AI driven labor crisis?

JOhnH

Touchdown! Greaser!
Joined
May 20, 2009
Messages
15,163
Location
Florida
Display Name
Right Seater
Home Airfield
KOMN
From CNBC:
20,000 job cuts at Meta, Microsoft raise concern that AI-driven labor crisis is here

Is this just a temporary aberration as the economy evolves, as it always has in the past?
Or just "click bait"?
Or is it the beginning of a whole different paradigm?

“This represents a fundamental structural shift rather than a temporary market correction,” said Anthony Tuggle, an executive coach and leadership expert.

I'm retired so it affects me to a lesser extent, but it could still have widespread repercussions.

What do you think?
 
A little of column A and a little of column B IMO. Tech companies massively over hired during COVID and are still shedding staff here and there. Some of this is just shedding that labor and slapping the AI label on it to make the market happy.

AI is legitimately a strong productivity tool for those skilled and knowledgeable in their craft (especially software obviously) to use it properly. I think the mediocre software engineer cranking out boilerplate business logic code or the office drone doing mostly repetitive administrative work is highly at risk. AI is at that level in terms of capability if managed and overseen by someone decent. A good senior software engineer can now do with AI what it would’ve taken a few low level juniors overseen by the same person. This is of course somewhat an existential problem because if you fire all the juniors, in 15 years you won’t have any seniors to take up the mantel, but long term thinking is not exactly the market’s or corporation’s speciality.

On the other hand, we’ve seen some pretty spectacular failures of companies that tried to go all AI too fast. We saw the failure of builder.ai entirely, Klarna went all in and fired their entire customer service department to replace them with AI, only to have to re-hire most back when it didn’t work.

I use AI on a daily basis as a tool in software development. I’m not all in on the “AI will destroy all jobs in 5 years” but I do think there is a class of jobs that are quite at risk. There will always be a place for the upper quartile of skill and intelligence IMO, but I personally do think there is a legitimate risk to a section of white collar jobs. That doesn’t mean that some new career track won’t come around to make up for it (like we saw in software itself as we got more advanced tools from assembly to high level languages), the question is will this happen too fast for the people losing their jobs to adapt?
 
So the question in my mind is how you get to be skilled an knowledgeable when the entry level jobs go away. The last 25 years of my career was in consulting where the newbs did stuff that required a lot of time and helped you learn the ropes and the higher ups went toe to toe with the clients. Over time you moved from spreadsheet or power point mule to running teams and client relationships. You were credible in the latter because you had time in the trenches that taught you how to move to the next level.

If I remember the Foundation Series by Asimov, it foretold the result of what might happen with AI.
 
long term thinking is not exactly the market’s or corporation’s speciality
Yep. Only potential saving grace would be a degree in AI operation, so entry-level training becomes college education vs on the job training.
 
So the question in my mind is how you get to be skilled a knowledgeable when the entry level jobs go away.
Every CEO right now: “Sounds like a problem for the next guy, the market wants AI slop right now.”
 
Last edited:
Musk says in 25 years we’ll only work if we want to. AI and robotics will do everything, we’ll be on a form of UBI and cost of living will be a fraction of what it is now. One thing is for sure, we need to get on top of this and have guardrails in place to protect those displaced from the workforce.

 
One thing is for sure, we need to get on top of this and have guardrails in place to protect those displaced from the workforce.

why?

sure it would be nice, but must we have guardrails?

having typed that question, I suspect it will rapidly go full SZ.

There are places the concerns etc can be freely discussed, just not here, but we can't mention the name of the shall-not-be-named forum)
 
Meh. During the industrial revolution they said machines would do everything and man would luxuriate on the beach. Nope. And then it was computers would do everything, but still no beach. Now, AI will do everything. Nope, just like machines and computers, AI will greatly improve the productivity of those who learn how to leverage new tech. Those who don’t will be pushed to more menial positions.
 
why?

sure it would be nice, but must we have guardrails?

having typed that question, I suspect it will rapidly go full SZ.

There are places the concerns etc can be freely discussed, just not here, but we can't mention the name of the shall-not-be-named forum)
Well there’s nothing political here to worry about but just like he stated on the vid, AI is progressing faster than what we’re prepared for. The guys (Hinton, Altman, Musk, etc) behind the creation of AI have all warned about white collared jobs losses. It’s anyone’s guess as to the extent of AI taking over those jobs but I don’t see CEOs making decisions to keep people employed. They care about the bottom line and just like in the lawyer example, companies will choose what’s cheapest.
 
Last edited:
I can certainly see accounts payable clerks, engineering drafters, legal secretaries, etc. being replaced by AI simply because many of those tasks can be easily enough done by AI. The accounting world has plenty of positions ripe for AI to replace. The positions it won't be able to easily replace are the higher level positions (like a Controller) where the knowledge and critical thinking will be tough for AI to parse.

As much as people want to believe robots and AI will take over the world, I'd ask them to show me a robot that will climb under a sink and replace a faucet or garbage disposal. We aren't there yet, and I doubt we will be there in 20 years, much less 5.

