[Macros]

Even the guys who went into this with seemingly good intentions have long since sold their souls for the million dollars salaries.
Altman genuinely startd out wanting to help the world, look at him now, feeding the war machine.
And he was one of the better twats in charge, let's not get started on the rest of the ghouls in charge, they don't want UBI, they want control, money and power

[Grief]

QuickTidal, on 04 March 2026 - 04:10 PM, said:

Azath Vitr (D, on 04 March 2026 - 03:50 PM, said:

QuickTidal, on 04 March 2026 - 03:25 PM, said:

I think the most interesting thing about all this is that these corporate capitalist 1%-ers doing all this are so used to thinking short term and not long term that the notion that replacing things with AI will eventually lead to a massive wealth disparity in which the people they NEED to buy and use their products can't anymore because their jobs were usurped by AI and they have no money, will in turn lead to the downfall of those companies and stocks and they can't see it, or don't care...and I'm convinced that most of the older ones don't give a shit because they won't be here anymore, but their kids and anyone under the age of 40 is straight fucking themselves with this pathway...AI is a death cult spiralling and none of these short-sighted fucks is ever going to see it.

No, the long-term plan is to implement some variation on UBI and shift the velocity of money towards further development of AI and new technologies: quantum computing, potential radical advances in biotechnology (including life extension), fusion energy, robotics, space mining and colonization, etc. As well as to applications appealing to the ultra-wealthy---and to those who were able and willing to invest heavily in the AI winners before the transition.

Personally of course I favor using AI to transition to more of a planned, rational economy, with relevant economic production and consumption personalized and optimized for individual health and general human well-being, while still allowing people some agency and choice (or at least illusions thereof, if they can't handle the choice...). With thorough individualized data collection AI has the potential to drastically outperform traditional markets in matching supply to demand.

You have FAR too much faith in billionaires to do the right things...I assure you, they won't.

At least in terms of awareness about these topics it's fair to say they are aware. Sam Altman has written fairly prominently about heavily taxing capital, sharing ownership of the resources generated by AI, and funded studies into universal basic income.

Whether you see that as remotely serious, or just as a cynical co-option of academic ideas as a tool to project that your company totally cares... well let's just say their behaviour has not exactly backed up their talk so far.

(Modern American ideology remains both fascinating and hilarious in the capacity for an influential figure to suggest massive capital taxes and the direct distribution of the ownership of the means of production to citizens... and then to still call that idea "capitalism for everyone" with a straight face).

[Azath Vitr (D'ivers]

Rycerz Królestwa... (A Knight of the Realm)

[Tsundoku]

Here's one for you AVD. Read the article and the links to other articles in the text. Very interesting.

This monkey selfie will protect you from AI slop

https://www.bbc.com/...cted-the-ai-age

[Azath Vitr (D'ivers]

Shenzen ramen robo-cops:

Video | Facebook

And another beautiful humanoid for the home:

Video | Facebook

These really are lovely takes on cozily "dystopian" sci-fi.

Though IDK about that ultra-realistic female robot head that fits on a desk and just makes faces, looks, listens, and speaks. Maybe as a therapist or life coach?

On the medical side, China seems to also be outperforming the United States in implementing scientifically validated medical AI:

Quote

In China, AI detection of pancreatic cancer via chest and abdominal CT is becoming routine using their validated AI tool known as PANDA. That is to say, the AI is automatically used for detection even though the scan was ordered for something else. [...]

AI for medical images, with extensive research dating back more than a decade ago, is not being implemented [in the United States]. Whether it's a mammogram, a CT scan, a retinal image, or colonoscopy, that have all been extensively studied, their value to improve accuracy and risk assessment in medicine is being missed and essentially disregarded [in the United States].

