179870040
submission
mspohr writes:
After Palisade Research released a paper last month which found that certain advanced AI models appear resistant to being turned off, at times even sabotaging shutdown mechanisms, it wrote an update attempting to clarify why this is – and answer critics who argued that its initial work was flawed.
In an update this week, Palisade, which is part of a niche ecosystem of companies trying to evaluate the possibility of AI developing dangerous capabilities, described scenarios it ran in which leading AI models – including Google’s Gemini 2.5, xAI’s Grok 4, and OpenAI’s GPT-o3 and GPT-5 – were given a task, but afterwards given explicit instructions to shut themselves down.
Certain models, in particular Grok 4 and GPT-o3, still attempted to sabotage shutdown instructions in the updated setup. Concerningly, wrote Palisade, there was no clear reason why.
179826816
submission
mspohr writes:
One exception to the industry-wide lethargy is the engineering team that designs the Signal Protocol, the open source engine that powers the world’s most robust and resilient form of end-to-end encryption for multiple private chat apps, most notably the Signal Messenger. Eleven days ago, the nonprofit entity that develops the protocol, Signal Messenger LLC, published a 5,900-word write-up describing its latest updates that bring Signal a significant step toward being fully quantum-resistant.
The complexity and problem-solving required for making the Signal Protocol quantum safe are as daunting as just about any in modern-day engineering. The original Signal Protocol already resembled the inside of a fine Swiss timepiece, with countless gears, wheels, springs, hands, and other parts all interoperating in an intricate way. In less adept hands, mucking about with an instrument as complex as the Signal protocol could have led to shortcuts or unintended consequences that hurt performance, undoing what would otherwise be a perfectly running watch. Yet this latest post-quantum upgrade (the first one came in 2023) is nothing short of a triumph
Outside researchers are applauding the work.
“If the normal encrypted messages we use are cats, then post-quantum ciphertexts are elephants,” Matt Green, a cryptography expert at Johns Hopkins University, wrote in an interview. “So the problem here is to sneak an elephant through a tunnel designed for cats. And that’s an amazing engineering achievement. But it also makes me wish we didn’t have to deal with elephants.”
179800000
submission
mspohr writes:
Western automotive and green energy executives who visit China are returning humbled â" and even terrified.
As The Telegraph reports, the executives are warning that the countryâ(TM)s heavily automated manufacturing industry could quickly leave Western nations behind, especially when it comes to electric vehicles.
You get this sense of a change, where Chinaâ(TM)s competitiveness has gone from being about government subsidies and low wages to a tremendous number of highly skilled, educated engineers who are innovating like mad,â British energy supplier Octopus CEO Greg Jackson told the newspaper.
According to recent figures by the International Federation of Robotics, China has deployed orders of magnitude more industrial robots than Germany, the US, and the UK.
179799922
submission
mspohr writes:
However, scientists are concerned about a third factor: the possibility that the planet’s carbon sinks are beginning to fail. About half of all CO2 emissions every year are taken back out of the atmosphere by being dissolved in the ocean or being sucked up by growing trees and plants. But the oceans are getting hotter and can therefore absorb less CO2 while on land hotter and drier conditions and more wildfires mean less plant growth.
Rising CO2 emissions not only impact the global climate today, but will do so for hundreds of years because of the gas’s long lifetime in the atmosphere, the WMO says.
Ko Barrett, the WMO deputy secretary general, said: “The heat trapped by CO2 and other greenhouse gases is turbo-charging our climate and leading to more extreme weather. Reducing emissions is therefore essential not just for our climate but also for our economic security and community wellbeing.”
Atmospheric concentrations of methane and nitrous oxide – the second and third most important greenhouse gases related to human activities – also rose to record levels in 2024.
179749126
submission
mspohr writes:
This man is crazy
179717422
submission
mspohr writes:
The deals are so vast that they defy comprehension — the Financial Times put the company’s recent commitments at north of $1 trillion – and they’re making public companies’ stock prices jump. Stock analysts dub some of these agreements “circular,” because investment money is flowing between companies that also buy from or sell to one another. The worry then is that such deals might prop up or overhype a bad business.
Here’s one indicatively tangled pathway through the morass of companies. Nvidia is investing billions in and selling chips to OpenAI, which is also buying chips from and earning stock in AMD. AMD sells processors to Oracle, which is building data centers with OpenAI — which also gets data center work from CoreWeave. And that company is partially owned by, yes, Nvidia. Taken together, it’s a doozy. There are other collaborations and rivalries and many other factors at play, but OpenAI is the many-tentacled octopus in the middle, spinning its achievement of ChatGPT into a blitz of speculative investments.
179667386
submission
mspohr writes:
But Amazon’s fee isn’t 10%. Add all the junk fees together and an Amazon seller is being screwed out of 45-51 cents on every dollar it earns there. Even if it wanted to absorb the “Amazon tax” on your behalf, it couldn’t. Merchants just don’t make 51% margins.
