I thought that was the free space.

He and Trump deserve each other.

They’re all towers. But the buttons are all pretty shallow with very light actuation force required.

And they all happen to be situated such that the corner which has the button is the corner furthest away from the desk, so when she jumps up onto the PC as a platform to get ready to jump onto the desk, her feet are all grouped up right in that corner.

And you can imagine that if she’s crouched down ready to jump, and I put my arm out to prevent her from jumping from the tower to the desk, that’s a lot of pressure all applied to her little toe beans.

It’s an unfortunate coincidence. But that experience, together with seeing this Mac Mini design, has made me wonder why we tend to put a button with such drastic effects right out in the open like this.

That sounds…

Easier to get almost right than actually learning the subject.

Much, much harder to get completely right than actually learning the subject.

So yes, basically the archetypal use case for LLMs.

I’ll just stick with covering it up. Without fail, if I leave it uncovered my cat will press it. She’s even held it long enough for a forced shutdown twice that I can think of.

Yes, but even pushing it will bring up a prompt, which is annoying. And also my cat has held it down long enough to force a shutdown on my media server before, as well as on my wife’s PC during Overwatch.

This is the dumbest timeline.

As someone who has to use heavy/taped-on little toys to cover the power buttons on my PCs or else my cat invariably opens a shutdown dialog in the middle of something… Thank you.

Sure. And you can buy a dirt bike cheaper than an ATV. Yet people still buy ATVs.

I’m not gonna do iOS dev or ML on a GMKtec no matter how cost-effective it is, just like I’m not gonna play x86 Windows games on a Mac even if I win a maxed-out unit in a giveaway.

Is there even a better ARM SoC? All I know of is the Snapdragon X Elites, which are either on par or slightly below the M4. And you can only get them in a laptop form factor at this point, cuz they cancelled the mini-PC dev kit.

True. It was just the first comparison I saw when I searched for M4 benchmarks.

Really, AMD isn’t even a fair comparison because we’re talking about an ARM SoC here. So maybe the Snapdragon dev kit that ultimately got cancelled?

It was supposed to be $900, for a special Snapdragon X Elite, 32GB RAM, and 512GB SSD.

cpubenchmark.net has comparisons to other X Elite chips, putting them pretty much on-par with the M4 or maybe just below it.

With the same amount of RAM and storage in a Mac Mini, you’re talkin $1200. So, $300 premium for a device that’s maybe 2-8% better, has retail support instead of being a dev kit, and… well, actually exists. It’s not a slam dunk for the Mini, but it’s clearly not a rip-off either.

M4 reportedly outperforms Intel’s Core i9-14900KS by 16%. That CPU alone is over $600.

This is where we need something other than copyright law. The problem with generative AI companies isn’t that somebody looked at something without permission, or remixed some bits and bytes.

It’s that their products are potentially incredibly harmful to society. They would be harmful even if they worked perfectly. But as they stand, they’re a wide-open spigot of nonsense, spewing viscous sludge into every single channel of human communication.

I think we can bring out antitrust law against them, and labor unions are also a great tool. Privacy, and a right to your own identity factor in, too. But I think we’re also going to need to develop some equivalent of ecological protections when it comes to information.

For a long time, our capacity to dump toxic waste into the environment was minuscule compared to the scale of the natural world. But as we automated more and more, it became clear that the natural world has limits. I think we’re headed towards discovering the same thing for the world of information.

Looks like certified mail, not notarized. But still, ridiculous. https://retrofitness.com/faq/

I’ll keep saying it: We’re watching the information equivalent of Kessler Syndrome in real time.

Good call, thank you.

Also: Referencing Wikipedia in this context is kinda funny.

So I put 2 and 2 together, and decide this whole thing is pissing me off.

Still waters run deep.

Unleashing generative AI on the world was basically the information equivalent of jumping headfirst into Kessler Syndrome.

“We can’t solve climate change by repeating our past behavior. Let’s ignore climate change and build a machine that regurgitates our past behavior.”

This just in: Author/professor/CEO whose books/classes/company are about manipulative technologies… voluntarily installs manipulative technologies.