Plus the dangerous long-term consequences attached to it.

Because there is almost nothing worthwhile about it. Environmentally, fiscally, and socially it's a disaster. For most purposes, it simply doesn't do what it is supposed to do, it is used as a pretext to offshore tech jobs, and it is only going to get smart enough to be dangerous, like members of the Trump Administration. And it creates a huge accountability crisis: if AI's "thoughts" can be put into action at the bot's own discretion, who's responsible for any subsequent harm? I would like it limited to use as a tool in science and medicine but without any agency. Just yesterday, I got an email from a friend who had to replace an entire AC unit because the main circuit board failed. I had "smart reply" in Gmail turned on by default and I've been amused by the AI suggestions for responses. But this time the AI created a completely fictional story about me having a similar issue with a printer two years ago. WTF!

and not true Gen AI, the better. These AI companies are selling mostly hype to buy time until they can produce actual AI sometime after we get (eventually) useful, working quantum computers.

For the average person, all it offers is the ability to make cool fake pictures.

For me, it makes way too many of my Google searches useless garbage, rendering it unreliable and thus a waste of my time. Just give me good links!! The 1 or 2 links AI gives to support its weird summaries are often absolute crap, and not reliable sources.

The wholesale adoption of AI for customer service chatbots is the worst. These chatbots NEVER solve the problem I am calling about. Utterly worthless. I always have to insist on a live agent to get what I need. So, they have not saved any labor for the company and have only annoyed the shit out of me and made me hate the company using the damn things.

Until they are "done" developing them.

They also don't know, and at the bottom of all of it is they don't really understand how they work anyway and from what I can tell, there is no way to understand them because of the way they work.

So they are just assuming some linear or predictable increase in utility when there is none.

But they keep shoveling resources at them and people are funding it out of planned ignorance because they will never know the truth.

Until it is known that they don't work and have no actual intelligence just regurgitation of what they have been trained with.

You can't even power these data centers in some cases for years because of that physics thing where you need to generate power and deliver it.

Which collides with the capitalism thing, where they believe there is enough money to be made and nobody is thinking about the power requirements dropping by orders of magnitude in the future, either from technology improvements or from a lack of need.

Deepseek made better software to do AI at a fraction of the CPU power of other models, that scared the industry, but they just continued the poor man's optimization which is to bury your software in faster and faster hardware rather than optimizing your software.

What happens to local ratepayers when they have no income from that 50GW nuclear power plant they spent billions to spin up because of data centers.... I'm sure it will just mean power costs drop to almost zero because of lack of demand.... no, they still need to pay for the infrastructure they built so guess who gets the bill.

I see a large pin on the horizon ready to burst this bubble and the next Great Depression...