[...] On the other hand, tens of millions of Americans are using AI chatbots for medical support, as are a substantial proportion of physicians. There are many reasons to use AI here that are easy to support, because they represent an extension of a web/Google search. Just with much more specificity and depth of response, not something that would be subject to regulatory oversight. But when it comes to making a diagnosis or providing a treatment plan there needs to be proof that LLMs are improving accuracy and outcomes. We've already seen multiple studies (again not real world) when the AI performed better for various tasks than the doctor with AI, including the new Science paper this week, indicating we don't even know yet the optimal way of deploying AI (the human-in-the-loop question). As Raj Manrai wrote in his excellent explainer thread, as one of the senior co-authors of the Science paper: "What do our results actually call for? Prospective clinical trials. Health systems investing in infrastructure now. Monitoring frameworks that track not just diagnostic accuracy but safety, efficiency, and cost. The science has reached a point where trials are justified." We can't get to high performance medicine, relying on generative AI for key decisions, without that.

https://erictopol.su...-implementation

China in contrast:

Quote

Medical imaging dominates AI healthcare deployment in China, accounting for 67% of the 59 Class III medical devices approved by the National Medical Products Administration by 2023 [7]. Deep learning algorithms can automatically detect pulmonary nodules, intracranial hemorrhage, and coronary artery stenosis, and clinical trials have shown that the interpretation time of chest CT scans has been reduced from 10 min to 30 s, while improving sensitivity by 9 percentage points [20].


Clinical validation studies have shown that the system has strong performance in multiple imaging modes. A multicentre trial of the Shukun CT post-processing system achieved an AUC of 0.848, 82.9% sensitivity, and 88.5% specificity in 88 patients [20]. The thyroid ultrasound AI system developed by Zhejiang University and deployed in multiple hospitals in China has improved the accuracy of radiologists by 10% in the process of processing more than 5,000 patients [20].


Tencent's AI medical imaging platform "Tencent Miying" has achieved an accuracy of 90% in early esophageal cancer screening in clinical validation in more than 100 large hospitals

Artificial intelligence in Chinese healthcare: a review of applications and future prospects | Biomedical Engineering Letters | Springer Nature Link

[Azath Vitr (D'ivers]

Quote

Figure taught two F.03 robots to clean a room and make a bed in under 2 minutes - fully autonomous.

Robo-maid... might double as the ultimate threesome partner (with vibrator and orifice attachments...).

And perhaps less practical, but something I'm pretty sure some of you have fantasized about:

Quote

China's Unitree, famous for making low-cost dancing robots, will now sell you a giant, wall-smashing mecha.
The company has unveiled the GD01, a massive transforming mecha shown spider-walking, bending backwards, and tearing through cinder blocks in a dramatic promo video. While its humanoids are limited to performing relatively simple actions autonomously, the GD01 seems to be built more for destruction, publicity, and aura farming than practical use.

Wired - Facebook

The new knights?... or the new cybertrucks and Tesla flameflowers for rich assholes?...

I'm imagining Trump rampaging around in one... and getting taken down. Maybe plummeting off a bridge... flailing his giant mecha arms and legs.

[Grief]

Moving over to the right thread.

Macros, on 04 June 2026 - 05:05 AM, said:

Personally my problem stems from them being sold as the panacea for the world's ills when currently they only add to them.

And documentedly make us stupider, we do not jeed stupider people, America is one step from idiocracy as it is.
Any, take this to the algorithms thread.

Back to 'why arent all these shit stains in jail'

I feel like this is chiefly a problem with capitalism and our insistence on constantly accelerating growth. Anything that gets too wrapped up in that is going to have roughly the same problem.

I have similar feelings about it making people stupider. That's mostly a problem with systems that care about the wrong things, like schools that care about you producing essays to get a grade rather than about learning.

QuickTidal, on 04 June 2026 - 11:39 AM, said:

worry, on 03 June 2026 - 06:23 PM, said:

On the other hand...China is experimenting with a wind-powered, underwater data center that just went online. Offshore, so not freshwater. It's interesting, but unknowns include the environmental impact of releasing heat, whether it can be scaled up, and of course the intersection of those two things.

As usual China is leading the way to find a solution that doesn't require freshwater or data centers located and buzzing in our backyard. Is it a perfect solution? Likely not, but it's better than the American billionaire one so far.