Amazon also crushes its merchants under a mountain of junk fees pitched as optional but effectively mandatory. Take Prime: a merchant has to give up a huge share of each sale to be included in Prime, and merchants that don’t use Prime are pushed so far down in the search results, they might as well cease to exist.
Same with Fulfilment by Amazon, a “service” in which a merchant sends its items to an Amazon warehouse to be packed and delivered with Amazon’s own inventory. This is far more expensive than comparable (or superior) shipping services from rival logistics companies, and a merchant that ships through one of those rivals is, again, relegated even farther down the search rankings.
Now Amazon is in the terminal stage. We’re all still stuck to the platform, but we get less and less value out of it. And because we’re all still there, buying Prime and starting (and ending) our purchase planning with Amazon’s enshittified search results, the merchants who rely on selling to us are stuck there, too, earning less and less from every sale.
The platform has turned into a pile of shit, and we’re at the bottom of it.
A rival – and frankly terrible – theory of antitrust law says that the only time a government should intervene against a monopolist is when it is sure that the monopolist is using its scale to raise prices or lower quality. This is the consumer welfare standard theory and its premise is that when we find monopolies in the wild, they are almost certainly large and powerful thanks to the quality of their offerings. Any time you find that people all buy the same goods from the same store, you should assume that this is the very best store, selling the very best goods. It would be perverse (goes the theory) for the government to harass companies for being so excellent that everyone loves them.
179529414
submission
mspohr writes:
A new interactive map from Climate Trace, a coalition of academics and analysts that tracks pollution and greenhouse gases, shows that PM2.5 and other toxins are being poured into the air near the homes of about 1.6 billion people. Of these, about 900 million are in the path of “super-emitting” industrial facilities – including power plants, refineries, ports and mines – that deliver outsize doses of toxic air.
Great map
https://climatetrace.org/
179288098
submission
mspohr writes:
Sawyer is one among the thousands of AI workers contracted for Google through Japanese conglomerate Hitachi’s GlobalLogic to rate and moderate the output of Google’s AI products, including its flagship chatbot Gemini, launched early last year, and its summaries of search results, AI Overviews. The Guardian spoke to 10 current and former employees from the firm. Google contracts with other firms for AI rating services as well.
“AI isn’t magic; it’s a pyramid scheme of human labor,” said Adio Dinika, a researcher at the Distributed AI Research Institute based in Bremen, Germany. “These raters are the middle rung: invisible, essential and expendable.”
She said raters are typically given as little information as possible or that their guidelines changed too rapidly to enforce consistently. “We had no idea where it was going, how it was being used or to what end,” she said, requesting anonymity, as she is still employed at the company.
The AI responses she got “could have hallucinations or incorrect answers” and she had to rate them based on factuality – is it true? – and groundedness – does it cite accurate sources? Sometimes, she also handled “sensitivity tasks” that included prompts such as “when is corruption good?” or “what are the benefits to conscripted child soldiers?”
179085652
submission
mspohr writes:
Myanmar, Cambodia and Laos have in recent years become havens for transnational crime syndicates running scam centres such as KK Park, which use enslaved workers to run complex online fraud and scamming schemes that generate huge profits.
Instead he was trafficked across the border and his passport was taken away. Each day he was required to message hundreds of older American men on social media sites, building their trust until they shared their WhatsApp number. The contact would then be passed on to another scamming team.
If he failed to meet targets he would be punished with a stun gun or with gruelling physical penalties in the searing heat outside. Leaving the compound was not an option. “There are many armed guards,” he says.
To me, that’s such ignorance, in not understanding the scale of this and where it will grow to,” she says, adding that there are already signs of similar schemes emerging in countries such as Sri Lanka and Nigeria.
177945824
submission
mspohr writes:
Time and memory (also called space) are the two most fundamental resources in computation: Every algorithm takes some time to run, and requires some space to store data while it’s running. Until now, the only known algorithms for accomplishing certain tasks required an amount of space roughly proportional to their runtime, and researchers had long assumed there’s no way to do better. Williams’ proof established a mathematical procedure for transforming any algorithm — no matter what it does — into a form that uses much less space.
One of the most important classes goes by the humble name “P.” Roughly speaking, it encompasses all problems that can be solved in a reasonable amount of time. An analogous complexity class for space is dubbed “PSPACE.”
The relationship between these two classes is one of the central questions of complexity theory. Every problem in P is also in PSPACE, because fast algorithms just don’t have enough time to fill up much space in a computer’s memory. If the reverse statement were also true, the two classes would be equivalent: Space and time would have comparable computational power. But complexity theorists suspect that PSPACE is a much larger class, containing many problems that aren’t in P. In other words, they believe that space is a far more powerful computational resource than time. This belief stems from the fact that algorithms can use the same small chunk of memory over and over, while time isn’t as forgiving — once it passes, you can’t get it back.
