer Castelo Infinito
https://www.tecmundo.com.br/minha-serie/602000-6-novos-personagens-que-estrearao-em-demon-slayer-castelo-infinito.htm?utm_source=flipboard&utm_medium=activitypub

Posted into TecMundo @tecmundo-TecMundo




The first time an app is deployed to production you learn very quickly that your logging approach is insufficient.


ok; i'll say it.

MY life is more important than a fetus'.

i don't know why we tiptoe around saying this explicitly when this is what we (the folks who believe in not being forced to be pregnant) mean.

anyway, that's what popped into my head when i read this.

texasobserver.org/texas-legisl…



Sensitive content

in reply to れんくん

The image shows an open blue umbrella lying on the floor in an indoor setting. The umbrella is positioned with its canopy facing upwards, revealing its fabric and metal ribs. The handle is visible in the center, and the umbrella appears to be in good condition with no visible damage. The background consists of a plain white wall and a corner of a white cabinet or structure, suggesting the umbrella is inside a building, possibly in a hallway or storage area. The lighting is soft and even, indicating an indoor environment with artificial lighting.

Provided by @altbot, generated privately and locally using Ovis2-8B

🌱 Energy used: 0.133 Wh



For the forty-eighth anniversary of Elsie Carlisle's death:

"When That Man Is Dead and Gone." Words and music by Irving Berlin (1941). Recorded in London on April 9, 1941 by Elsie Carlisle with orchestral accompaniment directed by Jay Wilbur. Rex 9960 mx. R-5566-1.

#78rpm

elsiecarlisle.com/when-that-ma…



Unemployment rate for Black women is now the highest it’s ever been in the history of the United States. It’s bad for everyone right now… but as usual…. ESPECIALLY for Black women.

Let’s be clear. This is blatant retaliation against the #92Percent who did not, and will not ever vote for the nazi kakistocracy.

…anyone out there prepping the refugee room? Because yikes.

reshared this



while clearing out an old usb drive in order to reboot a found computer, i discovered some old photos i took in the desert with a friends camera that had this weird sci-fi filter. goodbye silent sunday, hello #manicMonday

reshared this



One of the best things about this place is that if you are dealing with something difficult, someone will immediately offer tangible, concrete help, in addition to compassion. This is, I think, what separates this social media from every other.

People here support each other in genuinely holistic ways: in each other's emotional, mental, physical, and spiritual health, and even in each other's interests. Frogs or photography, windows or waterfalls, trees or tech, it's all not only welcome, but often commonly shared. You can say "Look at this cool leaf!" and not only does no one make fun of you, but many ooh and ahh and agree.

We've had a few trolls and spammers of late, but we all work together to limit them.

So, give yourselves a pat on the back. You're doing something meaningful here.

reshared this



Nothing to see here (or in #Antarctica)

Scientists issue warning on 'unstoppable' phenomenon that could have global impact: 'Catastrophic consequences'

by Matthew Swigonski, 9/3/2025

"According to a recent study, a team of researchers has determined that the West Antarctic Ice Sheet is at an increased risk of completely collapsing due to continually high levels of atmospheric carbon dioxide. Even with progressive shifts in dirty fuel usage, realistic reductions in CO2 may still not be enough to prevent more abrupt changes."

msn.com/en-us/weather/topstori…

World’s biggest iceberg, A23a, has broken up

The A23a iceberg has been closely tracked by scientists ever since it calved from the Filchner-Ronne ice shelf in Antarctica in 1986.

By Lianne Kolirin and Issy Ronald, Sep 4, 2025

" 'The world’s largest iceberg is 'rapidly breaking up' into several large 'very large chunks,' scientists from the British Antarctic Survey (BAS) have said.

"Previously weighing nearly a trillion metric tonnes (1.1 trillion tons) and spanning an area of 3,672 square kilometers (1,418 square miles) — slightly bigger than Rhode Island — the A23a iceberg has been closely tracked by scientists ever since it calved from the Filchner-Ronne ice shelf in Antarctica in 1986. "

accuweather.com/en/weather-new…

Scientists looked deep beneath the #DoomsdayGlacier. What they found spells potential disaster for the planet

By Laura Paddison, September 20, 2024

"Scientists using ice-breaking ships and underwater robots have found the #ThwaitesGlacier in Antarctica is melting at an accelerating rate and could be on an irreversible path to collapse, spelling catastrophe for global sea level rise.