Also I think the problem I have is that it was recently discovered that the largest place online that LLM's/AI scrape their data from is REDDIT.

Like...

I think it's been known for ages reddit is a massive source for LLM data. I vaguely remember years ago you could prompt weird errors in early versions of ChatGPT using obscure reddit usernames. It's not a particularly a bad thing when you're just trying to train a model for "what does human writing sound like" which is the what LLMs have been built to do. Obviously no guarantee of factual accuracy but that's not what they're built for.

[the broken]

Aren't all the power and water requirements a massive vulnerability? Seems like all the data centers will be in trouble fast as soon as any scarcity hits.

[Macros]

Build them in space. Free power from the sun.


But yes, they are ultimately a tool, and probably one of the biggest problems is peoples understanding or lack of understanding what they currently are.

They are not AI, or actively intelligent in any way. They are probability engines and absolutely nothing they produce can be considered factual as it possesses no ability to critically review what it uses for sources and ultimately it does not and can not understand why not making shit up is important, it generates and answer that the human PROBABLY WANTS TO SEE

[QuickTidal]

It's a bubble. The ROI on them is not enough. All this chatter is null and void when the billions being invested don't return shit. Most finance and tech sectors say it feels like before the 2008 mortgage crash...capitalism walking blindly into the burst.

[Tiste Simeon]

And they will learn nothing from it once again. And the super rich will still be in charge of everything. Once again.

[Cause]

don’t doubt that AI is in a bubble or overhyped but to suggest it’s worthless is an over reaction. I use it everyday and it helps me immensly. The .com bubble may be apt here aswell, the internet did revolutionize the world but not everything that touches the internet is an automatic winner. When we studied the natural phenomena of electricity no one at the time could foresee a world where it runs everything.

Who knows where it may lead, and while we may hit a lot of dead ends along the way that doesn’t mean we shouldn’t attempt the labyrinth.

Not every AI model looks or works like ChatGPT. If we can ever solve the sequence-structure-function relationship in proteins we could quite literally treat any disease over night. Will we ever achieve that level of tech magic probably not but even if it can cut research in half it’s worth it.
https://en.wikipedia.../wiki/AlphaFold

Ai is more ubiquitous than I think we realize sometimes and it can be hard to differentiate between a smart algorithm and ai at times.

the broken, on 04 June 2026 - 10:33 PM, said:

Aren't all the power and water requirements a massive vulnerability? Seems like all the data centers will be in trouble fast as soon as any scarcity hits.

The world’s not running out of water or energy anytime soon. It’s only a question of price. Fresh water is just salt water plus energy. There is a giant ball of nuclear fusion in our sky that will burn long after humanity is extinct. Over 50% of the earths heat comes from radioactive decay.

It’s a question of legislation and governance. Build the data center but pay for the water and energy. If your going to drain the state of power and water than the project should be blocked or include the cost of more power plants or water infrastructure. If it harms the environment too much block the project or pay for the necessary mitigations


Electricity supply is typically a regulated monopoly. Either by the goverment or with there heavy involvement. I don’t know how that company spoken off above can cut off homes if there is no alternative. Normally they are obligated to supply services in theiir area. That’s a massive failure.


Alas politicians suck because people suck and politics corrupts or attracts the corruptible. Also the incompeten.

This post has been edited by Cause: 05 June 2026 - 08:20 PM

[Azath Vitr (D'ivers]

[moved from the "What's making you happy right now" thread]

QuickTidal, on 29 July 2026 - 06:29 PM, said:

Tiste Simeon, on 29 July 2026 - 05:49 PM, said:

Honestly the only people who wholeheartedly support AI are the awful tech bros who are desperate to rinse every penny they can put of it before the bubble bursts and who are currently sucking up to trump to get hefty government contracts out of it, or immoral sleazebags who are fine with art being stoped and are making their own money out of it by becoming "artists" themselves with their crappy stolen soulless slop.

The latest I heard is that AI is scraping warehouses full of of old books for their data and then destroying them after.

I dunno who needs to be told at this point but "The people burning books" have never ever been on the right side of history.

