Category Archives: News

Article note: Interesting. Intel and AMD have each pumped at least a year of funding into this thing, and both backed out after their initial commitment, but there is now a vaguely-complete open source CUDA runtime for other targets. The HN thread seems to have a hard time understanding that Nvidia has (1) been co-evolving CUDA and their chips, so executing CUDA code on non-Nvidia-architecture parts will always be at a disadvantage. (2) Nvidia embedded a bunch of blackbox runtime behavior into CUDA that is necessary for the execution semantics but not exposed in the APIs, because it was convenient and/or to throw off competing implementations. (3) Nvidia has shown a great deal of willingness to spend resources preventing competing implementations (like when they bought PGI from ST) (4) Nvidia has been treadmilling the CUDA APIs for years, to iteratively improve user experience and/or to throw off competing implementations. No reason to think they'll stop. (5) The way _most_ of the user base works, CUDA looks more like a high-level assembly, not the language they work in, so there isn't actually that much CUDA code in the world. For the big money example, most of the AI boom has been built on higher level abstractions (...and reusing other people's code), so you only need to write a eg. torch backend for a different target, not for every user to rewrite a bunch of their own code. Rolling with that example, there are _already_ Intel/oneAPI and AMD/ROCm backends in torch that Intel and AMD should prefer to perpetually playing catch-up with CUDA.

Article note: Orbital (especially orbital2/brown) has been in the soundtrack of my life for decades, but I hadn't given their recent stuff a listen until I saw this. It's good. I'm really enjoying Optical Delusion.

Article note: LUnix is among the oldest and most impressive "Levering UNIX-like systems onto unsuitable potatoes" projects. 6502-likes are unsuitable for C-like languages, unsuitable for preemptive multitasking, unsuitable for addressing reasonable amounts of memory... and here we are, doing it on a even feebler variant.

Article note: "This general purpose tool exposed that several car manufacturers did negligent things in shipping products, so we're going to go after the tool instead of holding the manufacturers accountable."

Enlarge / A Flipper Zero device (credit: https://flipperzero.one/)

Canadian Prime Minister Justin Trudeau has identified an unlikely public enemy No. 1 in his new crackdown on car theft: the Flipper Zero, a $200 piece of open source hardware used to capture, analyze and interact with simple radio communications.

On Thursday, the Innovation, Science and Economic Development Canada agency said it will “pursue all avenues to ban devices used to steal vehicles by copying the wireless signals for remote keyless entry, such as the Flipper Zero, which would allow for the removal of those devices from the Canadian marketplace through collaboration with law enforcement agencies.” A social media post by François-Philippe Champagne, the minister of that agency, said that as part of the push “we are banning the importation, sale and use of consumer hacking devices, like flippers, used to commit these crimes.”

Article note: Interesting. I use VirtualBox because USB device passing actually-fucking-works (I can hot-plug a thing that I only have Windows drivers for into my Linux host running a Windows VM, with the vid:pid filter set, and it just gets delegated), and the libvirt front-ends are sort of unbelievably terrible where VirtualBox's is decent. I'd really like to migrate to a fully open source not-Oracle-licensed-plugin solution, so things that move the functionality closer together - even from the other direction like this - are exciting.

Article note: It's sort of amazing that a thing named for a vendor jail appliance is one of the last great examples of successful ecosystems built out of vendor independent open protocols/standards instead of some kind of rentseeking roach motel. Even though all those rentseeking roach motels are hosted on top of infrastructure that is previous successful vendor-independent open standards.

Article note: It's such a good (fictional) example. It's also such a classic "almost no one checked the references, and the person who propagated it misunderstood what they read." scenario.

Article note: The moderately priced, readily navigable, nearly one-stop-shop, largely ad-free option has fragmented into a bunch of overpriced rent-seeking entities with small and constantly shifting catalogs hidden behind wretched UX, with increasing ad penetration. The conditions that incentivized piracy returned, so piracy is ramping back up. Not a mystery. The media industry can build a large number of interchangeable distributors all of which can get to all the content via reasonable terms (like a retail or FRAND situation), or consolidate into a small number of services (studios produce content and sell via an independent distributor) or... try to squeeze and everyone goes back to piracy until the industry is forced to contract and reconsider, which seems to be the plan.

Article note: This awesome work by people who seem delightful, but unless we address the structural incentives that make fraud (and other flavors of chicanery like deceptive hyping, fad chasing, and simulation-encodes-bogus-assumptions tautology proving) more effective as career strategies than honest research work (and/or instructional work!), it's a losing battle.

Article note: Google remembered another useful product and killed it.

Enlarge (credit: Sean Gallup | Getty Images)

Google will no longer be keeping a backup of the entire Internet. Google Search's "cached" links have long been an alternative way to load a website that was down or had changed, but now the company is killing them off. Google "Search Liaison" Danny Sullivan confirmed the feature removal in an X post, saying the feature "was meant for helping people access pages when way back, you often couldn't depend on a page loading. These days, things have greatly improved. So, it was decided to retire it."

The feature has been appearing and disappearing for some people since December, and currently, we don't see any cache links in Google Search. For now, you can still build your own cache links even without the button, just by going to "https://webcache.googleusercontent.com/search?q=cache:" plus a website URL, or by typing "cache:" plus a URL into Google Search. For now, the cached version of Ars Technica seems to still work. All of Google's support pages about cached sites have been taken down.

Cached links used to live under the drop-down menu next to every search result on Google's page. As the Google web crawler scoured the Internet for new and updated webpages, it would also save a copy of whatever it was seeing. That quickly led to Google having a backup of basically the entire Internet, using what was probably an uncountable number of petabytes of data. Google is in the era of cost savings now, so assuming Google can just start deleting cache data, it can probably free up a lot of resources.