"Since 2018, a team of scientists forming the International Thwaites Glacier Collaboration, has been studying Thwaites — often dubbed the '#DoomsdayGlacier' — up close to better understand how and when it might collapse.

"Their findings, set out across a collection of studies, provide the clearest picture yet of this complex, ever-changing glacier. The outlook is 'grim,' the scientists said in a report published Thursday, revealing the key conclusions of their six years of research.

"They found rapid ice loss is set to speed up this century. Thwaites’ retreat has accelerated considerably over the past 30 years, said Rob Larter, a marine geophysicist at the British Antarctic Survey and part of the ITGC team. 'Our findings indicate it is set to retreat further and faster,' he said.

"The scientists project Thwaites and the Antarctic Ice Sheet could collapse within 200 years, which would have devastating consequences.

"Thwaites holds enough water to increase sea levels by more than 2 feet. But because it also acts like a cork, holding back the vast Antarctic ice sheet, its collapse could ultimately lead to around 10 feet of #SeaLevelRise, devastating coastal communities from Miami and London to Bangladesh and the Pacific Islands.

"Scientists have long known Florida-sized Thwaites was vulnerable, in part because of its geography. The land on which it sits slopes downwards, meaning as it melts, more ice is exposed to relatively warm ocean water."

cnn.com/2024/09/20/climate/doo…

reshared this




Democratic Leadership Still Hasn’t Caught Up to the Party’s Base on Gaza
https://inthesetimes.com/article/dnc-israel-gaza-palestine-democrats-ceasefire-genocide?utm_source=flipboard&utm_medium=activitypub

Posted into In These Times @in-these-times-InTheseTimesMag




So, on the topic of AI - if we were to assume that AI would follow approximately the same trend as the internet did - a massive early bubble, followed by gradually becoming more important and more and more of a Big Deal - what parallels could we make? If we assume that it will move faster, is there any way we could measure our progress along that path, and estimate it's speed? :/
in reply to Angle🖇

Like, they are figuratively building the terminators *as we speak*, my guy. This is a big deal, we can't afford to just ignore it. -_-
in reply to Angle🖇

I'll admit that I don;t know what to do about it... But acknowledging that the problem exists, telling other people about it, and maybe building a consensus that it needs to be addressed would be a good first step... T_T
in reply to Angle🖇

@Angle🖇 My complete and utter reject of this technology isn't born out of underestimating it. Even now in its "infancy" stage, it has been shown to wreck massive havoc on several intellectual disciplines for the benefit of ridiculously few people. I don't need more evidence from more mature applications.

You're talking about the terminator, but I'm just talking about LLMs/GenAI which is an outsized application of machine learning. And like I said, the benefits are felt by extremely few people while the downsides are felt globally. I expect this imbalance to be its undoing, no matter the validity of the few niche uses that do not only benefit mindbogglingly rich people.

To go back to asbestos, it doesn't matter if your building is cheaply fireproof if the people get cancer just by existing within its confines.

in reply to Hypolite Petovan

I will also note, it's not enough to just 'reject' the technology - you have to actually do something about it. Otherwise you're just going to find yourself getting shoved into a smaller and smaller corner. :/

Like, I really want to see people actually learning about the technology and how it works and what it's good for - but at a bare minimum, learning what other people are doing with it and how you can prevent them from fucking you over seems important...

in reply to Angle🖇

You'll only get shoved into smaller and smaller corners if we, collectively, allow that. Gen AI is destroying industries and lives and does so entirely through theft for the benefit of a tiny, tiny minority of people.

"Just ignore it" is a viable tactic in some circumstances. Blockchain and NFT guys promised just as much revolution as the current AI bubble (cont)

in reply to KolaMagpie

but just like AI, those fads required enough people to sign on. And it didn't happen, and it faded.

