Enlarge (credit: Getty Images | Chris Delmas)

Elon Musk's decision to rebrand Twitter as "X" wouldn't be complete without a change to the company's official Twitter account. The @X handle was already taken by a user who registered it over 16 years ago, but that wasn't much of an obstacle—Twitter simply took over the username and offered its longtime owner some merchandise but no monetary compensation.

San Francisco-based photographer Gene X Hwang was @X on Twitter from March 2007 until yesterday. "They just took it essentially—kinda what I thought might happen," Hwang told The Telegraph . "They did send an email saying it is the property of 'x' essentially."

Hwang confirmed to Ars today that "there was no financial compensation" offered to him. The company offered "to switch the @x account and its history/followers etc to a new handle once I select one that is available," Hwang told us. "They also offered some merch and to meet with the management team as well."

Enlarge (credit: OpenAI)

On Tuesday, OpenAI released an official ChatGPT app for Android, now available in the Google Play Store in four countries: the US, India, Bangladesh, and Brazil, with more coming soon. As a client for OpenAI's language model family, the GPT-3.5 and GPT-4 models run on the cloud and provide results to your Android device. It also integrates OpenAI's Whisper model for speech recognition.

ChatGPT, launched in November, is a conversational AI language model interface. As an AI assistant, it can help with summarization, text composition, and analysis. OpenAI bills its use cases as a way to seek "instant answers," "tailored advice," "creative inspiration," "professional input," and "learning opportunities."

A series of images showing OpenAI's proposed uses of the ChatGPT Android app, taken from the Google Play Store. (credit: OpenAI)

However, as we've noted in the past , ChatGPT is occasionally prone to confabulation (that is, making things up)—especially the GPT-3.5 model—so it's not entirely trustworthy as a factual reference. It can come in handy as a way to analyze data you provide yourself, though, so long as you're familiar with the subject matter and can validate the results.

Enlarge / The four companies say they launched the Frontier Model Forum to ensure "the safe and responsible development of frontier AI models." (credit: Financial Times)

Four of the world’s most advanced artificial intelligence companies have formed a group to research increasingly powerful AI and establish best practices for controlling it, as public anxiety and regulatory scrutiny over the impact of the technology increases.

On Wednesday, Anthropic, Google, Microsoft and OpenAI launched the Frontier Model Forum, with the aim of “ensuring the safe and responsible development of frontier AI models.”

In recent months, the US companies have rolled out increasingly powerful AI tools that produce original content in image, text or video form by drawing on a bank of existing material. The developments have raised concerns about copyright infringement, privacy breaches and that AI could ultimately replace humans in a range of jobs.

Enlarge / Welcome aboard the orbital HQ, readers! (credit: Aurich Lawson | Getty Images)

Welcome back to our series on how Ars Technica is hosted and run! Last week, in part one , we cracked open the (virtual) doors to peek inside the Ars (virtual) data center. We talked about our Amazon Web Services setup, which is primarily built around ECS containers being spun up as needed to handle web traffic, and we walked through the ways that all of our hosting services hook together and function as a whole.

This week, we shift our focus to a different layer in the stack—the applications we run on those services and how they work in the cloud. Those applications, after all, are what you come to the site for; you’re not here to marvel at a smoothly functioning infrastructure but rather to actually read the site. (I mean, I’m guessing that’s why you come here. It’s either that or everyone is showing up hoping I’m going to pour ketchup on myself and launch myself down a Slip-'N-Slide , but that was a one-time thing I did a long time ago when I was young and needed the money.)

How traditional WordPress hosting works

Although I am, at best, a casual sysadmin, having hung up my pro spurs a decade and change ago, I do have some relevant practical experience hosting WordPress. I’m currently the volunteer admin for a half-dozen WordPress sites, including Houston-area weather forecast destination Space City Weather (along with its Spanish-language counterpart Tiempo Ciudad Espacial ), the Atlantic hurricane-focused blog The Eyewall , my personal blog, and a few other odds and ends.

