The CEOs are lying and LLM-based AIs are greatly overrated. They're helpful, but they're more like an enhanced version of Stack Overflow. If you know what you're doing, they slow you down. For example, I know Java really really really well. When I have Claude Sonnet or Opus generate Java, it takes them waaay longer than it takes me to write it, so I can't be lazy and outsource it to them. It seriously takes the AI minutes to do something it takes me seconds to do.

OK, so what about things that take me minutes?...like writing a unit test?...well, that's my favorite use case for AI. I'd LOVE to see it succeed, but I work primarily in Java, a compiled language...and it is strict about getting things right, so I see the errors immediately. Python users who vibe code, just ship bugs and let their users find them. OK, so with enough tries, it barfs out a unit test. It looks pretty good...afterall, LLMs are top-notch guessers. Unfortunately, the unit test is completely wrong and useless...so I have to go make it actually test the code instead of testing bean getters and setters and stupid shit like that. The scary part is that it looks good. It looks correct. But it often isn't, so you have to evaluate line-by-line.

One of my coworkers is more bullish on AI and introduced over 20 bugs last week with his AI slop, including undoing half my fixes for the week. His boss is consider putting him on a Performance Improvement Plan for his AI use. He's not a dumb guy. He just didn't understand the pieces I worked on, didn't read my docs and comments, and was fooled by the AI when it undid all my code to make his component's test pass. He is in India and didn't wait for me to review the code and had someone in IST review it who knew even less.

The only powerful use case I've found for them is for scenarios where you need to work with a technology you used to know well, but have forgotten. As a backend software engineer, this would be front-end code, RegExes, obscure stored procedure method calls, etc. For RegExes, I write them maybe 2x a year...so I never am confident of things I write. I can review the code better than I can write it from scratch.

If you've never used a technology, the code is unreliable. At best, it might save you some time learning. For example, if I had to write something in C#, a language I barely touched 20 years ago, it might double the ramp-up time, but I'd still have to spend a lot of time learning the fundamentals of the language. It would take the place of a really well written book...helpful, but not a game changer.

The point being...AI doesn't tangibly save time. It might save a bit under some circumstances, but not enough to justify layoffs. The CEOs are full of shit. They're AI washing routine layoffs. Either they overhired, or they wanted to shut down products and features or they wanted to get rid of dead weight....but apparently it's more fashionable to overtly lie to investors? It baffles me why shareholders haven't filed a lawsuit against Beinhoff....or any other CEO.

I think they're arguing that they are the third thing, a message parlor.

Unions are a real-life strategy because they work. Divide-and-conquer is also a real-life strategy, because it works too.

Thus, I think the truth of your statement all depends on whether you look at this conflict between government and the the people, from the point of view of the attacker, vs the point of view of the defender.

Children do not have the maturity that is required for unfiltered access to the adult world

But they used to. In the 1980s, nobody dared to say in public, that 17-year-old me should not be allowed to visit public (or even university) (or even medical) libraries. (Or if someone did, they were still very obscure and unpopular, little more than a glimmer in the left's eye.)

If I may, could I narrow down which of these two things you think is best? First, there's exactly what you said above..

Kids have no right to use end-to-end encryption without parental consent

..but I've altered it:

Kids have a right to use end-to-end encryption unless denied by a parent

Did I make it better, or did I make it worse?

> All in all, it was just ...so american.

My father often said that if you want tits and explosions to watch with your brain turned off, you watch American. If you want something that doesn't spoon feed you, watch British. If you want to commit suicide, watch something Scandinavian.

Things have evolved a bit since those days, but it's not the worst general rule to start with.

And that the 'Newtonian' physics of Gravity wasn't.

I'm a huge buzzkill. I can watch Pacific Rim with a smile on my face, but tell me you've made something realistic when you haven't and it nukes suspension of disbelief for me completely.

For attracting venture capital, about 10 years ago.

OK, so they make their own fabs. Aren't the components in short supply? Aren't they paying more for ASML machines?...etching chemicals?...high-end silicon? It's like the Strait of Hormuz closing...the USA is now an oil exporter, so we're not impacted, right?...nope!!!...because everyone who is impacted is now buying from the USA and our gas prices went up nearly as much as everyone else's. OK, so Apple can make their own chips?....what's to stop them from jacking up the price since the market is willing to pay? They make more money, but they're not going to be virtuous and underprice their machines in a market that's willing to pay far more than they should.

The problem you get though is what I call the "California Cancer Warning Problem"
Basically, people can only pay attention to so many warnings. The more often people get false or trivial warnings, warnings where they have to continue to get things done as standard, the more likely they are to just plain ignore the warnings.

While hackers might be able to figure out a way to do something malicious without triggering the warning, the warnings back then were worse than useless, because they not only triggered for just about every document, users by default could not assess the document for safety without enabling the scripting. IE I couldn't by default open the document and look at the scripts to assess them (and some of them were only like a dozen lines) without enabling them.

Saying the warnings were necessary also ignores that there have been exploits that didn't even require opening a document to cause infection. Preview was enough.

Basically, if the hackers figured out something clever, just add that to the check. It would still be a better situation than what we had back then.

If you were correct, then this wouldn't apply to laptops and open-market RAM and storage. Consoles are not the concern...the fact that I have to pay a fuckton extra for SSDs and RAM and devices of all sort, even those not subsidized as a loss-leader, is the concern. I don't even want a PS5, but this does apply to everything I do want to buy in the next year that draws electricity.

This is a problem, well-documented, by the fact that the major AI vendors are buying all they can...it's a basic market problem. The only solution is to let the horrid bubble pop. My frustration is that they're subsidizing it in a circular manner and hiding from everyone how horrid their business model is.

LLM AIs will outlive this hype cycle and all the destruction they're causing. AI is here to stay. I just am sick of suffering because the major vendors are lying to investors and lying to their customers about both their business model and the capabilities of their technology.

Every time Hollywood sells a movie as 'realistic', it's turned out to be bullshit. The trade mags and entertainment reporters repeat the lie, but that doesn't make it true.

I'll be watching this movie soon, it looks fun. I will not expect them to get physics anywhere close to correct enough that someone with a decent high school physics class under their belt won't see where they got it wrong.

I was so proud of this robot that I went up and gave it a hug. Everyone around me started yelling at me to get away and then the robot tried to install a solar panel on me. Long story short, I ended up in the hospital with a restraining order against to stay away from the robot.

THIS ROBOT IS NOT FRIENDLY.

You probably can't see how because of the same vision problem that prevented you from reading my post.
"simulate materials without ever having to construct them."

Don't care. If your shit shows me ads, it's not getting into my kitchen. Note to self: Don't buy appliances from Samsung anymore.

Yes, I am vocal in how much I hate ads. I believe the CEOs of advertising companies should get one hit with a stick for every time their ad bothered someone even in the slightest.