Sam Altman seems to be making OpenAI way more non-profit than before: Even as the AI bubble becomes a mainstream talking point on Wall Street, tech companies continue to peddle the fantasy that AI is poised to become an almost magical money-maker. Case in point, OpenAI wants you to believe that by 2030, it’ll be raking in $100 billion a year just from ads alone — even though it’s currently struggling to reach just $1 billion. ↫ Joe Wilkins at Futurism The US tech giants fueling this “AI” bubble are trying to hide the true extent of their debt: Hidden debt at U.S. tech giants swelled eightfold in four years to an estimated $1.65 trillion as artificial intelligence investments ballooned, a Nikkei study shows, exceeding actual debt and making it tougher for investors to assess risk. The five companies’ hidden debt, which does not appear on balance sheets, totaled $1.65 trillion in the most recent quarter, exceeding the roughly $1.35 trillion in debt reflected on their balance sheets. The data includes some estimates. ↫ Kohei Yamada at Nikkei Asia The bubble is expanding to comical proportions: The American stock market is booming, thanks to artificial intelligence. Tech giants are borrowing billions to acquire AI talent, purchase chips and hardware, and construct data centers. And market watchers are starting to get worried. They see financiers bulldozing giant piles of money to private AI start-ups with no realistic path to profitability, tech companies reliant on other tech companies for revenue growth, and non-tech businesses without a lot to show for their AI investments. The value of AI-linked firms has climbed $27 trillion in the past three years—an astonishing amount, equivalent to 36 percent of the value of the entire U.S. stock market today. Although future earnings could justify those valuations, as Dominic Wilson and Vickie Chang of Goldman Sachs argued in a note to clients, the profit expectations require Panglossian optimism. No less an authority than Sam Altman is arguing that we are in an AI bubble. The International Monetary Fund is citing it as a significant risk to financial stability and warning about what might happen when it bursts: diminished investment, tighter credit, reduced consumption, disrupted trade flows. ↫ Annie Lowrey at The Atlantic I’m not worried, though. I have it on good authority that “AI” increases productivity by 10x, so surely, none of the above is a problem. Any day now, we will be inundated with waves of brand new, high-quality, valuable software. Any day now, existing software will increase in quality by 10x, leading to a huge surge in software sales. Any day now, productivity in factories will increase rapidly thanks to “AI” freeing up workers’ time, driving prices down 10x, leaving consumers with 10x more money to spend. Any day now, everyone will be able to produce the next Citizen Kane or write the next Anna Karenina, causing an explosion in magnificent, timeless art that will have historians of the future marvel at our civilisation’s ingenuity and artistry. In the meantime, these companies can just ask their “AI” how to become profitable. Should be table-stakes for a 10x force multiplier. I’m not worried.

One of the cool features of JPEG files is that there’s the option to save low frequency components first. This means that a partially downloaded image will be displayed at low resolution instead of being cut off. ↫ maurycyz.com Oh I know where this is going… Doing this, I can get Chrome to render around 90 frames before giving up. Other browsers like Firefox have more patience, but a 90 scan image seems to work almost everywhere. ↫ maurycyz.com Yes, you can abuse the mentioned feature to create a really odd type of video. Or animation? Well, it allows you to create something resembling a really low-resolution GIF. Useless, yes, but very novel.

The NLnet Foundation’s NGI0 Commons Fund supported an effort to add WPA3 support to OpenBSD, and the work’s payed off. All drivers which support PMF can use WPA3, which are: iwm, iwx, and qwx. So far, I have tested this patch on iwx AX200 only. I will roll out this patch to more of my devices now. Help with testing is welcome. There are both userland and kernel changes involved. ↫ Stefan Sperling Only the second implementation of WPA3 will be supported, which requires some explanation: WPA3 has a complicated history. There are two versions of WPA3. The initially standardized version suffered from side-channel leaks found by Mathy Vanhoef and dubbed “Dragonblood“. A revised and fixed version has been standardized and is mandatory in the 6 GHz band as of Wifi 6e (11ax) and mandatory on all bands as of Wifi 7 (11be). ↫ Stefan Sperling Obviously, WPA3 is a very welcome addition to OpenBSD.