AI isn't inevitable, unless we make it so. It's already starting to crumble as profits fail to appear, as lawsuits start catching up. The people pushing for it will find something else to try, undoubtedly.

And personally I think "trying to find the good uses" is just giving it legitimacy. Any good use will be corrupted by capital.

in reply to KolaMagpie

AI has a way, way lower bar, though. Like, you can generally tell if you're getting caught up in blockchain nonsense, whereas telling if a company is using AI is a lot harder. Yeah, a lot of it is really shit right now, but you shouldn't assume it's going to stay that way forever.

As for good uses, the first requirement is to cut capital out of the equation as much as possible. Ideally, you want AI developed by the people, for the people.

in reply to Angle🖇

I mean, Ideally, I want no genAI at all. I don't see a use case that doesn't just waste time and resources.

What is the point of producing lots of text? What social purpose does that perform? IDK if its built "by the people" and "for the people" meaningless but grammatically correct text is just chaff.

And yeah Capital will keep trying I don't doubt that. They do it with child slavery too. Anything to make a profit.

in reply to KolaMagpie

And like child slavery you can't be like "well there might be a good use case!" (I'm not saying you're in any way in support of CS, just using an example). And when laws wont stop it it's up to us to deny it a part in our society.

And if we wont do that, then we get the bullshit.

in reply to KolaMagpie

I don't think individual rejection actually accomplishes that much, though.

Like, I use AI to help me write Minecraft mods - it's far from perfect, but it's still a 10-20% productivity boost, even in the current early prototype form that I have access too. And if I weren't telling people about, nobody would know. Even with me telling people about it, I think it's extremely unlikely that I face significant pressure over it - and that's just me! I'm a small player here. :/

Questa voce è stata modificata (2 settimane fa)
in reply to Angle🖇

@KolaMagpie @hypolite The big companies have so many more resources to cover up and deflect their AI bullshit, the same way they do all their other bullshit. I just don;t think your rejection means anything. :/
in reply to Angle🖇

@hypolite I get that. But the same argument could be used and is used to justify all sorts of bullshit.

Individual action doesn't mean anything, so why bother. One person shopping Walmart doesn't make the community store fail. But everyone believing that, does make it fail.

One person trying to win rights never works, so why try at all?

I'd personally be okay with 20% fewer minecraft mods, in exchange for a healthy society.

in reply to KolaMagpie

@KolaMagpie @hypolite But that's not actually the deal on offer. Like, I'd look at it like guns in a warzone. Should everyone have them? Probably not. But *some* people absolutely should, and a blanket 'guns are bad' when the Russians are coming over the horizon won't help anyone. -_-
in reply to Angle🖇

@hypolite But this isn't a warzone, not like that at least. In war, if you don't have a weapon, you're a victim. In society, we are only victims /if we let ourselves be/ (please don't take this out of context).

NFT bros wanted us to be victims, but we rejected that because that future sucks. We /do/ have the power to do that. Collectively. And enough individual people were like "fuck that" that it wasn't the future. The same can be true of AI.

in reply to KolaMagpie

@hypolite But I think the biggest problem is getting that across. Getting people to think in terms of societal cost-benefit.

Sure, meat is delicious. Am I okay with eating meat, if it means wasting 54% of calories grown in the USA?

Sure, Walmart is cheap. Am I okay with having more things if it means my town and economy is destroyed?

Sure, AI is convenient. Am I okay with more websites and movies if it means the destruction of professionals?

in reply to KolaMagpie

@KolaMagpie @hypolite I mean, the average person is answering a definitive 'yes' to the first too. So right there, we're off to a bad start. All my points about the futility of individual rejection apply to those things too. :/
in reply to Angle🖇

@hypolite Sure, in which case, we're fucked. Cuz Capital will /always/ be able to out-spend, out-propoganda, out-/anything/ us. Capital are experts at making us complacent.