Enlarge (credit: AMD)

A recently disclosed bug in many of AMD's recent consumer, workstation, and server processors can cause the chips to leak data at a rate of up to 30 kilobytes per core per second, writes Tavis Ormandy, a member of Google's Project Zero security team. Executed properly, the so-called "Zenbleed" vulnerability (CVE-2023-20593) could give attackers access to encryption keys and root and user passwords, along with other sensitive data from any system using a CPU based on AMD's Zen 2 architecture.

The bug allows attackers to swipe data from a CPU's registers. Modern processors attempt to speed up operations by guessing what they'll be asked to do next, called "speculative execution." But sometimes the CPU guesses wrong; Zen 2 processors don't properly recover from certain kinds of mispredictions, which is the bug that Zenbleed exploits to do its thing.

The bad news is that the exploit doesn't require physical hardware access and can be triggered by loading JavaScript on a malicious website. The good news is that, at least for now, there don't seem to be any cases of this bug being exploited in the wild yet, though this could change quickly now that the vulnerability has been disclosed, and the bug requires precise timing to exploit.

Enlarge (credit: Evgen_Prozhyrko via Getty )

For more than 25 years, a technology used for critical data and voice radio communications around the world has been shrouded in secrecy to prevent anyone from closely scrutinizing its security properties for vulnerabilities. But now it’s finally getting a public airing thanks to a small group of researchers in the Netherlands who got their hands on its viscera and found serious flaws, including a deliberate backdoor.

The backdoor, known for years by vendors that sold the technology but not necessarily by customers, exists in an encryption algorithm baked into radios sold for commercial use in critical infrastructure. It’s used to transmit encrypted data and commands in pipelines, railways, the electric grid, mass transit, and freight trains. It would allow someone to snoop on communications to learn how a system works, then potentially send commands to the radios that could trigger blackouts, halt gas pipeline flows, or reroute trains.

Enlarge / An AI-generated image of a chatbot in front of library shelves. (credit: Benj Edwards / Stable Diffusion)

On Thursday, OpenAI announced a new beta feature for ChatGPT that allows users to provide custom instructions that the chatbot will consider with every submission. The goal is to prevent users from having to repeat common instructions between chat sessions.

The feature is currently available in beta for ChatGPT Plus subscription members, but OpenAI says it will extend availability to all users over the coming weeks. As of this writing, the feature is not yet available in the UK and EU.

The Custom Instructions feature functions by letting users set their individual preferences or requirements that the AI model will then consider when generating responses. Instead of starting each conversation anew, ChatGPT can now be instructed to remember specific user preferences across multiple interactions.

Enlarge / AlmaLinux lets you build applications that work with Red Hat Enterprise Linux but can't promise the exact same bug environment. That's different from how they started, but it's also a chance to pick a new path forward. (credit: AlmaLinux OS)

I asked benny Vasquez, chair of the AlmaLinux OS Foundation, how she would explain the recent Red Hat Enterprise Linux source code controversy to somebody at a family barbecue—somebody who, in other words, might not have followed the latest tech news quite so closely.

"Most of my family barbecues are going to be explaining that Linux is an operating system," Vasquez said. "Then explaining what an operating system is."

It is indeed tricky to explain all the pieces—Red Hat, Red Hat Enterprise Linux, CentOS, CentOS Stream, Fedora, RHEL, Alma, Rocky, upstreams, downstreams, source code, and the GPL—to anyone who isn't familiar with Red Hat's quirky history , and how it progressed to the wide but disparate ecosystem it has today. And, yes, Linux in general. But Vasquez was game to play out my thought experiment.

Enlarge / A Z16 Mainframe.

Mainframe computers are often seen as ancient machines—practically dinosaurs. But mainframes, which are purpose-built to process enormous amounts of data, are still extremely relevant today. If they’re dinosaurs, they’re T-Rexes, and desktops and server computers are puny mammals to be trodden underfoot.

It’s estimated that there are 10,000 mainframes in use today. They’re used almost exclusively by the largest companies in the world, including two-thirds of Fortune 500 companies, 45 of the world’s top 50 banks, eight of the top 10 insurers, seven of the top 10 global retailers, and eight of the top 10 telecommunications companies. And most of those mainframes come from IBM.

In this explainer, we’ll look at the IBM mainframe computer—what it is, how it works, and why it’s still going strong after over 50 years.