Speaking of OpenVMS and Alpha – and we like speaking about OpenVMS and Alpha, don’t we? – there’s now a port of DOSBox that runs on the Alpha version venerable operating system. Astr0baby has published both binaries and source code for the port, as well as a lovely set of screenshots to show it off working.

It was only a few weeks ago that we got a massively improved Alpha emulator, capable of running VMS, Windows 2000, and Tru64, including X11 support and a variety of other exciting features. Today, we’ve got another major emulation milestone (update: sadly, with “AI” support, so odds are this will fizzle out. Bummer!). The emulation space is going crazy, after my previous post on Windows booting on DEC Alpha es40 emulator, there is now another huge breakthrough in the emulation of other non-x86 CPU emulation. Yufeng Gao with help from gdwnldsKSC (the man behind the updated es40-fork) has released version 0.1 of his Intel Itanium (IA-64) emulator that boots the Itanium version of Windows Server 2003 and Windows XP 64-bit. No OpenVMS or HP-UX yet and Linux/BSD also don’t boot. But Windows is amazing already. ↫ Remy van Elst Much like Alpha hardware, Itanium hardware is quite hard to come by – especially Itanium workstations are a nightmare to find; I think I’ve only ever seen one or two Itanium workstation come up for sale on eBay in recent years, and their rarity obviously commanded hefty prices. The sooner we are able to run Itanium version of operating systems comfortably in a virtualised environment the better. As long-time OSNews readers know, my heart beats for HP-UX, but the Itanium versions of Windows and VMS would be of more interest to most people, I’m sure. Excellent news.

A country selector looks simple because the interface reduces geography to one field. The infrastructure behind it is harder. In regulated digital products, location can affect access, content, payments, and account actions. That is why sweepstakes casino software by NuxGame should be viewed through a systems lens: geography is not a preference, but a changing signal that software must interpret. A Country Selector Knows Almost Nothing A country selection tells a platform what a user claims, not necessarily where the current session originates. IP data can suggest a region, browser geolocation can add precision when permission is granted, and account information can provide context. None of these inputs should automatically be treated as perfect. The design problem is therefore about confidence. A platform needs rules for what happens when signals agree, conflict, or disappear. That requires a source-of-truth policy rather than another front-end field. Product teams also need to decide which actions can continue under uncertainty and which should stop until location is resolved. Location Is A Confidence Decision A useful geolocation layer asks more than “which country is this IP in?” The system has to understand how trustworthy each signal is, how recently it was collected, whether the current action requires a fresh check, and what happens when evidence conflicts across a session. Useful inputs and controls can include: More checks can improve confidence, but they also create friction and more failure states. That trade-off matters for sweepstakes casino software by NuxGame because product, compliance, support, and engineering may all depend on the result. A stricter policy reduces ambiguity while creating more cases that need explanation or review. Sessions Move Even When Accounts Do Not An account can remain unchanged while the user moves between networks or countries. Mobile devices switch from Wi-Fi to cellular connections, and VPNs can alter the apparent origin of traffic. Treating location as a permanent profile attribute therefore creates a mismatch between account state and session reality. The legal and commercial question is separate from the engineering one. Research into which countries allow iGaming can help a business understand where different models may be permitted, subject to current local rules and professional advice. Software must then translate approved market decisions into executable controls: when to check, what to restrict, and how to handle uncertainty. Caching Can Make Regional Rules Dangerous Caching is useful for static assets, localization files, images, and content metadata. It becomes more sensitive when the cached value controls whether a user can perform an action. A location decision that was valid earlier may no longer describe the current network or device context. This creates a performance-versus-authority trade-off. Rechecking everything continuously would increase requests, latency, and operational cost. Trusting an old result indefinitely is simpler but weaker. A better design defines expiration rules according to risk: some content can tolerate stale regional data, while sensitive actions may require a newer location decision. The Back Office Needs To Explain The Decision Geolocation failures are difficult to support when the system stores only “allowed” or “blocked.” Operations teams need enough context to understand which signal was used, when it was collected, whether signals conflicted, and which rule produced the outcome. Otherwise, support can only repeat what the front end says. Auditability creates another balance. More detailed logs make investigations easier, but location data can be sensitive and should not be exposed widely. Access controls, retention policies, and clear event records therefore matter as much as the geolocation engine itself. The implementation should make decisions explainable without giving every employee access to raw location data. Geolocation Belongs In Product Architecture Geolocation should not be added at the end as a country dropdown plus an IP lookup. It touches session management, permissions, caching, logging, support workflows, regional configuration, and the interface shown when access changes. Once those dependencies are visible, the architectural problem becomes much clearer. For sweepstakes casino software by NuxGame, the practical lesson is simple: location should be treated as a decision with evidence, freshness, and consequences. A dropdown can collect a preference. A production system has to decide what is true enough to act on—and remain capable of explaining that decision when the answer changes. Resources:  NuxGame Legal iGaming Markets Guide – regional iGaming market considerations;  W3C Geolocation – browser geolocation standard;  IETF RFC 8805 – IP geolocation feed format;  MDN Geolocation API – browser geolocation implementation guidance.