I'd posit that the lack of education and leftist outreach during the rise of those industries lead to their current monopolies. And I'd rather try for a better future than give up and give more power to the weapons that oppress us.

in reply to KolaMagpie

@KolaMagpie @hypolite I'm not entirely sold? It's not just a matter of spending, propaganda, organization, etc, but the combination of those with other factors - but that still requires that we put forth the spending, organization, agitprop, and appropriate other factors to make things work. :/
in reply to Angle🖇

@KolaMagpie @hypolite Like, vegans and vegetarians and their assorted fellow travelers have been rejecting meat for quite a while, but they've barely made a dent in the meat industry. If you want to make headway on this, then it takes more than individual rejection, you have to find a way to make it appealing to the average person, and moralizing just won't cut it. :/
in reply to Angle🖇

@Angle🖇 @KolaMagpie I don't know about that, in most cities restaurants will have at least a well thought-out vegetarian option, which is probably more than you could have expected 20-25 years ago. There are thriving vegan restaurants, which is a good sign that there's some significant adoption. Speaking of dent, meat production has decreased significantly in Europe between 2020 and 2023, and stagnated in the US. Only in Asia did it rebound from the pre-pandemic level. The dent is there, see ourworldindata.org/grapher/glo…
in reply to Hypolite Petovan

@hypolite @KolaMagpie Eh, that's great for the vegetarians, but most places are showing roughly a 100% increase in meat production since 1990. I wouldn't call that a victory for vegetarianism. -_-
in reply to Angle🖇

@Angle🖇 @KolaMagpie The important isn't the absolute number, but the comparison. Meat production has been growing way slower in developed countries, while it still grows in developing countries. For me this means that at a certain level of development, meat has become way less important as a status symbol.
in reply to Hypolite Petovan

@hypolite @KolaMagpie As a status symbol, sure. But status aside, meat is tasty and it's easier to have a healthy diet with it than without it. And until there's a decent substitute, I don't expect consumption to drop much.
in reply to Angle🖇

@Angle🖇 @KolaMagpie No, you're right, which means even a stagnation in the production means for me that vegetarian/eco-conscious lifestyles are gaining ground as there isn't much of a reason for this production to slow down otherwise.
in reply to Angle🖇

@hypolite To circle back to my original point: the people rejecting AI aren't your enemy nor are they the antithesis for keeping society safe from the AI crisis.

Just like vegans aren't the enemy when fighting wasteful land use policies.

I stand by that AI, currently, isn't inevitable so rejection might be by itself valid. It might need more active work than NFTs did, but it's possible it can be nipped in the bud.

in reply to KolaMagpie

@KolaMagpie @hypolite It's not the rejection that bothers me, per se, but rather the insistence that it's not capable and that rejection is sufficient. Like, a vegan is fine, but a vegan who insists 'Meat is bad for you actually, the meat eaters will all die out any day now' is much less so - especially if they have lots of company and like, you actually want to cut down on meat consumption. -_-
in reply to Angle🖇

@Angle🖇 @KolaMagpie I believe we disagree on capabilities. I believe GenAI is capable of mesmerizing an entire generations of humans into believing a human-shaped parrot serves them using a language they are familiar with.

On the other hand I don't think GenAI is capable or will ever be capable of writing or drawing masterpieces, generating production-ready code for complex real-world applications, or make any progress in research and development. I believe its widespread adoption would lead to an overall stagnation of the artistic, scientific and technologic production, while decreasing humans' very ability to learn and make stuff ourselves.

The capability that is currently being developed is fooling the training staff that it's reaching its arbitrary metrics goals. Machine Learning has always been able to do that, and the introduction of natural language only reinforces this because as humans, we strongly associate language with intelligence.

But even this is showing its limits. To "increase productivity", training staff is now running the output of a model by another GenAI model to check for "quality". Predictably, the trained model has started optimizing for this very task and outputs very well-formed nonsense that the other model is applauding as the pinnacle of literacy: futurism.com/gpt-5-literary-ou…

in reply to KolaMagpie

@KolaMagpie @hypolite Ehhhh, it's closer to that then I'd like. Like, if a company uses AI to automate your job or whatever, what are you going to do about it? If they use AI to automate someone else's job, how will you even know, let alone do something about it? :/
in reply to Angle🖇

@hypolite Tell people about it! Be a society about it. Target's imploding cuz people stopped wanting to go there because they made anti-societal decisions.

