In the sense of making a list of who to deport, who to watch for “woke ideas” and might become a problem, who needs some religious purification. They Nazi loved records too, especially those about religion.

A lot of the improvement came from finding ways to make it bigger and more efficient. That is running into the inherent limits, so the real work with other models just started.

as they’ve been designed to

Well, designed is maybe too strong a term. It’s more like stumbling on something that works and expand from there. It’s all still build on the fundaments of the nonsense generator that was chatGPT 2.

wouldn’t buy that brand anymore so they would never do it again

Assuming there is perfect information in the market. In reality there is heavy information asymmetry.

It also assumes free competition while we have every market dominated by a few players buying up everyone else, often with cartel like behavior.

They did that to drive up short term ad revenue and it worked, and joy was in their greedy little hearts. They also did figure out that poisoning search results drives away users, and that search is kinda the fastest gateway drug to their entire ecosystem.

But they’re stuck. Fixing search would lower their ad revenue, and stock holders would kill them for that.

"past performance is not indicative of future results¨

Why connect to the Internet with it then

To check if your subscription to their models is still valid of course

Announced since 2014

AI trained on carmageddon

These articles also never mention how this is going to work in the fabric of a city.

Say optimistically a 15 second separation between vehicles, 4 passengers per vehicle, means one ¨sky-lane¨ can transport 60/15 * 60 * 4 => about 1000 people per lane. That´s worse than a car lane, and an order of magnitude behind bicycle lanes and subways. You can stack multiple lanes vertically, but then it becomes more fun with traffic control for take-off and landing and still a bicycle lane transports more people.

So far these things seem better as a city hopper, or used in a rural setting, but the current ones don´t have the range for that either.

Even in cyberpunk these flying cars are only for big corpos and the rich.

Much greater models

It’s Musk, so the new model is simply “the more money you have, the more you receive”

And it’ll be used to suppress wages, because “you’re not making new stuff, just fixing some problems in existing code.” That you have to rewrite most of it is conveniently not counted.

That’s at least what was tried with movie writers.

Then twins walk into the airport and the system crashes.

And facial recognition has been known to fail with very dark skin so that’ll turn into a racial issue real quick.

Amazon of the “just walk out” shops powered by AI, that turned out to be s lot of Indian workers in the background.

Popular media doesn’t really investigate other ways of organizing society. It’s always the standard job, shop and money with only rare exceptions. It just swaps dollars with credits, uses hovercars instead of normal cars and calls it a day. Hell, they were shoveling coal into a reactor in Rebel Moon to power their space ship. I don’t expect that much from entertainment and can live with the black mirror doom porn. Maybe I give writing my own stuff another try, but I’m not a good writer.

Honestly I’m with black mirror. These days when a new tech appears, the mind immediately wonders how this will be abused it obstructed by other parties. That’s not the show’s fault but of what keeps happening in real life.

Black Mirrors shows the future, extrapolated from current systems and events. It’s not about new paradigms that might be possible.

And it will say people of lower economic classes and foreign origin are the biggest risk. Will action be taken to help these people? Nah. This is just going to be a technological excuse for some good ol’ repression.

how you never have anyone who can actually “manage” the AI workers.

You just use other AI to manage those worker AI. Experiments do show that having different instances of AI/LLM, each with an assigned role like manager, designer, coding or quality checks, perform pretty good working together. But that was with small stuff. I haven’t seen anyone wiling to test with complex products.