Just for the record, it's Nintendo suing this time, not Sony. Sony is only responsible for the precedent, which says that emulation is legal.

Nintendo may be repeating their strategy here, though.

I mean, that's fair, but you can still bankrupt the devs and it's not always easy to replace them, nor will a lot of people be eager to step up if they think they might be next.

> Nothing illegal was done.

The main precedents in emulation legality are Sony v. Connectix and Sony v. Bleem, where Sony lost both cases, but drove the defendant out of business via the lawsuit. Make of that what you will.

Windows 11 is a lot worse than Windows 10 at handling headphones, from what I've seen. You plug some in and instead of just working, some apps don't get the memo and sometimes the system doesn't figure out that you did anything at all.

They were supposedly just poisoning the prompts. You'd say "show me vikings" and they'd change it to "show me vikings black" or such, including a random ethnic heritage on the end. Some people exposed this by asking for comic book style images, which got words from the prompt in the text boxes, or by adding "and a sign which says" to the end of their prompts.

I guess later they might add the words to the start, but similar creative prompting would likely help expose this as well.

Ah, I must have been lucky with my applications and the GPU. I actually started with Mesa OpenCL which seemed fine at first, but there were timeouts on my longer-running kernels, and ROCm has none of that.

I do fairly simple but heavy numerical stuff, and it turns out AMD cards are much better for these uses. For example, double precision float speed is only half of single precision, whereas DP on Nvidia consumer cards is much slower. It's easy to check this as Nvidia also runs OpenCL, so in my experience ROCm isn't particularly slow, quite the contrary.

I've refused to learn CUDA as I don't want my code to be at the mercy of a single GPU maker. The project looks interesting at first glance, but it seems like they'd just be playing catch-up with new CUDA developments. Open standards are much nicer, and besides OpenCL, I've got the impression that ROCm itself (which is open source) provides a lot of CUDA-like higher-level functionality.

This doesn't seem right at all. I'm using ROCm drivers for OpenCL applications right now on an RX 6600, an affordable consumer GPU. ROCm is open source (see the Wikipedia link in the summary) so there isn't an immediate danger of dropping support for a given hardware; you can always fork it and backport things etc.

AMD also provides closed source drivers so perhaps you're referring to them?

The idea of 3DNow instructions hasn't gone anywhere, it's just that Intel pushed their own version called SSE, which was eventually adopted by AMD as well. From what I understand, NPUs, tensor cores etc. are just continuing the trend of wider SIMD units and shouldn't be too application-specific. It's just marketing that likes to name them after the most popular application, just like AMD's floating point SIMD unit was named for 3D graphics.

The way I see this, it's just SIMD units getting wider. Some of us have been working with large vectors/matrices for decades, and wonder what's so "neural" about the latest matrix multiplication unit. It's nice to se the wide SIMD trend continue, as long as the hardware doesn't get too application-specific, so it remains useful after the current AI craze wears off.

I've had it with these motherfucking eels on this motherfucking hovercraft!

Honeywell? In Europe? What a load of Bull.

There was a good example of this problem in the recent /. article on bird-identifying binoculars. They don't seem particularly useful for learning, since they don't tell how they came up with the results. If it's not a completely black box, it should be possible to extract this "how", but it probably wouldn't make much sense to humans. Current AI makes a very bad replacement for teachers, as it cannot teach the actual process but only gives answers.

I remember looking at Gemini some time ago, and I thought it would be a great fit for my minimalist website. But I don't really see much point in it besides the geek factor. If I'm trying to reach a wide audience with my website, it should be accessible in a standard way without extra apps.

If you like the minimalist geek esthetic of the early web, you can still do it using the same HTTP and HTML we used back then; genuine vintage tech instead of this wannabe retro approach. There's also no need to include social media integration or tracking.

we are programmed to deceive.