It’s already happened in legal spaces…
They all individually turned to AI instead of fresh lawyers, fresh lawyers couldn’t get any experience…
And now firms are realizing there’s no new “middle” lawyers to hire to do the things AI can’t.
As an added bonus, AI prices keep going up. Just like with Uber, they intentionally lost money at first and planned to jack up rates when there was no competition to make profits.
The problem is AI still isn’t profitable and still can’t really fully replace the most junior personnel anywhere.
They fell for snake oil, they just haven’t fully realized it yet
Just like with Uber, they intentionally lost money at first and planned to jack up rates when there was no competition to make profits.
They’re intentionally still losing money because of the heavy investments to rush ahead in capabilities.
They’re intentionally still losing money because
Nvidia wouldn’t be funding them if AI wasn’t the biggest buyer of Nvidia chips…
It’s an Ouroboros, if they charged enough to be profitable that would kill demand which would kill data center rollout, which would kill Nvidia’s business model and then they wouldn’t fund AI companies and kill the AI company stock.price causing a sell off.
Becoming profitable would erase their business model, and all their personal wealth tied up in their own stocks.
We’d all be better off if we just paid them to hand briefcases full of money to each other instead.
The tech hit a plateau but they were in too deep already so they’re doubling down until they crash
Why are companies being allowed to engineer economies is the part of the narrative we need to be questioning.
Lemmy’s been saying that for a few years
Yeah, not really a new or surprising realization at all.
But Wall Street and execs need money NOW!
You don’t need to be an MIT AI expert to work this out ffs
I am an IT guy at a financial company, I have been put in charge of rolling out Copilot pro licenses to anyone who wants them, I have yet to roll one out for myself.
I learn far more by doing the work myself than to just ask an AI.
The most I use an AI for is like an advanced search engine.
If I am coding something in a web project, then I might ask it something like.
“I want to make a loading bar similar to the humanity theme from Ubuntu 9.04, how can I set up a custom style in CSS?”
I then check the sources and work from there.
I may not be as productive as other people using AI to a larger extent, but I have a better understanding of the end product.
To be perfectly frank, I find AI responses to be quite annoying and mostly just dumb, though it does get good sources.
For technical work, AI can definitely be helpful, but as a tool, not replacement. I work in software development and have used AI to tell me why test runs were running out of memory, after a few suggestions that didn’t work, it did give me the right answer (it was a wrong and unnecessary annotation). Also things like “what is the syntax for configuring a proxy with this-or-that command”.
The thing about technical work is that no matter if the solution is made by humans or AI, you still have to, and can, verify if it actually works. If you ask AI factual questions you genuinely don’t know the answer to and have no way of verifying, it’s your own fault if they are wrong.
