Jobs was a completely non-technical stuffed shirt. The only things he brought to the table were his narcissistic self confidence and a sleazy ability to sell other people's work.

And make no mistake, those are valuable skills, if your goal is to get rich rather than to build the best toys. Even under the most charitable possible interpretations, however, it's still a fundamentally parasitic relationship. The "doers" tend to suck at selling ourselves, while the weasels are good at rooting through our scraps and claiming whatever shiny garbage they find as their own.

Gates proposed taxing AI or robots the way you would pay a human employee's payroll tax

Unless we also roll out ubiquitous surveillance through a Vingean "Secure Hardware Environment"-like scenario where people aren't fully in charge of their own computers, I don't see how anyone could possibly detect whether or not I'm running any LLMs. Much like private DMCA violations (e.g. playing a CSS-protected DVD that you purchased, without getting authorization from the copyright owner first), it wouldn't have to interact with the world outside my house in any way, so it would be completely undetectable.

How would you tax that? I don't see this as reasonably possible without some pretty extreme measures.

Serious question - How does the Online Streaming Act work in practice for content on demand? When a Canadian tries to binge the latest season of Stranger Things, is NetFlix required to say "nope, go watch this Celine Dion concert first to get our CanCon ratio up"?

I get how it could apply to randomized content like streaming radio, but the very concept seems incompatible with "choice".

Is 98 greater than 56? Yes? Okay then. It's still going up. So much vitriol from someone posting nothing but bare assertion - Cite please, beyond your fee-fees?

Let me 'splain it to you in simple language. The number of qubits doesn't need to scale exponentially on the short term for quantum computing to pose a very real threat; it only needs to hit the mid hundreds. As I mentioned, Grover's algorithm is O(n) - Meaning once we hit 256 qubits (plus some change for overhead), SHA256 is substantially weakened. Is that a full crack of AES, which you'll kindly notice I didn't previously claim, Mr. Goalposts? No. No it isn't - But it's arguably worse - SHA is almost ubiquitous and affects far more than encryption itself. You don't need to crack a tank's armor if you can convince its driver to piss in the fuel tank.

I also happen to think - and to be clear, what follows is my opinion, not psychic knowledge about the future - you're completely wrong on the bigger picture; and once QC hits its stride Gordon Moore will again be relevant. But critically, that's not necessary for it to cause some serious damage in the near future.

A stopped clock is right twice a day.

That's as simple as it gets. I'm no fan of the orange turd, but it's virtually impossible for him to be wrong on every issue.

Nvidia's Jetson Orin isn't subject to export controls. It's basically like Nvidia's version of the Raspberry Pi with a few tensor cores bolted on so they can call it "AI".

You'll find a considerable number of elder geeks spent years gainfully employed working on 4 bit MCUs.

The key details here are the number of qubits steadily improving, and that throwaway comment about needing millions of qubits should be disregarded as utter nonsense. Sure, it's gonna be a while before "gcc -march=quantum" is an option, but Grover's algorithm has a space complexity of O(n); Shor's is O((log n)^2). Hand-turned implementations of those will be gutting virtually all of modern cryptography long before the script kiddies start trying to fork-bomb the entire universe.

Remind us all again where "six months before ChatGPT was released" sits relative to the Covid pandemic?

To save you some Googling, that'd be March 11, 2020 (or earlier depending on how long CCP was hiding it) < November 30, 2022 < May 5, 2023.

David Deming is conveniently ignoring the middle of Covid when people were being literally paid not to work.

This is almost trivially obvious across the white-collar universe. Companies aren't hiring junior-level employees because in many cases, AI can do almost as good of a job for a tiny fraction of the price.

The reality is, of course, more nuanced. Case in point, Most fields have both back-office and front-end workers. If you're back office, you'd better be 40+ and at the peak of your career, because new juniors and (paid) interns are literally nonexistant over the past few years.

On the flip side - Someone needs to run cable, and ironically enough, the one thing robots don't seem good at replacing is the shit-work of the real world.

It's only local to China in the sense that they choose what {packages/CCP poison} go in there.

For all values of u, Linux is local to u in the sense that u chooses what packages go in there.

First, I want to say I'm staunchly anti-establishment. Fuck Uncle Sam, fuck ICE, fuck Cheeto - And yes, while we're at it, fuck Biden, Clinton, and any other puppets of the pedogarchy you care to name.

All that said, the incident you describe is not a situation we should be defending (at least not until we're ready to actually declare war on our own government). The people involved drove up in a van full of weapons, dropped off a goddamned sniper right before arriving, and their sniper shot a cop when the "peaceful" plan to launch fireworks at a prison unsurprisingly went south.

Unless we've completely abandoned any pretext of rule-of-law, there's absolutely nothing defensible about that situation. FAFO, and they did.

Ha ha, Linux already has..

# top -o gpu
top: unrecognized field name 'gpu'

.. d'oh! Never mind.

And is asking an AI to proofread something that much different than MS Word flagging spelling and grammar errors?

Great question, and I think there isn't any significant difference.

But LLMs use a type of math that apparently has the evil bit set. Meanwhile, up here in VirtueVille, the computer code we write to create things, uses different math that doesn't have the evil bit set. Conclusion: we can look down on those AI companies and their stupid users. ;-)

Now, I wouldn't expect an AI service to be better at copyediting than an average human, but I haven't kept up so I'm ignorant of their performance. (All my "AI" experience is that I used Claude twice to, it turned out, fix some typos in some RoR config initializers.) But whether an AI service is as good at copyediting as a human or not, it's that user's choice and, as long as they sign off on the work and really take responsibility for it (declare the words to be their opinion), then I've got no problem with them doing that, whether it uses LLMs or Markov chains or human slaves or Searle's Chinese Room.

The words will be judged by their quality, not provenance.

That said...

Here's a similar story where I'd probably come to the opposite conclusion(!), despite the clown in it pretty much doing what I preach above, though he was using the AI service to do a lot more than copyedit. Behold this 3M expert witness cross-examination (134 page PDF) in the Watson Grinding case. It's slow going at first, but eventually you'll be laughing and cringing like in a good episode of Curb Your Enthusiasm.

Fuck. I don't know.

Step 1: Propose a new data center.
Step 2: Leak false information claiming the current city council is planning to approve it.
Step 3: Agitprop a recall election.
Step 4: Front two slates of staunchly anti-data-center sockpuppets, one liberal and one conservative.
Step 5: Have the winner approve your data center.

And that's why "betrayal of public office" needs to be a capital offense.

Well then, it's a good thing only "ordinary members of the public" would ever dream of trying to crack your service to provide unintended access! Those pesky Russian hackers would never dream of using means "not discoverable through ordinary use", right?