Sure, the rest of the market is barely keeping pace (4%) with inflation (3.4%) YTD; but but but, NVidia is shiny and AI is going to save us all! Pay no attention to the billionaires promising us a post-scarcity world while building massive survival bunkers in the background.

In server-grade hardware, sure. And you paid $$$$$ for it.

The earliest "consumer" grade RAID cards, by contrast, were basically just softRAIDs with a built-in IDE controller - And that's still true to a large degree; the "RAID-mode" in your gaming PC's BIOS is still merely offloading the hard work to the CPU.

I'm not claiming Linux' softRAID implementation was or wasn't faster than cheap RAID hardware, but I have no problem believing it could at least be comparable.

You're thinking of what amount to ultra-SFF PCs. In the truly embedded space, having a fixed (and usually tiny) amount of RAM on-die is the norm rather than the exception.

That said, Meta clearly isn't targeting 68HC05s here, so the distinction is largely meaningless.

Yes, pedant, I left out the word "qualified". I took it as a given the very people just laid off are qualified to do their own jobs. If that's my greatest sin today, I'll call it a pretty damned good day.

"Anyone I disagree with must be an rsilvergun sockpuppet". That's so broken I'm not even sure what kind of logical fallacy to label it as. Poisoning the well? A variant of Reductio ad Hitlerum? Genetic fallacy by proxy?

It's funny - Thanks to your efforts, I suspect virtually every /.'er knows who rsilvergun is at this point. And I can promise you, the vast majority of us find him far less obnoxious than his stalkers.

I don't think the OP's point was that the solutions are not ever useful, but the "nobody cared for a long time" part. If no one cares about a problem, then no one is spending their time trying to solve it. Whether or not it may be useful in some way is not the point, but if no humans are trying to solve it, then humans are not going to solve it.

So now we have AI that will simply focus on whatever it is told to focus on, looking at problems esoteric enough that no one wanted to waste time on them, and finishing some of them.

You're missing the point here entirely, and ironically you are actually providing counter-evidence.

It appears that ChatGPT, etc, are taking information provided to them by people like you in the form of conversing with the agent and providing files, and using that information to generate output for others. This was shown where a mathematical solution ChatGPT came up with was pretty much exactly like an unpublished solution that had been discussed with ChatGPT by a couple researchers

Do you really think you are the only person in the world discussing that particular thing with the AI agent? You say you're very familiar with the literature in that field. Has your brain literally been built upon every word of that information that has ever been published, like an LLM? Of course not. So it doesn't matter how familiar you think you are with the literature, an LLM's "database" of that literature is beyond anything you could achieve.

Your example does not prove the LLM didn't connect existing dots, or produce that solution from existing information, either published or unpublished that was provided to it in chat sessions like yours.

There's quite a lot of nuance here, I guess, which isn't unexpected. It's quite a useful anecdote. At the very least, we can't just wave it away like it's not useful. It may generate sub-standard paintings or completely miss the point of art, but if it makes your proofs better, it seems insane to not have it help.

It's explicitly fraud, because in sponsoring H1-Bs, a company needs to attest that it's unable to fill open positions with available domestic talent. If they just fired 6000 domestic workers, there are clearly 6000 domestic workers available for those roles.

Granted, it's almost never enforced, which is what most of us in IT have been screaming about for decades.

And I may just need to buy a lottery ticket tonight while dodging flying pigs, because the current administration has managed to hit "stopped o'clock" for the second time!

That's kind of the point of most Malthusian scenarios. No, the population won't really grow exponentially forever - But it will grow exponentially within the availability of any constraining sub-exponential resources (e.g. food or physical space).

What happens when we hit that limit varies, but a common natural pattern is exponential growth followed by population collapse, rinse wash repeat. That's all well and good when we're talking about squirrels or deer; not so pleasant when we're talking about 4B humans starving.

We need companies with some balls to put technophobic politicians in their place. Don't appeal, just maliciously comply: "Okay, we've blocked all of France. Welcome back to the dark ages."

I'm pleasantly surprised to see a rare sane decision from courts who think their word is physical reality, but this shouldn't even have come up in the first place.

I think the original article was somewhat tongue in cheek.

And those corrections require atomic clocks.

You think LLMs would write "Slashdot doesn't have table support, so it's hard to do side-by-side well, but:"? You think LLMs like writing sentences that end in conjunctions that express incomplete thoughts? Do you think LLMs write "high frequences", "naively pairs", call H.265 X.265 (I was thinking of what you refer to it as in ffmpeg), imbalanced parens after "lacks motion vectors",etc?

Don't get me wrong, I am a LLM user, I absolutely do use it sometimes for search, fact checking things I'm writing, spelling/grammar checks (clearly not that time, lol), etc. But I write my own posts.

Oh, you think this is about doing the right thing... the US government... the current US government? I mean it's cute that you believe that.

Adobe, MS, et al. are being targetted because they've got deep pockets. A few million given to the right orange madman and this problem goes away. They don't even have to do it in secret any more.