As of writing, the Zilog Z80 processor was officially launched 50 years ago, in July of 1976, less than 4 years after the last human had walked on the moon, decades closer to WWII than to the present day, roughly at a half way point between the Kennedy assassination and the fall of the Berlin wall, closer to the Korean war than to 9/11 which is itself an event that happened a quarter of a century ago. (Sorry…) The processor was extremely successful, being used in many 8 bit microcomputers, including early personal computers, home & hobby computers, as well as many embedded, industrial applications. Together with the 8080 & 8085 that it is binary compatible with, it contributed to creating a de facto hardware standard for 8 bit micros, allowing a de facto software standard of CP/M, and Microsoft BASIC. ↫ David Oberhollenzer The only device I actively remember using with a (sort-of) Z80 in it was the Game Boy, but most likely I’ve used a ton more over the decades that I don’t remember or simply was never ware of. I did a little surface-level digging, and there we are: the TI-83, one of Texas Instruments’ stupidly popular and eternally overpriced graphing calculators, release in 1996. I was part of the first wave of high school children in The Netherlands for whom a TI-83 graphing calculator was mandatory. During my high school years I used that thing extensively, for far more than just math class – I programmed applications for and on it, and played so many games on it. A friend and I even bought a communication cable so we could play competitive 1v1 Bomberman in class. Good times, made possible by the Z80.

Rumours had been circulating for a while, but now it’s official: OnePlus is effectively retreating from the European and US markets. Today, our hearts are undoubtedly heavy and mixed with emotion. As part of the proactive global strategy adjustment, OnePlus has decided to conclude new product rollouts in Europe and North America. ↫ OnePlus statement Once OnePlus’ co-founder Carl Pei left the company (and founded Nothing), things have been feeling shaky for OnePlus, once the undisputed darling of the more technical part of the Android crowd. Their phones got more expensive, their minimalist, close-to-stock Android version got progressively worse, and they started lagging in updates, too. My OnePlus Watch 3, for instance, which was promised to get WearOS 6 at some point, but never got it – meanwhile, WearOS 7 has already been released. No, this news is not particularly surprising. Luckily, the company claims it will honour its warranty and update support obligations for existing products in Europe and the US, which is nice, but also something they’re legally obligated to do (at least in the EU). A snag here is that the only update path the company offers is to ColorOS, from its parent company Oppo, which many more traditional Android and OnePlus users certainly won’t be happy about. Something is better than nothing, I suppose, and I’ll reserve judgment until I see what ColorOS 17 will be like on my other OnePlus product, a OnePlus Pad 3. It’s just one more victim of western markets (illegally) consolidating on Apple and Samsung (while a few Pixels rummaging in the margins).

