1. Replace workers with AI.
2. Now instead of wages you pay a company for AI services.
3. Despite the likely decline in quality of the work, suppose you become dependent as a company on this service. Suppose you make it work.
4. AI is heavily subsidized by venture capital, its priced lower than the cost to provide the service to attract early adopters (and to lock companies like you in.)
5. Inevitably the AI bubble bursts, AI services jack up their prices.
How is paying rent better than wages?
That's the plan.
Fire everybody.
Use genAI to produce mediocre results for less money.
When the price goes up, "fire" the genAI and rehire the desperate former workers who never found another job for even less.
@jrdepriest @futurebird A lot of people do find other jobs though.
I think Hal Pomeranz a bit further down has the right of it. They simply aren't thinking long term. Tomorrow does not exist. Only today does. Profits today, burn down tomorrow.
There is no plan. It's all "move fast and break things" without consideration of the fact that one of the things they're breaking is the boat that keeps them above water.
@nazokiyoubinbou @jrdepriest @futurebird
I would argue that introduction of a new technology at a company simply takes time.
(Not saying that AI stuff makes sense.)
But a lot of companies first have to introduce a technology to see that it doesn't bring benefits.
Of course, they could have properly tested the technology beforehand. But nobody does that, especially not for software.
"Let's introduce it and then figure out how to use it later".
Especially when the business magazines tell you that AI is the next big thing.
@wakame @jrdepriest @futurebird Meanwhile the Fediverse celebrates "push to prod Friday" as a meme joke for a reason.
"Move fast and break things" admits inherently that stuff will break. It assumes that only haste matters no matter what the cost. But it admits there will be a cost.
I think the running scam of pretending that LLMs are AI is proof that it was always a false belief. When that bubble pops (likely soon now!) those companies investing so heavily in it are going to be hit *HARD*.
@nazokiyoubinbou @jrdepriest @futurebird
For stuff like customer service or press statements, LLMs are "good enough".
As in: "We don't want to pay for it anyway, so quality isn't an issue, only costs."
(Well, until they invent cash-back guarantees and other stuff that costs the company money.)
@futurebird It's also seen as a way to shift blame when something goes wrong. I don't think that this is a primary motivation, but I would say that it's more prevalent than one would think.
These landlords want to be seen as "responsible owners" when things are going well but as soon as something goes poorly, they want someone/thing else to yell and scream at (whom they don't have to look in the face) to get things going. It's about being able to shift the blame and causality away from themselves when something goes wrong. And, to be frank, most business types are too dumb to realize that they've been sold on a promise of "new efficiencies" that will never be delivered.
What is that old saying, "A poor workman blames his tools?" I believe that this is one of the main ideas driving this "AI" adoption in business - what these people really want are tools for which they can take credit but then also blame as being "faulty" when their own ineptitude causes problems.
That's all that this whole "AI" bubble is about - they don't want to augment or improve human capability or even efficiency, they want automated slaves to do their bidding 24/7/365, and screw the actual human workers. Having a workforce that won't talk back and demand rights is more valuable to them than hitching their wagons to the providers of those tools.
It has never been about long-term efficiency or sustainability. It has always been about short-term, quarter-to-quarter profit and being able to get out with their money before that rent really comes due.
@futurebird Darned good points. (Except LLMs are not AI and it's not ok to let them keep calling them such.)
A company that has workers who are trained and experienced in working with its needs has greater independence. By relying on another company to do everything for them, they actually not only lose that independence, but they risk that the other company now has the means to simply replace them...
Actually, I believe I have heard examples of this happening in the past (albeit not LLMs.)
@futurebird
I have never met a C-suite executive who could think more than a quarter ahead.
You're assuming that those executives will be around for more than a quarter, and won't have cashed out and are sipping margaritas by their Olympic-sized pool.
Automated checkout was supposed to reduce the number of cashiers needed at stores. But now they need a person to fix the machines and people to help the people using the self-checkout... meaning it didn't really reduce labor costs much at all.
Things like checking ID, reversing charges all require an employee.
And in most stores they still need to have some regular checkout since some customers just CANNOT do it.
@virtualinanity @futurebird I literally double-scanned an item by accident at one of those & I couldn't delete it without an employee. They simultaneously want to trust customers to self checkout but don't trust them at all.
Not to mention the plexiglass corrals at the entrance & exits now to reduce theft but just makes the place feel un-welcoming. 🤦♀️
@kentenmakto @futurebird Good quote.... Puts me in mind of a great movie, Chaplin's anti-fascist "The Great Dictator" :
"....You are not machines, you are not cattle, you are men! You, the people, have the power to make this life free and beautiful, to make this life a wonderful adventure. Let us use that power. Let us all unite!..." (worth reading the whole speech).
..also used to good effect in Paolo Nutini's "Iron Sky"
Even if, for the sake of argument, you ignore the real cost of LLM compute and assume the bubble won't burst -- your whole business has now changed to be just reselling LLM compute: You've become a middle-man that adds little to no value and only exists until your customers realizes they can just buy directly from the source, or your LLM provider decides they'd rather keep your profit margin for themselves.
@futurebird management has been aggressively pushing to opex vs. capex for at least the last 30 years.. from most employees being contractors, to magic beans, it's all the same
short term gains with terrible long-term prospects. but they've learned they won't be held accountable. at worst they might have to change jobs and probably end up making more.
Furthermore:
Tech companies who stop using traditional means and, instead, move to using LLM services have given up their ability to do work over to the LLM vendors.
Once the subsidy from VCs ends the price will go up drastically - what choice will they have but pay it? They laid off all of their workers who knew how to make things happen.
fmc01
myrmepropagandist
J. R. DePriest 
:EA DATA. SF:
Nazo
Mauser II
Enola Knezevic
Kévin ⏚
Tor Iver Wilhelmsen
Jim Infantino
LovesTha🥧
Rodney
Kobold
XCM
sollat
➴➴➴Æ🜔Ɲ.Ƈꭚ⍴𝔥єɼ👩🏻💻
Jefflon "SynAck" Appleberg 
Hal Pomeranz
Matt T.
canleaf08 ⌘ ✅
Raven667
Chip Unicorn
elizabeth worm🔅
Steven Hoefer
wasabi brain
Sherri W (SyntaxSeed)
small, primitive, wingless
Patrick Morris Miller
Graeme 🏴
Suzanne she/her
Skjeggtroll
Not🐧A🐧Convicted🐧Felon
Dan
Kofi Loves Efia 
Rpsu (326 ppm)
Dean F.
Light
cy
Lorgo Numputz
JWcph, Radicalized By Decency
hex