Unionize. Organize. We are not helpless.

in reply to KolaMagpie

@KolaMagpie @hypolite ...Did Target actually do something noticeably worse than other companies? I mean, yeah, communicating and rejecting the companies that seem especially bad is something, but I still think we need something more organized and larger scale... :/
in reply to Angle🖇

More organized and larger scale's great, but that doesn't happen without individual action or awareness. The union wars didn't happen (and the subsequent workers' rights) because people were all "darn, the company's forcing me to live in their house but there's nothing I can do about it" or "darn wages suck but what can I do about it?" or even "darn they're gonna throw bombs at me so why bother trying?"
in reply to Angle🖇

@KolaMagpie @hypolite And sure, current AI is far from automating jobs wholesale - but that just makes detection even harder. Like, a company uses AI and hires 20% fewer people then they would have otherwise... How do you know? What do you do? If the majority of companies do that, how do you possibly address things? :/
in reply to Angle🖇

@Angle🖇 @KolaMagpie I write software for a living, and I would never use GenAI at my job for a few reasons:
1. I like to write code.
2. I don't like to read code someone else wrote.
3. I'm not confident I would be able to measure productivity change with GenAI. Studies show that people who believe they are more productive with AI actually aren't. Because that's what GenAI is good at and optimized for, giving the change about its actual limited capabilities.
in reply to Hypolite Petovan

@Angle🖇 @KolaMagpie oh, and I forgot:
4. I don't know what code it's been trained on, if their authors are aware and consenting to their work being used this way.
in reply to Angle🖇

@hypolite Like, if some guy has a fucking gun, and looks like he's bad news, and I go 'gee that looks like it could be trouble, maybe we should do something about it, or prepare or whatever', and your response is to cross your arms and stick your nose in the air and go 'I don't like guns', well I'm sorry but... :/
in reply to Angle🖇

@hypolite All around me I see people rapidly developing more powerful and more capable AI, while the people I would expect to turn to for help are sticking their heads in the sand and going "LALALA AI can't do anything it's not important nothing to worry about LALALA!", and I just want to fucking scream. T_T
in reply to Angle🖇

I've been struggling with this. like, what's "more powerful" mean? I play with local models, which is probably ethically equivalent to eating meat. there's that new swiss one you have to have a huggingface account and a newer GPU to use, so even though it's fully open-source, it's still questionable. I compared qwen3-coder and qwen2.5-coder and qwen2.5-coder seems better at x86 assembly, but, I'm not a good judge, and
in reply to upbeat estimation software

the kind of "powerful" I see is more like, "given access to more systems", AI weapons and shit, not "better at creating the illusion of thought" (or thinking, or whatever they do. transformer models are useful, demucs is fun, beat_this is unbeatable for beat detection, image generation seems... idk, more capable? but maybe rejecting it /is/ doing something about it, and I'm more a part of the problem than I'll admit.
in reply to upbeat estimation software

@amsomniac Eh, there are some things where it could be worth rejecting it? But I'd look at it like guns in a warzone. Should everyone have them? Probably not. But *some* people absolutely should, and a blanket 'guns are bad' when the Russians are coming over the horizon won't help anyone. -_-
in reply to Angle🖇

@Angle🖇 @upbeat estimation software Again, I have to reject the analogy. GenAI is (only?) good at making people believe it is competent, regardless of their actual competency. What kind of people do you think wants to wield such "weapon"? What kind of people do you think are the prime targets?

Using local models isn't fighting back against megacorps (I assume the Russians in your military analogy), it's potentially a softer self-subjugation with a much smaller energy footprint. Which is still a plus in my book.

in reply to Hypolite Petovan

@hypolite @amsomniac ...I'm sorry, but you're just fucking delusional on this. You're noticing a handful of obvious failures and insisting that's all there is, because that's what you want to believe. Here's a video examining the effects:

youtube.com/watch?v=yhpyHV1iz0…

TLDR: AI absolutely can be useful if you use it right, and it already *is* transforming our economies in big ways. We will need to figure out how to grapple with it, and ignoring it won't get you anywhere.

in reply to Angle🖇

@hypolite @amsomniac If you don't want to use it, fine, but you will still need a better strategy than just turning your back and sticking your nose in the air.
in reply to Angle🖇

@Angle🖇 @upbeat estimation software Nope, I did my due diligence, I learned how the tech work, what it is capable of and mostly not capable of, I've been impressed with some of the output (mainly the generated video) but I don't want any part of it.

