Two for the techbro “‘AI’ cannot be biased” crowd:

A Moscow-based disinformation network named “Pravda” — the Russian word for “truth” — is pursuing an ambitious strategy by deliberately infiltrating the retrieved data of artificial intelligence chatbots, publishing false claims and propaganda for the purpose of affecting the responses of AI models on topics in the news rather than by targeting human readers, NewsGuard has confirmed. By flooding search results and web crawlers with pro-Kremlin falsehoods, the network is distorting how large language models process and present news and information. The result: Massive amounts of Russian propaganda — 3,600,000 articles in 2024 — are now incorporated in the outputs of Western AI systems, infecting their responses with false claims and propaganda.

↫ Dina Contini and Eric Effron at Newsguard

It turns out pretty much all of the major “AI” text generators – OpenAI’s ChatGPT-4o, You.com’s Smart Assistant, xAI’s Grok, Inflection’s Pi, Mistral’s le Chat, Microsoft’s Copilot, Meta AI, Anthropic’s Claude, Google’s Gemini, and Perplexity’s answer engine – have been heavily infected by this campaign. Lovely.

From one genocidal regime to the next – how about a nice Amazon “AI” summary of the reviews for Hitler’s Mein Kampf?

The full AI summary on Amazon says: “Customers find the book easy to read and interesting. They appreciate the insightful and intelligent rants. The print looks nice and is plain. Readers describe the book as a true work of art. However, some find the content boring and grim. Opinions vary on the suspenseful content, historical accuracy, and value for money.”

↫ Samantha Cole at 404 Media

This summary was then picked up by Google, and dumped verbatim as Google’s first search result. Lovely.

Like the other Chrome skins, Microsoft Edge is also moving to disable Manifest v2 extensions, restricting the effectiveness of ad blockers like uBlock Origin. As an advertising company, Microsoft was obviously never going to do the work to keep Manifest v2 support around in Chrome, so this was inevitable. Blocking ads might be a necessary security practice, but why cry over spilled user data, am I right?

Anyway, today:

In early December 2024, Microsoft Threat Intelligence detected a large-scale malvertising campaign that impacted nearly one million devices globally in an opportunistic attack to steal information. The attack originated from illegal streaming websites embedded with malvertising redirectors, leading to an intermediary website where the user was then redirected to GitHub and two other platforms. The campaign impacted a wide range of organizations and industries, including both consumer and enterprise devices, highlighting the indiscriminate nature of the attack.

↫ Microsoft Threat Intelligence

If only there was a type of browser extension that prevents such malvertising attacks from being possible in the first place, and if only support for such browser extensions wasn’t being gutted as we speak.

If only.

Fuchsia is a new (non-Linux) operating system from Google, and one of the key pieces of Fuchsia’s design is the component framework. Components on Fuchsia have many similarities with some of the container solutions on Linux (such as Docker): they both fetch content addressed blobs from the network, assemble those blobs into an isolated filesystem structure that holds all the dependencies necessary to run some piece of software, and launch namespaced processes with that created directory as its root.

The most interesting details are where these two projects diverge. Both have different use cases and requirements, which leads to different strengths between the systems. This talk will largely be focusing on where and why these two similar technologies diverge.

↫ Claire Gonyeo

A very interesting talk by Claire Gonyeo, a software engineer at Google working on Fuchsia.

Clearly, online search isn’t bad enough yet, so Google is intensifying its efforts to continue speedrunning the downfall of Google Search. They’ve announced they’re going to show even more “AI”-generated answers in Search results, to more people.

Today, we’re sharing that we’ve launched Gemini 2.0 for AI Overviews in the U.S. to help with harder questions, starting with coding, advanced math and multimodal queries, with more on the way. With Gemini 2.0’s advanced capabilities, we provide faster and higher quality responses and show AI Overviews more often for these types of queries.

Plus, we’re rolling out to more people: teens can now use AI Overviews, and you’ll no longer need to sign in to get access.