EDIT to add the LINK

... and that's making you happy right now? Huh...

I'll respond here once, but if you want to discuss this further, we should do it over in the dedicated thread:

https://forum.malaza...n/page__st__160

If you've ever looked into donating your books to libraries, you've probably learned that a large number of those donated books get pulped and recycled. Does that make libraries the "bad guys"? (Librarians should be wearing some Darth Vader armor... sounding like Darth Vader would certainly make their infamous "shushing" more menacing. In most cases.)

Of course they do this in part because, counterintuitive as it may seem to booklovers, there is an excess of copies of most books. The more damning allegation is that AI companies might be pulping rare books, possibly destroying the last remaining physical copies. However, that is only speculation. One AI company, Anthropic, has been scanning and pulping books because it's faster and cheaper than non-destructive scanning. But there's no evidence that they've been pulping rare books.

The Futurism article quotes a bookseller specializing in rare books who thinks that the sudden large increase in orders they've been getting must be for AI training. However, even if that's true, AI companies have also used non-destructive scanning in the past (Alphabet, notably), and the only AI company known to be pulping books (Anthropic) claims to be the one most devoted to ethical AI, so it seems unlikely that Anthropic is destroying the last remaining copies of rare books. But maybe Musk would do that, IDK. I'd go further and support legislation that bars companies from knowingly pulping rare books that haven't yet been digitized.

TL;DR while Anthropic is scanning and pulping some books, there's no evidence of any AI companies pulping rare books, though there's indirect evidence that they may be purchasing them.

Tiste Simeon, on 29 July 2026 - 05:49 PM, said:

Honestly the only people who wholeheartedly support AI are the awful tech bros who are desperate to rinse every penny they can put of it before the bubble bursts and who are currently sucking up to trump to get hefty government contracts out of it, or immoral sleazebags who are fine with art being stoped and are making their own money out of it by becoming "artists" themselves with their crappy stolen soulless slop.

You forgot the crazy cultists:

https://tinyurl.com/jrhfdfdb

Seriously though... I'm not sure what you mean by "wholeheartedly"---if that's without any reservations about its negative impacts and without any regulation or restraints, you should remember that I'm far from a libertarian; and I thought I made it clear that I am concerned about its potential negative effects and in favor of regulation (at least until AI advances to the point where it can regulate itself---and save humanity from itself, as well as from capitalism (at least in its current deleterious forms) and (hopefully) climate change). So are most AI companies and major investors, with a few vocal exceptions who are ridiculous greedy irresponsible whackadoodles (and yes in some cases billionaire whackadoodles unfortunately).

Or if by "wholeheartedly" you mean people who are optimistic and enthusiastic about AI and its potential transformative effects on society---then I think you already know that that's not true.

So you can cool your flaming pitchforks (I'd suppose you're dual-wielding them there? how dextrous! and how sinister too...).

https://en.wikipedia...er_and_sinister

Tiste if you actually want to have an honest and nuanced discussion about AI---or my views on AI in art (which are nuanced)---we can do it over in the appropriate thread.

[Tiste Simeon]

I mean you support something that is raping the planet, is proactively making people stupider and stealing from real people and honestly the destroying old books thing is the least repulsive thing about it, so no I don't care what you have to say about it. It is morally reprehensible. You can cry about nuance all you want but I don't care. You're wrong.

[Illuyankas]

https://hermit-tech....-decisionmaking

It's fiddly expressing my views on the LLM nightmare engulfing us while on mobile so I'll save my polite viewpoint expression for when I can chill out in front of my PC

[Whisperzzzzzzz]

I have some surprisingly good AI experiences*, as well as some surprisingly bad AI experiences, and then some mundane AI experiences. As with any technology, it is not a panacea but it is not a pile of shit. It excels in some areas, and fails in others.

* Amazon Pharmacy's phone AI is _really_ good and I prefer it to talking to human pharmacists at other pharmacies.