Out of all the bloody things Microsoft could release as open source, they chose the world’s weirdest IRC client they shipped in the late ’90s that nobody used or even remembered? What on earth is happening? Microsoft Comic Chat is a Microsoft-developed Internet Relay Chat (IRC) chat client released in 1996 that rendered conversations as automatically generated comic strips. Instead of plain text, users communicated through cartoon avatars with messages displayed in speech bubbles inside dynamically composed comic panels. The application used an expert system to determine character placement, gestures, facial expressions, balloon shape, and panel layout in real time. It shipped as part of Internet Explorer 3.0 and was later bundled with Windows 98 and MSN before being discontinued in the early 2000s. ↫ Comic Chat’s GitHub page Not only is the original source code now available on GitHub, there’s also a modern, updated version that can make use of larger displays and higher resolutions. There’s a deliciously ’90s website for it, too.

OpenBSD parts ways with an architecture: OpenBSD will no longer be developed for loongson. The reasons are exactly what you’d expect. The last compiler update unfortunately does not work on mips64el, with clang 22 built with clang 19 being apparently functional, but clang 22 rebuilt with the previous clang 22 hitting deterministic SIGSEGV on various files. I don’t have the time and energy to try and debug this (which is likely an endianness problem, as octeon appears to run happily with clang 22), especially when it takes 10 days for clang to rebuild itself on these machines; and switching back to gcc 4 won’t help much as modern software in ports will require a working C++>=11 compiler to build anyway. ↫ Miod Vallat If I got my facts right, this does not affect the newer LoongArch, which is an entirely different architecture that isn’t supported by OpenBSD at all. Similarly, the other MIPS-based architecture OpenBSD supports, Octeon, remains supported and thus isn’t affected either.

After I mentioned a Jurassic Park anecdote the other day, I watched the movie again. I must have seen it at least ten times now. This time, I researched every computer/software I spotted. ↫ Fabien Sanglard We are all aware of the infamous “This is a UNIX system, I know this!” meme, but many more computers make their appearance in Jurassic Park, and Fabien Sanglard documents all of them. Apparently, there’s even a Motorla Envoy running Magic Cap on Dennis Nedry’s desk, which I almost find more exciting than the SGI powerhouses he uses. What’s also quite interesting – but not surprising – is that all of the computers used in the movie were real. The value of all of this hardware combined, when adjusted for inflation, adds up to about $4 million. A lot of money, but don’t you worry your pretty little heart, as SGI and Apple all loaned this hardware to the studio. They didn’t have to pay anything for it.

As a former translator with two rock-solid university degrees in the subject, there was never a universe in which I would not talk about Twitter’s new autotranslation feature turning the tamest things into hardcore pornography. Elon Musk’s AI chatbot Grok has long garnered a reputation for experiencing horrifically racist meltdowns, enabling child abuse, and doxxing users’ home addresses. It should come as no surprise, then, that its supposed “translation” is a piece of work, too. In April, the almost-trillionaire’s social media platform X instated automatic AI translations for all of its users — and the results certainly speak for themselves. As writer and author Parker Molloy pointed out in a recent post on Bluesky, the Grok feature is “taking some interesting liberties” with people’s otherwise sincere posts. Screenshots show how Grok completely botched translations by coming up with shocking and decidedly NSFW AI hallucinations. ↫ Victor Tangermann at Futurism The sloppy translations this garbage software comes up with are honestly quite hilarious when taken in isolation. It’s adding translations that are straight-up hardcore pornography descriptions to entirely tame material that has absolutely nothing to do with pornography. The description of a video of some guy making coffee is translated into “man masturbates and jerks off to his own coffee during commercial flight”. We all know how this happened. There’s a lot of pornography on the internet, and Grok being the worst autocomplete among autocompletes, it was probably fed a lot of pornography, without any limitations or guardrails. The end result is obvious: some random videogame video is now a “cumshot video with my stepmom”. It would be absolutely hilarious if it wasn’t horribly dangerous. I’ve explained countless times that “AI”-based translations are going to get people killed – probably already have, but we just don’t realise it yet – and it’s not hard to see how a slopmachine turning innocuous things into hardcore pornography can do just that. There are countless places in the world where a woman unknowingly sending a pornographic message to her parents or whatever can get her hurt – or worse. I hadn’t even considered this particular way “AI” translations could get people hurt. Sadly, we will most likely never know the full extent to which “AI” translations will get people hurt and killed. When your grandmother takes her medicine in the wrong way because the “AI”-translated leaflet was unclear or downright wrong, and she ends up in the hospital because of it, will you ever find out what caused it?