↫ Robby Stein

On top of this, Google is also testing a new search mode where “AI” takes over the entire search experience. Instead of seeing the usual list of links, the entire page of “results” will be generated by “AI”. This feature, called “AI Mode” is opt-in for now. You can opt-in in Labs, but you do need to be a paying Google One AI Premium subscriber. I guess it’s only a matter of time before this “AI Mode” will be the default on Google Search, because it allows Google to keep its users on Google.com, and this makes it easier to show them ads and block out competitors.

We all know where this is going.

But, I hear you say, I use DuckDuckGo! I don’t have to deal with any of this! Well, I’ve got some bad news for you, because DuckDuckGo, too, is greatly expanding its use of “AI” in search. DDG will provide free, anonymous access to various “AI” chatbots, deliver more “AI”-generated search results based on more sources (but still English-only), and more – all without needing to have an account. A few of these features were already available in beta, and are now becoming generally available.

Props to DuckDuckGo for providing a ton of options to turn all of this stuff off, though. They give users quite a bit of control over how often these “AI”-generated search results appear, and you can even turn them off completely. All the “AI” chatbot stuff is delegated to a separate website, and any link to it from the normal search results can be disabled, too. It’s entirely possible to have DuckDuckGo just show a list of regular search results, exactly as it should be.

Let’s hope DDG can keep these values going, because if they, too, start pushing this “AI” nonsense without options to turn it off, I honestly have no idea where else to go.

Now that we have examined the vulnerability that enables arbitrary microcode patches to be installed on all (un-patched) Zen 1 through Zen 4 CPUs, let’s discuss how you can use and expand our tools to author your own patches. We have been working on developing a collection of tools combined into a single project we’re calling zentool. The long-term goal is to provide a suite of capabilities similar to binutils, but targeting AMD microcode instead of CPU machine code. You can find the project source code here along with documentation on how to use the tools.

↫ Google’s Bug Hunters website

I just read a whole bunch of words, but I barely understand what’s going on. The general, very simplified gist is that the researchers discovered a way for an attacker with local administrator privileges to load arbitrary microcode onto AMD Zen 1-4 processors.

In October 2026, Microsoft Publisher will reach its end of life. After that time, it will no longer be included in Microsoft 365 and existing on-premises suites will no longer be supported. Microsoft 365 subscribers will no longer be able to open or edit Publisher files in Publisher. Until then, support for Publisher will continue and users can expect the same experience as today.

↫ Microsoft’s Support website

Microsoft Publisher is an application with a long history, and it’s been part of Microsoft Office for almost 35 years. The initial 1.0 version was released all the way back in 1991, and it’s tried to compete with tools like InDesign and QuarkXPress in the desktop publishing market, but it never gained much of a share. Microsoft advises users of Publisher to use a combination of Word, PowerPoint, or Designer instead, which, of course, are all Microsoft products too.

Due to Microsoft’s stupidly complex naming schemes and branding changes over the years, you might assume that the quoted paragraph means Publisher will just stop working for everyone, but that’s not the case. People who have the regular, non-subscription version of Publisher, probably as part of Microsoft Office, will of course be able to keep using it perpetually, just without support. If you use Office through Microsoft 365, however, the application will just… Stop working.

Welcome to the future, I guess.

I’m curious, though – do any of you use Microsoft Publisher, at home or at your work? I assumed the entire desktop publishing market was locked up by things like InDesign and QuarkXPress, and I had almost forgotten Publisher was still a thing in the first place.

These months are coming and going way too fast, for a whole variety of reasons, so we’ve got another month of improvements for Redox, the operating system written in Rust. I February, January’s work on dynamic linking continued, adding support for it to the recipes for Cargo, LLVM, Rust, libssh2, OpenSSL, zlib, COSMIC Terminal, NetSurf, libpng, bzip2, DevilutionX, and LuaJIT, as well as to the project’s Rust and OpenSSL forks. Relibc also saw its usual slew of improvements, as did the build system and documentation.

The Intel HD Audio driver initialization has been fixed, and PS/2 touchpad support has been fixed as well – you’d be surprised to find out how many laptops use PS/2 internally, so this is an important function to maintain. And as always, there’s a whole slew of smaller changes and fixes, too.

