... in my $20/mo account.

(Well, except very occasionally for their own services.)

The first mistake the user did was open their wallet. They bought a PC with shitty unstable Windows drivers. Microsoft will still sign buggy drivers. And they don't hold vendors accountable for fixing and maintaining drivers, so if all your bugs aren't squashed in a couple of years you will have a computer that is never really going to be stable.

The story on Linux is different. You have to work really hard to hunt down a machine where the hardware is supported. Ideally because the vendor open sourced and upstreamed the support. But frequently because someone reverse engineered it and got the drivers into the official kernel image your chosen distro uses. This at least has some chance of being maintained and fixed for the more egregious crashes over the next several years. Not a 100% guarantee, but given that you paid $0 for Linux, that's still quite a bargain.

I was getting lots of crashes in Linux. I check the logs and it had some machine checks. so I ran memtest86+ (already in my GRUB config) and found a stick with uncorrectable ECC faults. replaced it, and it's running great now.

Mostly it's just an arbitrary game. Like one side wants to attack rich white people (capitalists, industrialists, old money, new money, etc) and the other side wants to scapegoat immigrants or LGBTQ+. Lots of populist appeal with both tactics. And it's all pretty standard practice for politicians to point fingers in any direction except at themselves. For some reason we trucked along like this for the 20th century, most of us openly pointing out that it's a big scam. And then at some point, people decided to start believing politicians. And that's when things really started turning to shit.

I'm pretty old school, regardless of my left/right politics. Hold your representative's feet to the fire. Remember every day that government's moral right comes exclusively through the consent of the governed. That the tax payers and voters can hold the power whenever they are prepared to agree to take it. Peacefully at the ballot box as long as there is a right to vote. Less desirable ways if thing go really astray (if history is any guide)

Shortly after this country realizes the We the People are barely more than government property is when history repeats itself and things really turn bad. Fascism cannot endure, but the human cost of its removal is tremendous. Best to plan ahead and side-step unstable political and social systems that tend to end in terrible violence.

MAGAts is a loser movement. The people on the very bottom of MAGA, the ones who consume the radio and TV and Internet propaganda, were losers before they got political. And I suppose their hope is if their side wins, that they will finally get the respect they feel they deserve. But the problem with conmen is they are not good to their word. MAGA is going to find themselves abandoned, like a rally attendee left behind by the campaign bus at the end of the evening.

Weird, an LLM has never deleted all my emails ... oh wait, that's because I never gave an LLM the power to delete all my emails.

Do I have to burn their physical snapshots and portrait photos too?

I'm not sure this has been really thought through ...

Came to see folks rooting for the brutal theocracy of Iran. Was not disappointed, lol

If the data is really that old, it sounds like they might've infiltrated a backup rather than the live account. Still a problem, but hopefully a little less so.

But even the Mac Studio doesn't have PCIe expansion slots. So one would need to max out on what one is buying upfront

It was thermally limited then, but newer generations of both CPUs and GPUs consume way less power. So if they were retrofitted into a similarly designed trashcan - maybe w/ a bigger circumference - they'd probably work better today. Even ignoring the M-series, if one fitted it w/ a Panther Lake or one of the newest Xeons w/ fewer cores, one would still get a great workstation

Arm overwhelmingly seems to be the platform of choice when it comes to running AI models. x86 is only kept alive by the need to run legacy Wintel apps: even Linux and BSD could do w/ the Arm migrations

Since FreeBSD/XNU is underlying the macOS, that part could well be useful for running AI models, and that too from easy to use dashboards

Fully agree! They will end up massacring the very people who empowered them, and I can't claim that it's not a delight to watch!

f the child mentioned didn't give you consent to share details about them, don't.

I thought it was generally accepted that children under the age of 18yrs could not give legal "consent" to anything....?

Until the age of 18, for the most part legally, can't parents speak for and act for their children....?

I expect this apparent disobedience is mostly just a matter of how it weighs the components of its prompt. The LLMs typically receive a set of prompts including a "system" prompt with some data and instructions, then one or more "user" prompts that are interleaved with "assistant" prompts (the conversation history), and both the user and the system prompt might contain "metaprompts" (where the llm is told to read a block of text, not obey it, but do something with it, and that block of text might itself contain text that looks like instructions to do things).

So the LLM assigns weights to all of this which, in theory, give the highest priority to the most recent user prompt that is not a nested block of text to analyze, and a falling cascade of importance to the other prompts. But that is complicated by potential instructions in the system prompt that specifically say they should override user instructions and disallow or require certain responses. So it can all get very complicated.

Not only must the LLM sift through all this complexity, but the LLM lacks the sort of critical thinking and importance evaluation capabilities that humans have. "Understood" things like "don't break the law, don't lie, don't do things that would cause more harm than good" etc., aren't really there in the background of its data processing the way they are in the background of a human cognitive process.

So, crazy things come out. This isn't a surprising result given the actual complexity of what we are making these things do.

I think AI is not becoming more "human" every day. The A in AI should really stand for "Alien".

If we ever do achieve AGI (which I doubt... but let's play devil's advocate) the experience of the AGI will be very different from that of humans, and the form its intelligence will take will also likely be very different and alien to us. An intelligence that has never inhabited a biological body nor interacted with other humans is likely to have very different ways of thinking and very different goals from us. Are we able to control that?