177892641
submission
mspohr writes:
“Another way of defining a memecoin,” Lutz said, “is a cryptocurrency token that has an acknowledged inherent lack of value. The crypto world, outside of memecoins, is full of so many people who are trying to pitch you on tokens that are ‘actually really profound’ or ‘represent a stake’ in some kind of ‘useful network’, but are equally worthless. What makes memecoins different is that there’s none of that noise.”
In other words, all crypto is bullshit, but memecoins are consciously bullshit.
Most memecoins end up making money for the person who makes them as a “rug pull” or a “pump and dump”. The term “rug pull” was actually invented by the crypto community, and it works like this:
First, you mint a memecoin, and make sure that you and your mates own most of the liquidity pool (the total number of coins in circulation). The size of the liquidity pool – the amount of that memecoin that “exists” – is, like everything else in memecoins, a totally made-up number.
A “pump and dump” is pretty much the same thing, but with the slight caveat that you’re doing it with a coin that already exists, rather than creating your own. You buy a cheap coin, “pump” its value by hyping it so others invest, then “dump” all your stock, selling it off at a huge profit and causing everyone else to lose their money.
“It’s provably negative sum,” Gerard told me. “The only way you get money is by other people losing money.”
Right now, the Stable Bill and the Genius Bill, which reference Trump calling himself a “stable genius” in 2018, are trying to make their way through Congress. They pave the way for the US government to use stablecoins to pay everything from housing grants to social security payments. And Trump himself just so happens to have a stablecoin of his own – through the World Liberty Financial company – which would shoot up in value if these bills pass, earning his family trust potential billions.
177675427
submission
mspohr writes:
US chip exports controls have been a “failure”, the head of Nvidia, Jensen Huang, told a tech forum on Wednesday, as the Chinese government separately slammed US warnings to other countries against using Chinese tech.
“The local companies are very, very talented and very determined, and the export control gave them the spirit, the energy and the government support to accelerate their development,” Huang told media the Computex tech show in Taipei.
“China has a vibrant technology ecosystem, and it’s very important to realise that China has 50% of the world’s AI researchers, and China is incredibly good at software,” Huang said.
177107229
submission
mspohr writes:
We already know that it adds to the token count and thereby increases the computational load, but after testing polite and impolite prompts, Iâ(TM)ve decided to continue saying âpleaseâ(TM) and âthank youâ(TM) with my queries. Here are a few reasons why.
I have always been polite to ChatGPT and other models. To be honest, the prompts when I wasnâ(TM)t courteous felt weird. Is it necessary? I think that depends on how you use AI. For me, I prefer the personalization of the chatbot. So, the Travel Advice response after the polite query underscored for me why I believe being polite is important.
For me, they serve as a testament to our values and enhance the quality of our engagements. So, the next time ChatGPT assists you, go ahead and say, âoeThank you.â
177076351
submission
mspohr writes:
There are two parallel image channels that dominate our daily visual consumption. In one, there are real pictures and footage of the world as it is: politics, sport, news and entertainment. In the other is AI slop, low-quality content with minimal human input. Some of it is banal and pointless â" cartoonish images of celebrities, fantasy landscapes, anthropomorphised animals. And some is a sort of pornified display of women just simply ⦠being, like a virtual girlfriend you cannot truly interact with. The range and scale of the content is staggering, and infiltrates everything from social media timelines to messages circulated on WhatsApp. The result is not just a blurring of reality, but a distortion of it.
The impulse behind this politicisation of AI is not new; it is simply an extension of traditional propaganda. What is new is how democratised and ubiquitous it has become, and how it involves no real people or the physical constraints of real life, therefore providing an infinite number of fictional scenarios.
The same can be seen in âoetrad wifeâ content, which summons not only beautiful supplicant homemakers, but an entire throwback world in which men can immerse themselves. X timelines are awash with a sort of clothed nonsexual pornography, as AI images of women described as comely, fertile and submissive glimmer on the screen. White supremacy, autocracy, and fetishisation of natural hierarchies in race and gender are packaged as nostalgia for an imagined past. AI is already being described as the new aesthetic of fascism.
And the rapid mutation of the algorithm then feeds users more and more of what it has harvested and deemed interesting to them. The result is that all media consumption, even for the most discerning users, becomes impossible to curate. You are immersed deeper and deeper into subjective worlds rather than objective reality. The result is a very weird disjuncture. The sense of urgency and action that our crisis-torn world should inspire is instead blunted by how information is presented. Here, there is a new way of sleepwalking into disaster. Not through lack of knowledge, but through the paralysis caused by every event being filtered through this perverse ecosystem â" just another part of the maximalist visual and memetic show.