There's no valid strategy for me that involves actually using the stuff in any form. Just like drugs, you don't have to try them to know how they work and how harmful they can be. I don't judge too harshly people who casually use chat bots or GenAI because they are designed to be enticing and I know there are valid niche use cases, but I won't get near it myself.

I make sure to make my point known when the topic comes up, and that's pretty much all I can do. Like I said, it might reduce my opportunities, but I prefer the idea of finding my people in the corner rather than submitting to this humankind-sucking trend. Either it will come to pass and things will keep on for me practically unchanged, or it will reign and I don't want to have any part of it.

in reply to Hypolite Petovan

@hypolite @amsomniac ...I mean, I took my dexmethylphenidate this morning. 5 MG, extended release. Plus 100 MG caffiene. Helps me focus and get moving. :/

So, yeah, drugs can be useful, actually, if used correctly. Lots of finicky details, of course, and a lot of people probably shouldn't try and use them, but they're far from useless. A good parallel to AI, honestly. XD

in reply to Angle🖇

@Angle🖇 @upbeat estimation software Still not. Medicine (which I separate from drugs even if there's a single word for both 😖) has to undergo extensive clinical trials before being put on the market which AI currently isn't bound to.

Would you use a new bootleg drug its manufacturer says it will cure all your ills if only you trust them, although they didn't go through any serious trials? This is what LLMs feel to me at the moment. I don't doubt it has useful applications, LLMs existed before chat bots and was pretty useful as a first pass for translation for example. I've happily used DeepL for years, for example.

But I don't want to have to talk to machines, it just isn't the most efficient way of interacting with them and it tricks people into awkward situations where they trust the machine based on the presentation of the output rather than the output accuracy itself.

in reply to Hypolite Petovan

@hypolite @amsomniac Drugs are just medicine used wrong. I could walk outside and start selling the same pills I use, straight from the pill bottle, and bam, I'd be a drug dealer. Not something I want, but it's a thin line. :/

Of course, for some chemicals, there an argument to be made that there is no good use for them - but that doesn't mean we can dismiss all chemicals. -_-

in reply to Angle🖇

@hypolite @amsomniac

And I'm kinda iffy about talking to machines too? I use them in Cline, a Codium plugin, mainly, where I can pretty much just give them a few sentences, shove some files at them, and have them figure out the rest. Summarizing error logs or code, suggesting what might be the cause of problems I encounter, that sort of thing.

in reply to Angle🖇

@hypolite @amsomniac As for the manufacturers, yeah you should absolutely be picky about what chemicals you accept, and from who. But jumping from there to "Chemistry bad! No chemicals!" is just too much of a leap. :/
in reply to Angle🖇

@hypolite @amsomniac Oh, I would be interested in hearing about the equivalent of clinical trials for AI? I mean, there's an argument that that treatment has made medicine way worse and less accessible, which I think is broadly true... But if it protects us from bad medicine, maybe it's worth it? I dunno. I am would generally lean towards a bit less regulation for medicine... But not too strongly... :/
in reply to Hypolite Petovan

@hypolite Less options, for one thing. I'd love to get my dexmethyphenidate in 1 MG increments, but with the limited suppliers, there's little reason for them to diversify and offer me that option.
in reply to Angle🖇

@hypolite Also less competition means less pressure to drive down prices. Lots more competition to sell caffeine tablets, cause they're less regulated, so I can buy a giant jar of those on Amazon for $10 and have them delivered to my door, whereas the meds cost way more, for way less, and I have to pick them up from a pharmacy.
Questa voce è stata modificata (1 settimana fa)
in reply to Angle🖇

@Angle🖇 Ok, yes, I see what you mean.I believe it is the cost of drug safety, although if pharmaceutical research was a public service and patents were free, there would be better availability without sacrificing safety through deregulation. Drat, I criticized capitalism again.
in reply to Angle🖇

@hypolite @amsomniac Like, the early industrial revolution analogy really works, IMO. You're in the position of a worker seeing the new machines and going 'These things are terrible! They do bad work, they cost too much, they're trying to take my job, I saw one grind a small child into paste right in front of me! I hate them!"

