Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror

Comment Re:The enshittification begins (Score 2) 39

It may not be attracting to you but it is certainly popular. ChatGPT has about 800 million active users as of April https://www.demandsage.com/chatgpt-statistics/. Now, that's using data in part from OpenAI, but other metrics which are not from OpenAI paint a pretty similar picture. ChatGPT's website is one of the world's 10 most visited websites according to Similarweb and has been consistently that way for over a year now https://www.similarweb.com/top-websites/ . Whatever problems ChatGPT has, lack popularity is not one of them.

Submission + - JWST spots a strange red dot so extreme scientists can't explain it (sciencedaily.com) 1

alternative_right writes: The discovery of strange, ultra-red objects—especially the extreme case known as The Cliff—has pushed astronomers to propose an entirely new type of cosmic structure: black hole stars. These exotic hybrids could explain rapid black hole growth in the early universe, but their existence remains unproven.

Submission + - Scientists just teleported information using light (sciencedaily.com)

alternative_right writes: Quantum communication follows a similar idea, but individual photons act as the information carriers. A zero or one is encoded through the direction of the photon's polarization (i.e., their orientation in the horizontal and vertical directions or in a superposition of both states). Because photons behave according to quantum mechanics, their polarization cannot be measured without leaving detectable traces. Any attempt to intercept the message would be exposed.

Teleportation requires the photons to be nearly identical in properties such as timing and color. Producing such photons is hard because they come from separate sources.

At the University of Stuttgart, the researchers successfully teleported the polarization state of a photon from one quantum dot to a photon produced by a second quantum dot. One dot emits a single photon and the other generates an entangled photon pair. "Entangled" means the two photons share a single quantum state even when physically apart. One photon from the pair travels to the second quantum dot and interacts with its photon. When the two overlap, their superposition transfers the information from the original photon to the far-away partner of the entangled pair.

A key element of this achievement was the use of "quantum frequency converters," devices that adjust small frequency mismatches between photons.

Submission + - The internet works thanks to a shared infrastructure that nobody owns (elpais.com) 1

alternative_right writes: In the 21st century, every government should understand that ensuring software sovereignty and security is part of its job, not only for themselves but also for businesses, society, and researchers. In the 21st century, software is the invisible infrastructure of our everyday life, like roads and bridges. Everything runs on software, and a significant portion of this is made possible by open source, which is maintained by people selflessly. If this open source breaks down, it’s as if a road or bridge collapses: everything else becomes much more complicated and dangerous.

Submission + - The internet works thanks to a shared infrastructure that nobody owns (elpais.com) 1

alternative_right writes: In the 21st century, every government should understand that ensuring software sovereignty and security is part of its job, not only for themselves but also for businesses, society, and researchers. In the 21st century, software is the invisible infrastructure of our everyday life, like roads and bridges. Everything runs on software, and a significant portion of this is made possible by open source, which is maintained by people selflessly. If this open source breaks down, it’s as if a road or bridge collapses: everything else becomes much more complicated and dangerous.

Submission + - AI now writes 25% of Google's code, says CEO Sundar Pichai (arstechnica.com) 1

yuvcifjt writes: On Tuesday during Google's Q3 2024 earnings call, CEO Sundar Pichai revealed that AI systems now generate more than a quarter of new code for its products, with human programmers overseeing the computer-generated contributions, showing how AI tools are already having a sizeable impact on software development.

"We're also using AI internally to improve our coding processes, which is boosting productivity and efficiency," Pichai said during the call. "Today, more than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers. This helps our engineers do more and move faster."

ArsTechnica goes on to state:

Google developers aren't the only programmers using AI to assist with coding tasks. It's difficult to get hard numbers, but according to Stack Overflow's 2024 Developer Survey, over 76 percent of all respondents "are using or are planning to use AI tools in their development process this year," with 62 percent actively using them. A 2023 GitHub survey found that 92 percent of US-based software developers are "already using AI coding tools both in and outside of work."

"Whether you think coding with AI works today or not doesn't really matter," posted former Microsoft VP Steven Sinofsky in September. Sinofsky has a personal history of coding going back to the 1970s. "But if you think functional AI helping to code will make humans dumber or isn't real programming just consider that's been the argument against every generation of programming tools going back to Fortran."

Submission + - Oracle's credit status under pressure (latimes.com)

Bruce66423 writes: 'A gauge of risk on Oracle Corp.’s debt reached a three-year high in November, and things are only going to get worse in 2026 unless the database giant is able to assuage investor anxiety about a massive artificial intelligence spending spree, according to Morgan Stanley.'

First sign of the boom's inevitable collapsing?

Submission + - Scientists may have found dark matter after 100 years of searching (sciencedaily.com)

alternative_right writes: A University of Tokyo researcher analyzing new data from NASA’s Fermi Gamma-ray Space Telescope has detected a halo of high-energy gamma rays that closely matches what theories predict should be released when dark matter particles collide and annihilate. The energy levels, intensity patterns, and shape of this glow align strikingly well with long-standing models of weakly interacting massive particles, making it one of the most compelling leads yet in the hunt for the universe’s invisible mass.

