ForgottenFlux

The Debian Publicity Team will no longer post on X/Twitter. We took this decision since we feel X doesn't reflect Debian shared values as stated in our social contract, code of conduct and diversity statement. X evolved into a place where people we care about don't feel safe. You are very much invited to follow us on https://bits.debian.org/ , on https://micronews.debian.org/ , or any media as listed on https://wiki.debian.org/Teams/Publicity/otherSN

So taking data without permission is bad, now?

I'm not here to say whether the R1 model is the product of distillation. What I can say is that it's a little rich for OpenAI to suddenly be so very publicly concerned about the sanctity of proprietary data.

The company is currently involved in several high-profile copyright infringement lawsuits, including one filed by The New York Times alleging that OpenAI and its partner Microsoft infringed its copyrights and that the companies provide the Times' content to ChatGPT users "without The Times’s permission or authorization." Other authors and artists have suits working their way through the legal system as well.

Collectively, the contributions from copyrighted sources are significant enough that OpenAI has said it would be "impossible" to build its large-language models without them. The implication being that copyrighted material had already been used to build these models long before these publisher deals were ever struck.

The filing argues, among other things, that AI model training isn't copyright infringement because it "is in service of a non-exploitive purpose: to extract information from the works and put that information to use, thereby 'expand[ing] [the works’] utility.'"

This kind of hypocrisy makes it difficult for me to muster much sympathy for an AI industry that has treated the swiping of other humans' work as a completely legal and necessary sacrifice, a victimless crime that provides benefits that are so significant and self-evident that it's wasn't even worth having a conversation about it beforehand.

A last bit of irony in the Andreessen Horowitz comment: There's some handwringing about the impact of a copyright infringement ruling on competition. Having to license copyrighted works at scale "would inure to the benefit of the largest tech companies—those with the deepest pockets and the greatest incentive to keep AI models closed off to competition."

"A multi-billion-dollar company might be able to afford to license copyrighted training data, but smaller, more agile startups will be shut out of the development race entirely," the comment continues. "The result will be far less competition, far less innovation, and very likely the loss of the United States’ position as the leader in global AI development."

Some of the industry's agita about DeepSeek is probably wrapped up in the last bit of that statement—that a Chinese company has apparently beaten an American company to the punch on something. Andreessen himself referred to DeepSeek's model as a "Sputnik moment" for the AI business, implying that US companies need to catch up or risk being left behind. But regardless of geography, it feels an awful lot like OpenAI wants to benefit from unlimited access to others' work while also restricting similar access to its own work.

Sony's game publishing arm has done a 180-degree turn on a controversial policy of requiring PC players to sign in with PlayStation accounts for some games, according to a blog post by the company.

Sony hasn't lost hope that players will still go ahead and use a PlayStation account, though, as it's tying several benefits to signing in.

The change is a major about-face for a handful of single-player titles after Sony faced considerable backlash from many angry PC players about the PlayStation account requirement to play the games.

If you were hoping for a respite from rising streaming subscription fees in 2025, you’re out of luck. Several streaming providers have already increased monthly and/or annual subscription rates, continuing a disappointing trend from the past few years, with no foreseeable end.

Subscribers have generally seen an uptick in how much money they spend to access streaming services. In June, Forbes reported that 44 percent of the 2,000 US streaming users it surveyed who “engage with content for at least an hour daily” said their streaming costs had increased over the prior year.

Deloitte's 2024 Digital Media Trends report found that 48 percent of the 3,517 US consumers it surveyed said that they would cancel their favorite streaming video-on-demand service if the price went up by $5.

Similarly, in a blog post about 2025 streaming trends, consumer research firm GWI reported that 52 percent of US TV viewers believe streaming subscriptions are getting too expensive, “which is a 77 percent increase since 2020.” A GWI spkesperon told me that the data comes from GWI's flagship dataset and surveying people from over 50 global markets. Its methodology is available here.) GWI added that globally, the top reason cited by customers who have canceled or are considering canceling a streaming service was cost (named by 39 percent of consumers), followed by price hikes (32 percent).

“Pay TV packages and inflation have increased at similar rates in recent years. But over the past two years, streaming has gotten much more expensive relative to both,” eMarketer’s report says.

The narrative that OpenAI, Microsoft, and freshly minted White House “AI czar” David Sacks are now pushing to explain why DeepSeek was able to create a large language model that outpaces OpenAI’s while spending orders of magnitude less money and using older chips is that DeepSeek used OpenAI’s data unfairly and without compensation. Sound familiar?

Both Bloomberg and the Financial Times are reporting that Microsoft and OpenAI have been probing whether DeepSeek improperly trained the R1 model that is taking the AI world by storm on the outputs of OpenAI models.

It is, as many have already pointed out, incredibly ironic that OpenAI, a company that has been obtaining large amounts of data from all of humankind largely in an “unauthorized manner,” and, in some cases, in violation of the terms of service of those from whom they have been taking from, is now complaining about the very practices by which it has built its company.