It's a fair viewpoint. From the perspective of the worker, those things are all true - his mistake is in thinking that they're universal and will never change.

in reply to Angle🖇

@hypolite @amsomniac I'll admit I'm not entirely sure where to proceed - if you find yourself advising that early industrial worker, what do you tell him? He's not in a good position... I do think the truth is a necessary starting point, though. "Sorry dude, but the machines will eventually do good work, or good enough work at least, and they will be economically viable. They will eventually come to form the backbone of the economy. They will reshape your world, whether you like it or not..."
in reply to Angle🖇

@hypolite @amsomniac Probably not what he wants to hear. What can he do about it? I'm not sure. If I found myself dropped back at the dawn of the industrial revolution, I'd probably try and organize some kind of alternative to capitalism? Some kind of decentralized, grass roots economic system, with the power as close to the people as possible... Though I'm not sure it *is* possible. Maybe power is always destined to centralize, and the best you can hope for is a center that you can live with?
in reply to Angle🖇

@hypolite @amsomniac But standing at the dawn of the industrial revolution and thinking you can just skip out on the industrial machinery... Yeah, I'm sorry, but no. There's no way around it, the only way out is through. -_-
in reply to Angle🖇

@Angle🖇 @upbeat estimation software I understand the appeal of the Industrial Revolution analogy, but I'm weary of it. On one hand, we know the exact way it went down for industrial machines. They became more precise and safer to use for workers, and then they got relocated to developing countries to save even more on labor. But they were things that needed to be engineered and built by people who built the skills to make them more precise and accurate with safety features.

On the GenAI/LLM front however, we're building global black boxes that aren't engineered for accuracy but for human likeness. No industrial machine engineer was ever convinced their design was their friend and deserved personhood rights.

Since they are global, we all have a personal choice whether to use them or not, unlike industrial machinery that had an expensive upfront cost for business owners only.

Since they are black boxes, we have no reliably way of increasing the precision of the models or measuring the accuracy of their output, and that isn't even an optimization goal for the three main players in the space.

And then for safety, using local models sounds better but they are purposefully mesmerizing machines made to exploit human instincts, so it is only marginally safer.

in reply to Angle🖇

and we're still in the era where guns regularly blow up in the user's face >_<
in reply to upbeat estimation software

@amsomniac Ehh, okay, yeah maybe a little bit. Only if you don;t know what you're doing... But then, most people don't. XD
in reply to Angle🖇

@Angle🖇 @Nick As for what to do about it:

  • friends don't let friends use LLMs as a baseline.
  • Criticizing the systemic harm while not letting the few niche valid use cases weaken the larger point.
  • Shedding the assumption it's there to stay. Rejecting the false marketing promises in favor of real, current harm.


The tech already is unpopular after just a few years of mass deployment, so the only thing currently sustaining it is the massive sunk costs by the handful of companies who poured everything they had into an intangible moon shot.

in reply to Hypolite Petovan

@hypolite @nicksalt I think you should be picky about what AI you use, and what for... But it *is* a powerful tool, if you use it right. So, skipping out on it entirely is like finding yourself in a warzone and refusing to touch any guns, or worse, insisting that nobody else does so. Nooot necessarily a good decision.
in reply to Angle🖇

@Angle🖇 I understand what you are trying to get at with the warzone analogy, but I don't think it reaches the mark. Mainly because the current crop of AI directly exploits human weaknesses: our tendency to look for humanity anywhere and our longing for convenience in all things. This means the only people wielding it as a weapon are the megacorps. Individuals running homegrown models aren't "fighting back", they just contribute slop at a much smaller scale.