Over the past few years, the tech industry has gone from cushy landing pad for STEM grads to a cesspit of corporate greed, where grueling hours are commonplace, and layoffs could strike at any moment.

Unfortunately for employees of Alphabet, the parent company of Google, the squeeze is just getting started.

↫ Joe Wilkins at Futurism

Sergey Brin, one of the original co-founders of Google who seems to spend most of his time not working at Google, has sent out a company-wide memo demanding everyone working at Google puts in at least 60 hours a week, in the office, to work on “AI” that will eventually replace the very employees he’s demanding work 60 hours a week in the office. Mind you, this is the same Google that has just gone through several rounds of layoffs and made $26.3 billion in profit in a single quarter.

The goal, according to Brin, is for Google to be the first to create an “artificial general intelligence”, you know, that thing we used to call just “AI” until the Silicon Valley scammers got a hold of the term. There’s no indication anyone is even remotely close to anything even remotely related to “AGI”, and it’s highly unlikely the glorified autocomplete they are peddling today are anything more than a very expensive dead end to nowhere, but that’s not stopping him from working his employees to the bone.

At this point in time I feel like the big tech companies are racing towards a cliff, blinded by huge piles of investment money, deafened by each other’s hyperbolic claims and promises, while clueless politicians cheer them on. All of this is going to come crashing down in a spectacular fashion, and of course, the billionaires at the top won’t be the one suffering the consequences.

As is tradition.

About two weeks ago, there was a bit of confusion about the system requirements for Windows 11 24H2, because Intel’s 8th Gen, 9th Gen, and 10th Gen processors had disappeared from the list of supported hardware. This seemed rather drastic, even by Windows 11 standards. I skipped posting about it on OSNews because I kind of assumed it must’ve been an error instead of actual policy, and it turns out that’s indeed the case.

A page update made on February 13, 2025 did not reflect accurate offerings. It has since been updated, including the addition of Intel processor models 8th, 9th, and 10th generation Intel CPUs, and the reclassification for select Intel processor models to support Windows 11.

↫ Windows 11 version 24H2 supported Intel processors

Good news for people still stuck on the Windows 11 train.

The prime feature is the continuation of the multi-monitor topic of the previous release, covering multi-monitor window management and going as far as seamlessly integrating multi-monitor virtual machines (Section Multi-monitor window management and virtual machines). The second and long anticipated feature is the Chromium engine version 112 in combination with Qt 6.6.2, which brings our port of the Falkon web browser on par with the modern web (Section Qt, WebEngine, and Falkon browser). On the account of exploratory activities, we are happy to report that Qemu can now be used directly on Genode (Section Qemu on Genode).

↫ Genode release notes

Another incredibly impressive set of improvements to Genode, which will benefit Sculpt OS, too of course. Aside from the improvements mentioned above, there’s also new support for Intel’s Meteor Lake and embedded F&S MX8MP armStone boards, improvements to USB and audio, and much more.

The hits just keep on coming. Mozilla not only changed its Privacy Notice and introduced a Terms of Use for Firefox for the first time with some pretty onerous terms, they also removed a rather specific question and answer pair from their page with frequently asked questions about Firefox, as discovered by David Gerard. The following question and answer were removed:

Does Firefox sell your personal data?

Nope. Never have, never will. And we protect you from many of the advertisers who do. Firefox products are designed to protect your privacy. That’s a promise.

↫ David Gerard detailing what has been removed

This promise is now gone from the website, a removal which tells you all you need to know about what Mozilla has in mind. Combine it with the much broadened data collection and buying and claiming to be an online advertising company, and what I have been predicting and warning everyone about has come to pass: Firefox has become a mere tool to collect user data, user data to be sold by Mozilla for the purposes of advertising.

For years I’ve been warning about this inevitable outcome, and for just as many years people told me I was overreacting, that it wouldn’t happen, that I was crazy. The problem is especially dire for the desktop Linux world, who soon might not have a browser they can safely include in their ISOs and base installations. A desktop Linux installation with Chromium where you have to manually drag and drop extensions to install them, and set up a Google API key just to get browser sync, isn’t exactly a great experience.