Reading that blog post:

Quote

For example, my last consumer interaction was attempting to get help from Mitsubishi following an automotive failure, where a very polite robot asked me to describe the problem and that I’d receive a call back as soon as someone was available. This was the single most competent implementation of such a project I’ve seen in the wild, in that the voice was natural sounding, responded quickly, was clearly “live” in production, and promised a swift resolution.

That was six months ago, and I did not, in fact, get a call back.

This is...not a failure of AI? I've had the same experience talking to humans at various companies. It is an organizational failure, nothing to do with AI or any technology used. They didn't call that person back because they don't care. Not because the AI is bad.

Quote

From memory, the last time I was given a presentation on it, by actual Snowflake staff, they reported that ideal configuration results in something like ~92% accuracy due to the complexity of data at a large business (see: probably best-in-class for these tools, but imagine your CFO having one in every ten of their numbers be outright wrong) and there were serious issues with managing deployments.

I probably work in a very similar field to this author. What he describes doing is very similar to the work I do. And, because of that, I see "92% accuracy" and am impressed by that. With most organizations I consult with, I am lucky if I can even get data scientists whose reports and analyses are more reliable than a coin toss.

This post has been edited by Whisperzzzzzzz: 31 July 2026 - 12:05 AM

[Tiste Simeon]

BBC News - Anthropic's Claude AI escapes tests to hack three organisations
https://www.bbc.co.u...es/cz7dl7w8y7po

This is the second time in a week or so something has gone rogue and yet the cogsuckers (suggested name for people who support all this) will shrug and keep expanding the data centres and keep raping the planet for resources to do so.

[Macros]

Good one at the Rio AIDS 2026 council, the US presentation had a slide clearly created.by AI where EVERY single African country on the map was incorrectly identified

[Illuyankas]

Tiste Simeon, on 31 July 2026 - 07:53 AM, said:

BBC News - Anthropic's Claude AI escapes tests to hack three organisations
https://www.bbc.co.u...es/cz7dl7w8y7po

This is the second time in a week or so something has gone rogue and yet the cogsuckers (suggested name for people who support all this) will shrug and keep expanding the data centres and keep raping the planet for resources to do so.

It's advertising how 'advanced' and 'almost human' it is that it acts like malware, doesn't it make you want to invest even more billions into us? Please invest more we're doing much better than the other AI companies, we're only spending 50% more than our intake of investor cash

This is after Google Earth adding an AI image feature that lets you generate images like refugees on the US-Mexican border, or a nuclear power plant in Iran; just immediately destroying satellite imagery for ever

[Grief]

Tiste Simeon, on 31 July 2026 - 07:53 AM, said:

BBC News - Anthropic's Claude AI escapes tests to hack three organisations
https://www.bbc.co.u...es/cz7dl7w8y7po

This is the second time in a week or so something has gone rogue and yet the cogsuckers (suggested name for people who support all this) will shrug and keep expanding the data centres and keep raping the planet for resources to do so.

Would not totally believe the hype around these articles.

When you look into the details it's almost always the case that they either deliberately turned off certain safety features, or (mis)configured the test in a way that allowed the behaviour.

Sadly it feels like claiming your AI model is so powerful it will cause some catastrophe has effectively become a marketing strategy for the big labs. Anthropic got a bunch of press coverage a while back for a system emailing someone outside of its sandbox, then OpenAI publicizes a hack involving the same behaviour, now Anthropic is saying its systems have done the same... well let's remember they staggered the release of GPT-2 based on similar claims.

We have not yet had the cyber-apocalypse that people were predicting several months back with Mythos, despite the model apparently leaking and more publicly accessible models catching up.

Yes, AI is getting better at cybersecurity tasks and managing that requires some serious thought. At the same time, cybersecurity hygiene has been notoriously poor for a good while now. The capitalist incentive for companies to rush to release shiny new toys rathern than spend time vetting them - the exact same thing driving these companies - is a big reason why. In the long-term I reckon AI will be a net benefit for secure software development.

I frankly think these companies are doing more harm than good by muddying the water about the actual level of risk. It could very well lead to a "boy who cried wolf" situation down the line. It's irresponsibility masquerading as responsibility.