Haiku has another buy month of development activity to detail, and there’s a big ticket item this time, even if the developers themselves don’t consider it so. The thing that should be the biggest news item this month is that the GSoC 2024 work to port “NVMM”, the NetBSD Virtual Machine Monitor (which runs on more than just NetBSD, despite the name), providing hardware-accelerated virtualization support for QEMU, was finally merged. Unfortunately it still doesn’t fully work, so it’s still disabled by default: hence, it’s only a minor news item, unfortunately. ↫ waddlesplash on Haiku’s website It may not work due to – so far – not well-understood problems causing any complex virtualised operating system to crash in a variety of ways, but since these problems seem related not to NVMM but Haiku itself, I still think this is a big piece of news. If the problems can be addressed, Haiku will have proper virtualisation, which is crazy to think about. There’s a forum thread in case you wish to help out with this effort. Other than this major news, there’s the usual list of small fixes and changes, including preliminary work on USB Ethernet support, which, when working, could be very welcome news for people whose onboard Ethernet doesn’t work with Haiku. The team also believes a beta 6 might actually be released this August, but once again I’d like to underline that Haiku’s nightlies work just fine, and you really don’t need to wait for a beta.

I have yet to see any of these creepy camera glasses Facebook (and a few other companies) are selling. One of the many benefits of living in Arctic Sweden, where people are reserved, keep their their distance, and try not to draw attention to themselves, is that new technology fads don’t really permeate society here. The odds of me spotting one of these creepy predator glasses in my remote town are incredibly slim, and to me, that’s a feature, not a bug. Meanwhile, in places where these creepy things can actually be found in the wild, a backlash is thankfully growing. Will Kujawa, a freelance video producer, said that he has been thinking about buying a pair of Meta glasses with prescription lenses to film behind the scenes content during his shoots, but the online backlash has given him second thoughts. He says he was “blown away by how mean some of the people were” in response to his social media posts about considering buying a pair. “I saw all these comments about if you wear those glasses you’re basically a predator or a creep, and I was like, ‘oh, maybe it’s not a good idea to have those,'” he told Engadget. But he says he understands why people have concerns. “I didn’t really think that through all the way … there are a lot of times where it’s not appropriate to wear cameras on your face. And even though I would have no intention of do anything creepy with them, it didn’t even occur to me other people just assume that automatically.” ↫ Karissa Bell at Engadget I can maybe see a use for these things in specific professional environments, but even then, obviously not ones made by Facebook, one of the, if not the creepiest companies in technology history. If I were to see anyone out here in the real world using one these things, I, too, would automatically assume that the guy (statistically speaking) wearing them is a creep. I can only imagine what the people most often targeted by creepy men would think encountering some rando wearing these. Clearly, these things should be made illegal outside of specific professional environments where they could potentially be useful. While it’s impossible to stop tools like these from making their way into the hands of creeps, it at least provides the justice system with a clear method of nailing them to the wall. They didn’t get Al Capone for any of his violent crimes – they nailed him for tax evasion.