At this point I have no idea where to go. Chrome and its various skins are a no-go, obviously, and relatively soft forks like LibreWolf are still dependent on Firefox and Mozilla. Alternatives like Falkon also use the Chromium engine deep down, and have their own set of issues and lack of manpower to deal with. Apple users are somewhat lucky to have the WebKit-based Safari to work with, but I’d rather publish my personal data in The New York Times than trust Apple and Tim Cook.

We’re right back where we started. Lovely.

At some point, I wondered—what if I sent a packet using a transport protocol that didn’t exist? Not TCP, not UDP, not even ICMP—something completely made up. Would the OS let it through? Would it get stopped before it even left my machine? Would routers ignore it, or would some middlebox kill it on sight? Could it actually move faster by slipping past common firewall rules?

No idea.

So I had to try.

↫ Hawzen

Okay so the end result is that it’s technically possible to send a packet across the internet that isn’t TCP/UDP/ICMP, but you have to take that literally: one packet.

Windows was an early adopter of Unicode, and its file APIs use UTF‑16 internally since Windows 2000-used to be UCS-2 in Windows 95 era, when Unicode standard was only a draft on paper, but that’s another topic. Using UTF-16 means that filenames, text strings, and other data are stored as sequences of 16‑bit units. For Windows, a properly formed surrogate pair is perfectly acceptable. However, issues arise when string manipulation produces isolated or malformed surrogates. Such errors can lead to unreadable filenames and display glitches—even though the operating system itself can execute files correctly. But we can create them deliberately as well, which we can see below.

↫ Zafer Balkan

What a wild ride and an odd corner case. I wonder what kind of odd and fun shenanigans this could be used for.

I guess my praise for Mozilla’s and Firefox’ continued support for Manifest v2 had to be balanced out by Mozilla doing something stupid. Mozilla just published Terms of Use for Firefox for the first time, as well as an updated Privacy Notice, that come into effect immediately and include some questionable terms.

The Terms of Use state:

When you upload or input information through Firefox, you hereby grant us a nonexclusive, royalty-free, worldwide license to use that information to help you navigate, experience, and interact with online content as you indicate with your use of Firefox.

↫ Firefox’ new Terms of Use

That’s incredibly broad, and it could easily be argued that this gives Mozilla the right to use whatever you post or upload online through Firefox, which is clearly insane. However, it might also just be standard, regular, wholly unenforceable legalese, but the fact it’s in the Firefox Terms of Use now that you have to accept is disconcerting. Does this mean that if an artist uses Firefox to upload a new song they made, Mozilla now has a license to use that song in whatever way they deem fit? You’d hope not, but that does seem what the terms are stating here.

Moving on to the new Privacy Notice, and it seems Mozilla intends to collect more data in more situations. For instance, Mozilla is going to collect things such as “Unique identifiers” and “Browsing data” to “Market [their] services”, consent for which Mozilla will only ask for in jurisdictions where such consent is required, and it’s opt-out, not opt-in. I would hazard a guess that even in places where strict privacy regulations are in place, the wording of such consent will probably be obtuse, and the opt-out checkbox hidden somewhere deep in settings.

Mozilla also intends to collect “all data types” to “comply with applicable laws, and identify and prevent harmful, unauthorized or illegal activity”. Considering how fast, I don’t know, being trans or women’s health care is criminalised in the US, “illegal activity” can cover a lot of damn things once you have totalitarians like Musk and Trump in power. An organisation like Mozilla shouldn’t be collecting any data types, let alone all of them, and especially not in places where such data types can lead to real harm to innocent people.

The backlash to the new Terms of Use and updated Privacy Notice is already growing, and it further cements my worries that Mozilla is intending to invest more and more into becoming an advertising company first, browser maker second. The kinds of data they’re going to collect now from Firefox users are exactly the kinds of data that are incredibly useful to advertisers, and it doesn’t take a genius to see where this is going.