I mostly use the Google search AI
(the default Google results behavior,
the Summary and Dive Deeper).
It says all kinds of crazy shit and even
contradicts itself in the same sentence.
It gives me broken wrong shit about 60% of the time.
When it says something correct, it turns out to be a
literal verbatim copy of what Wikipedia.

YMMV.

And these LLMs are not going to improve.
This whole craze is going to be a very drawn
out and painful lesson for the humans.
For example, they have you sold.

People who want quick answers.
And don't care if the answer is wrong or not.

People who want ideas.
And can't think of any themselves.

People who struggle with proper tone or form.
And need to copy an answer off someone/something, and don't care if it's the correct answer.

People who are busy.
And don't care if they get the correct answer.

People who don't like to read manuals.
Can't be bothered to read or think and don't care if they get the correct answer.

People who are buried in make-work at the office.
And by definition don't care if they get correct answers.
Because their superiors who invented the make-work for them to do also don't care if they get the right answers. But if the AI gets a tiny biy better, those make-work people and their managers will be fired and have no means by which to support themselves. If I were in that category, I would not be singing the praises of AI right now.

There are a host of other kinds of people who want AI. I'm in several of these categries.
People who rely on computers to spell "categories" but can't even be bothered to press the button to automatically fix their mistake when it has been outlined by the computer in red ink.

None of the above is new, just that now there's a machine to "help". It's amazing the world works at all, isn't it? But now we have AI which can propagate and perpetuate the above qualities. Is that a "good" technology? I don't see any answer to the upcoming existential crisis where at least a third of the population is unemployable. What will happen to the huge underclass?

Bend Over.
You sound like you need a good fishing.

(a) I did that fine previously without AI
(b) Nobody is following any of the links that supposedly support the conclusions of the AI;
nobody is reading any source material,
they just believe whatever the AI says
(c) The source material is quickly turning
into AI-generated slop, such that
(d) Humans can no longer access original, correct information sources. It is becoming impossible.

How is that better?

I am sure that ChatGPT would confirm to you that ChatGPT (aka A.I. Gore) invented the Internet....

https://www.hoover.org/researc...

Face scan everyone who enters the store.
The police can match them by driver's license.
(Not to mention all the other scan/ID stuff the store can provide in addition to the entrance ID and activity record).

https://www.youtube.com/watch?...

OK. You're not crazy.
Just calm down.
Everything will be fine.
We have someone on the phone who would like to talk to you.

Oh, I've seen that!

Isn't this in San Francisco, where they refuse to prosecute anyone for theft under $800 each time? And where for some strange reason, there are suddenly thousands of incidents of people walking out of grocery, convenience, pharmacy, electronics stores, department stores, and everything else? Sometimes in gangs of rampaging youths, and more often individuals? Casually wheeling out all manner of merchandise, all day long, because it's a lawless city and a free-for-all?

Maybe the problem is not with the doors...

I bet there are customers who are armed, as a matter of their own personal policy. When I go to the store, I am always carrying. Based on the local police reports where I live, the petty criminals who would steal shit at the grocery store are also usually armed (but illegally).

Not sure why you want armed guards at Safeway. No amount of groceries is worth anyone's life. And legally, Safeway cannot stop, search, or impede you -- certainly cannot lock you in -- unless they have probable cause. (Not just "reasonable articulable suspicions", but the higher legal burden.) For example, if they observed you concealing merchandise under your clothing. But are they going to risk violence over this? And if they are mistaken, and attempt to detain you, they are going to be facing civil and probably criminal charges. Their felonies will include: false imprisonment, assault, and battery, for starters. Shopkeeper laws are not going to shield them from that. And now you want them to pull a gun?

A better idea might be to have security cameras, and possibly follow them discretely out the door and snap some photos of them and their car plates. But even doing that might be too risky. What if they pull a gun because you are following them? You want to start a gun battle in front of your store? Over some tomatoes and beans? Just because they might be crazy, does that mean it's worth dying over? (Not to mention maybe killing bystanders.) You're nuts.

You could just call the police, and give them the security camera footage of them stealing, leaving the store, and maybe even getting in their car and driving away. You might even snap the license plate from a safe distance without them knowing you are doing it.

The police can do facial recognition on them, if their faces were caught anywhere along the way (assuming they have a driver's license or something). If they are wearing a hoodie and a mask, that's a problem.

Armed guards are very expensive, and a liability. You can be sure that the store has done the math.

If it's that bad in the location, maybe you should lock the ENTRANCE, and not let in sketchy people. Make them ID or face-scan in order to get in the store. You could even automate that.

Amazon scans everybody who walks into their grocery stores. And in the fancy stores (if they have any of those experiments still going), your every move is tracked by video-AI.

https://www.youtube.com/watch?...

Nobody knows what kind of school it was, or what it taught to its 30 Gifted Students. We know that the school did not close - it moved. Nobody knows where it moved to.

We don't know why the school upset the neighbors, or what things they saw happening over there in the neighborhood. Was it VTOL jets making noise? Magnetic anomalies?

If I had to guess, they moved it from the Zuckerberg estate in Palo Alto to someplace in Westchester County, New York.

The funny thing is, it's not the "X" (formerly Twitter) school, it's the "Meta(human)" (formerly Facebook) school.

The two spends from the gaztrillion investments are the production of the AI chips, and the clouds (and power plants) to run them. Maybe the answer will be to repurpose those for something that (a) works and (b) is profitable. Besides the bogus LLM crap that is driving all this nonsense, there are actual AI applications that can use that infrastructure. Maybe there will be enough of that to take over. Although one wonders why, if that could be true, it isn't what has created this situation. Maybe it's just the hype factor.

They may have to call it something other than 'AI". A lot of it was called ML until recently.

I started in AI in 1980 and remember the first AI Hype cycle, which was followed by AI Winter. We had to rebrand things and deny that any AI was involved -- even though we were doing the exact same technology. (Because it worked. It just didn't live up to the hype that marketers and investors put on it.)

In general, one would get fired for doing anything "AI" or bringing in any products or technology associated with "AI". There was no worse thing you could say than the word "AI". And we won't even talk about starting up and getting investment in an "AI" company.

All the AI-powered companies (using a variety of "AI" technologies) suddenly claimed that they were not using AI. No Siree, There's No AI in here at all!

One mundane but funny example is one that overlapped with the .COM Hype that came right after. We were developing a consumer facing web site, and programmed it in Lisp. We did not want anyone to know it was in Lisp. (Really for two reasons. First, Lisp was bad/evil because it was heavily associated with the AI Hype. Second, because it was our secret weapon and we didn't want competitors to know.) So what we did is this. We hacked the outermost last resort error handling code, which would capture all errors and give the user a "Sorry, broken" error message. In the logs we would see the real error, backtrace, and everything. What the user would see was a generic Java error message (a NullPointerException). Didn't matter what the real error was - the actual error had nothing to do with any null pointer, and there was no Java anywhere in the entire stack.

I am the first poster, and I do not know what an "rsilvergun" is. (Maybe I am misreading the syntax, and you are responding to and ddressing "rsilvergun", rather than identifying eir as the "first poster"?) Or perhaps you mean the second poster, who brought up banks, the subject which has dominated the thread.

If you are indeed referring to my post, I guess it's nice to know there is now one person who agrees with me....

Translation:"AI is too big to fail, and we are already planning on bailing out those trillionaires."