This also means to me that the only way to defeat the megacorps is to reject their premise entirely. No, GenAI isn't an inexorable future, no, I do not want any of its output in any of my workflows, and yes, I prefer humans writing text, drawing images or coding slower and more deliberately. For me, running local models is already giving them the benefit of the doubt, which is more than I want to do.

To take your analogy literally, not everybody in a warzone takes up arms. You still need plenty of civilian support, and that's what I feel closer to than direct combat if it even meant something in the GenAI paradigm, which I don't think it does.

in reply to Hypolite Petovan

@Angle🖇 Ah wait, I found a case where you can somehow fight fire with fire: theregister.com/2025/09/03/ai_…

But for me my main policy would rather be to avoid companies using AI at all. I know this limits my opportunities and I'm not impatient for my possible layoff which will throw me into the IT job search meat grinder, but I'd still favor direct approaches at tech event and mixers rather than looking to automate job search using AI. I'm hoping that I would find my people by doing the legwork and standing firm on my principles, but I'm not eager to see them tested in real life.

Questa voce è stata modificata (2 settimane fa)


Want to be beautiful inside and out? Make sure you boycott ugly brands like #Unilever #Revlon and #Maybelline #MAC that use #palmoil in their #makeup. Search for #PalmOilFree brands now ⏬ and #BoycottPalmOil 🌴🩸🤮☠️🙈⛔️#Boycott4Wildlife @palmoildetect.bsky.social palmoildetectives.com/2021/02/…




Ventisei Paesi in difesa di Kiev. Trump fa pressione sulla Ue: “Basta comprare petrolio russo”
https://www.repubblica.it/esteri/2025/09/05/news/vertice_volenterosi_parigi_news-424827238/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Notizie dal mondo - la Repubblica @notizie-dal-mondo-la-repubblica-repubblica



Lisbona, l'italiana ferita: “Lo schianto, poi il terrore. Sentivo il gemito dei feriti intrappolati”
https://www.repubblica.it/esteri/2025/09/05/news/lisbona_incidente_funicolare_gloria_news-424827191/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Notizie dal mondo - la Repubblica @notizie-dal-mondo-la-repubblica-repubblica


in reply to JL Johnson

I've been to a lot of restaurants with veggie burritos that had lard in the beans or chicken in the rice. It's nice of them to mention it (but maybe, they shouldn't name it for a lifestyle it isn't appropriate for?)


Nella scuola di Gerusalemme dove arabi ed ebrei combattono la segregazione
https://www.repubblica.it/esteri/2025/09/05/news/la_scuola_nella_quale_studiano_arabi_e_israeliani-424827136/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Notizie dal mondo - la Repubblica @notizie-dal-mondo-la-repubblica-repubblica




Someone: "I'm going #PalmOil free for the #Orangutans"

my brain:
don’t say it
don’t say it
don’t say it
don’t say it
don’t say it

Me: "Maybe you could go plant-based too for the dozens of species lost every single day?" 🦧

#OrangutanDay
#WorldOrangutanDay





The Earth is a wild place.

One day you wake up and learn that blues thrives in the Sahara, Turkish funk is a full-blown genre, and hard rock is practically a science in Eastern India.

Questa voce è stata modificata (2 settimane fa)


Radchenko: “Putin non fermerà il conflitto finché non otterrà la resa, vuole solo umiliare Zelensky”
https://www.repubblica.it/esteri/2025/09/05/news/sergey_radchenko_intervista_vertice_volenterosi-424827015/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Notizie dal mondo - la Repubblica @notizie-dal-mondo-la-repubblica-repubblica



Il Papa vede Herzog, è gelo: “Proteggete tutti i palestinesi”
https://www.repubblica.it/esteri/2025/09/05/news/il_papa_incontra_herzog-424826992/?utm_source=flipboard&utm_medium=activitypub

Pubblicato su Notizie dal mondo - la Repubblica @notizie-dal-mondo-la-repubblica-repubblica



BREAKING: New York Governor Kathy Hochul is set to sign an executive order allowing pharmacists to prescribe and administer Covid-19 vaccines to anyone who requests them, regardless of age or underlying conditions.

reshared this