OpenAI is currently being sued by the New York Times for training on its articles, and its argument is that this is perfectly fine under copyright law fair use protections.

“Training AI models using publicly available internet materials is fair use, as supported by long-standing and widely accepted precedents. We view this principle as fair to creators, necessary for innovators, and critical for US competitiveness,” OpenAI wrote in a blog post. In its motion to dismiss in court, OpenAI wrote “it has long been clear that the non-consumptive use of copyrighted material (like large language model training) is protected by fair use.”

OpenAI argues that it is legal for the company to train on whatever it wants for whatever reason it wants, then it stands to reason that it doesn’t have much of a leg to stand on when competitors use common strategies used in the world of machine learning to make their own models.

Last week, Copilot made an unsolicited appearance in Microsoft 365. This week, Apple turned on Apple Intelligence by default in its upcoming operating system releases. And it isn't easy to get through any of Google's services without stumbling over Gemini.

Regulators worldwide are keen to ensure that marketing and similar services are opt-in. When dark patterns are used to steer users in one direction or another, lawmakers pay close attention.

But, for some reason, forcing AI on customers is acceptable. Rather than asking "we're going to shovel a load of AI services into your apps that you never asked for, but our investors really need you to use, is this OK?" the assumption instead is that users will be delighted to see their formerly pristine applications cluttered with AI features.

Customers have not asked for any of this. There has been no clamoring for search summaries, no pent-up demand for the revival of a jumped-up Clippy. There is no desire to wreak further havoc on the environment to get an almost-correct recipe for tomato soup. And yet here we are, ready or not.

Without a choice to opt in, the beatings will continue until AI adoption improves or users find that pesky opt-out option.

Signal has announced new functionality in its upcoming beta releases, allowing users to transfer messages and media when linking their primary Signal device to a new desktop or iPad. This feature offers the choice to carry over chats and the last 45 days of media, or to start fresh with only new messages.

The transfer process is end-to-end encrypted, ensuring privacy. It involves creating a compressed, encrypted archive of your Signal data, which is then sent to the new device via Signal's servers. Despite handling the transfer, the servers cannot access the message content due to the encryption.

With the introduction of a cross-platform archive format, Signal is also exploring additional tools for message transfer to new devices or restoration in case of device loss or damage. Users can begin testing this feature soon, with a wider rollout expected in the coming weeks.

Signal has announced new functionality in its upcoming beta releases, allowing users to transfer messages and media when linking their primary Signal device to a new desktop or iPad. This feature offers the choice to carry over chats and the last 45 days of media, or to start fresh with only new messages.

The transfer process is end-to-end encrypted, ensuring privacy. It involves creating a compressed, encrypted archive of your Signal data, which is then sent to the new device via Signal's servers. Despite handling the transfer, the servers cannot access the message content due to the encryption.

With the introduction of a cross-platform archive format, Signal is also exploring additional tools for message transfer to new devices or restoration in case of device loss or damage. Users can begin testing this feature soon, with a wider rollout expected in the coming weeks.

Bitwarden users who store their email account credentials within their Bitwarden vaults would have trouble accessing the sent codes if they are unable to log in to their email.

To prevent getting locked out of your vault, be sure you can access the email associated with your Bitwarden account so you can access the emailed codes, or turn on any form of two-step login to not be subject to this process altogether.

Bitwarden users who store their email account credentials within their Bitwarden vaults would have trouble accessing the sent codes if they are unable to log in to their email.

To prevent getting locked out of your vault, be sure you can access the email associated with your Bitwarden account so you can access the emailed codes, or turn on any form of two-step login to not be subject to this process altogether.

Despite Microsoft's push to get customers onto Windows 11, growth in the market share of the software giant's latest operating system has stalled, while Windows 10 has made modest gains, according to fresh figures from Statcounter.

This is not the news Microsoft wanted to hear. After half a year of growth, the line for Windows 11 global desktop market share has taken a slight downturn, according to the website usage monitor, going from 35.6 percent in October to 34.9 percent in November. Windows 10, on the other hand, managed to grow its share of that market by just under a percentage point to 61.8 percent.

The dip in usage comes just as Microsoft has been forcing full-screen ads onto the machines of customers running Windows 10 to encourage them to upgrade. The stats also revealed a small drop in the market share of its Edge browser, despite relentlessly plugging the application in the operating system.

Last month, Ente launched https://theyseeyourphotos.com/, a website and marketing stunt designed to turn Google’s technology against itself. People can upload any photo to the website, which is then sent to a Google Cloud computer vision program that writes a startlingly thorough three-paragraph description of it. (Ente prompts the AI model to document small details in the uploaded images.)

If you don’t want to upload your own picture, Ente gives people the option to experiment on Theyseeyourphotos using one of several stock images. Google’s computer vision is able to pick up on subtle details in them, like a person’s tattoo that appears to be of the letter G, or a child’s temporary tattoo of a leaf. “The whole point is that it is just a single photo,” Mohandas says. He hopes the website prompts people to imagine how much Google—or any AI company—can learn about them from analyzing thousands of their photos in the cloud in the same way.