Submission + - Trump says Venezuelan airspace should be considered closed (reuters.com)

sinij writes:

U.S. President Donald Trump said on Saturday the airspace above and surrounding Venezuela should be considered "closed in its entirety," but gave no further details, stirring anxiety and confusion in Caracas as his administration ramps up pressure on President Nicolas Maduro's government.

The war with Venezuela is not in US interest and goes against explicit election promise not to start more pointless wars.

Submission + - Torrent Giant YTS Returns to .LT Domain After .MX "Vanishes" (torrentfreak.com)

cristiroma writes: After five years of operating in relatively stable waters, popular torrent site YTS is on the move again. The site’s Mexican domain name, YTS.mx, suddenly stopped resolving yesterday and has effectively vanished. While the official reason remains unclear, the YTS operators decided to return to one of their previous homes, YTS.lt.

With millions of regular users, YTS is arguably the most visited torrent site on the internet today.

The current operators ‘unofficially’ took over the YTS brand in 2015 after the original group threw in the towel. Since then, it has amassed a rather impressive user base.

After adopting one of the most iconic piracy brands, YTS faced its fair share of legal troubles. In 2019, the popular torrent site and its operator were accused of mass copyright infringement in multiple lawsuits filed by filmmakers in the United States. Surprisingly, YTS managed to settle these lawsuits to live another day, although that came at a price.

YTS also dealt with various domain name challenges. When the site first entered the scene, it was operating from the YTS.ag domain name, which it traded in for YTS.am a few years later. In 2019, the torrent site moved to YTS.lt, which it swapped for the YTS.mx domain in 2020.

Legal Pressure?
While it is apparent that the .MX domain name issues are serious, it is not immediately clear what caused them. TorrentFreak asked Registry .MX for a comment on the situation, but the organization did not immediately respond.

Submission + - Scientists make sense of shapes in the minds of the models (foommagazine.org)

Gazelle Bay writes: It was at least since 2021, according to the authors of a preprint from March, that researchers began to see something interesting on the insides of their models.

Also known as an AI program, created from a neural network architecture, a model processes a word by learning to represent it as an arrow or a vector within a high-dimensional space. The directions of these words—which each end up at one single point—become the model's carriers of information.

While these spaces are already strange in their vastness, often consisting in thousands of dimensions, researchers were noticing something even more peculiar; sometimes, inputs would form clouds of points that were distinctively shaped, looking for example like 'Swiss rolls,' or cylinders, after being projected back down to just three dimensions, using standard methods. Over the next few years, they started to see other cloudy shapes, too: curves, loops, circles; helixes, torii; even trees and fractal geometries.

That models might learn to organize information in shapes did not necessarily surprise people. It was natural to think that a model might learn that certain categories of inputs could all be clumped together; like inputs describing calendar dates, or colors, or arithmetical operations.

But in 2023, when others discovered a new method for understanding the insides of their models, called sparse autoencoders (SAEs), the observations began to seem a little odder. This method, which quickly gained traction, was suggestive of a very different picture—that the most important concepts a model learned, like love, or logic, or the identities of different people, were highly fragmented, each one tearing off in a very different direction. But why then were certain inputs found close together?

Almost as soon as this hint of contradiction surfaced, it was quelled by other findings. Both the study from March as well as an October study, by researchers at the company Anthropic, have shown that models learn shapes in ways that compliment these other tendencies, suggested by other methods. As a consequence, we are increasingly making sense of why models learn to make shapes in the high-dimensional minds that they live in.

"There's a lot of confusion, but it also feels like there's been a lot of progress," said Eric Michaud of the Massachusetts Institute of Technology (MIT), who spoke to Foom in an interview. "I don't know where it's all going to go. But overall, it feels healthy."

Submission + - Conde Nast fined €750,000 for placing cookies without consent (noyb.eu)

AmiMoJo writes: In December 2019, noyb had filed complaints against three providers of French websites, because they had implemented cookie banners that turned a clear “NO” into “fake consent”. Even if a user went through the trouble of rejecting countless cookies on the eCommerce page CDiscount, the movie guide Allocine.fr and the fashion magazine Vanity Fair, these websites sent digital signals to tracking companies claiming that users had agreed to being tracked online. CDiscount sent “fake consent” signals to 431 tracking companies per user, Allocine to 565, and Vanity Fair to 375, an analysis of the data flows had shown.

Today, almost six (!) years after these complaints had originally been filed, the French data protection authority CNIL has finally reached a decision in the case against Vanity Fair: Conde Nast, the publisher behind Vanity Fair, has failed to obtain user consent before placing cookies. In addition, the company failed to sufficiently inform its users about the purpose of supposedly “necessary” cookies. Thirdly, the implemented mechanisms for refusing and withdrawing consent was ineffective. Conde Nast must therefore pay a fine of €750.000.

Conde Nast also owns Ars Technica.

Slashdot Top Deals

"I think Michael is like litmus paper - he's always trying to learn." -- Elizabeth Taylor, absurd non-sequitir about Michael Jackson

Working...