Sent from my SM-S938U1 using Tapatalk
 
Stopped at a Wendy's today and successfully placed my order talking to a machine that was able to decipher what I was trying to say first time.
 
blue collar and white collar work will switch places.
AI can do the data analytics but it can't fix a toilet or run wire through your attic.

At some point a conversation about universal basic income will start to take the main stage.

I am doing things with Claude in hours that would have taken a team of developers a few weeks.
A shift is coming. But people probably said that about the internet when it first came out.
I was in college when it came out so I just used it to look at boobs.

Now AI will look at the boobs so we don't have to. Or something.
 
I am doing things with Claude in hours that would have taken a team of developers a few weeks.
Sure, but... if it's generating that many lines of code are you able to review it all in a reasonable amount of time and understand what it is actually doing (like, really understand what's happening and not just seeing satisfactory output)?

Much of the work of 'coding' is not the actual creation of the code. It's QC, security review, UAT, stress testing, etc... magnified by orders of magnitude if sensitive data is passing through.
 
AI can do the data analytics but it can't fix a toilet or run wire through your attic.
There has already been a trend towards Community Colleges and Trade Schools. For quite a while the push that you MUST have a college degree has been an issue. So I think we will see a lot more "trade" type education taking over as AI takes over the analytical or other "computer" type jobs. And yes, a degree in AI will be a good thing to have if you get a job before the 300% overkill also graduate.

"Hi, welcome to Joey's Burgers, I have a doctorate in AI, how can I help you?"
 
What I'm not seeing yet is AI taking the jobs of middle managers and leaders. I'd say it's easier to replace the job of a CTO or CIO with a few agents working in coordination. That, however, is going to take a while to happen (if it does) because those same people control the use of AI in their companies.
 
Musk says in 25 years we’ll only work if we want to.
Musk says a lot of things. His batting record on predicting how long technology will take to mature or when things will happen is not great.


Doesn’t help that his K-hole nowadays is deeper than the Mariana Trench…

All the big tech CEOs are frothing at the mouth about how capable and disruptive their models are/will be. Elon keeps saying that Grok is at PhD++ level of knowledge in every field. Actual reasoning tests, though, don’t seem to back that up. There are several interesting professional mathematicians on YouTube that routinely test these models, and while they sometimes can do pretty impressive things, they generally fall well short of the claim. There was another interesting video by a grad student at Cornell who TAs their intensive freshman CS class (for a freshman class it is amazingly intensive). He had 3 different frontier models in late 2025 try and take the class alongside his actual students. None of them did well in longer term work that builds on top of itself.

I am relatively confident the CEOs are right that it will be disruptive. How disruptive, the nature of that disruption and its impact on society, and the timeline of that disruption, however, is much less clear to me. In general I take most anything said to me by a salesman (and in the end, these CEOs are salesmen) with a massive grain of salt.
 
Might be time to bust out that old copy of Vonnegut's Player Piano.
 
Sure, but... if it's generating that many lines of code are you able to review it all in a reasonable amount of time and understand what it is actually doing (like, really understand what's happening and not just seeing satisfactory output)?

Much of the work of 'coding' is not the actual creation of the code. It's QC, security review, UAT, stress testing, etc... magnified by orders of magnitude if sensitive data is passing through.
AI can do that, too.
 
I have a frig magnet that says:

To Err is human
To really screw things up requires a computer.

I think it needs to updated to:

To Err is human
To really screw things up requires a computer.
And to completely insidiously destroy things requires AI
 
Tech companies massively over hired during COVID and are still shedding staff here and there.
I've heard this here and there. I don't have the numbers to prove it, but my understanding is that, even after the latest cuts at Meta/Facebook, their headcount is still higher than it was pre-COVID.

"We're doing massive layoffs because we're replacing people with AI" plays much better to Wall Street than "We're doing massive layoffs because we way over-hired."
 
I've heard this here and there. I don't have the numbers to prove it, but my understanding is that, even after the latest cuts at Meta/Facebook, their headcount is still higher than it was pre-COVID.

"We're doing massive layoffs because we're replacing people with AI" plays much better to Wall Street than "We're doing massive layoffs because we way over-hired."
if CEOs admit they messed up with hiring they look bad, the stock price takes a hit, and they only get 20 mil this year. Slap the AI label on it, suddenly they are a pioneer and innovator, stock price rockets, and they get 50 mil. Pretty easy calculus for them.
 
Remember when Mohammed Saeed al-Gorithm invented the internet and all the good tech jobs were outsourced? Yeah, me neither. I reckon it's gonna be like that.

It also reminds me of when Intel convinced the world of the inevitability of IA64 and all of the RISC players blinked first.

It's also a bit like the AlGoRiThM fad a few years ago; suddenly, everything is an algorithm! Duh. Now, suddenly, everything is machine learning. Crystal duh.

It is also a new turn on the old fire-and-motion FOMO cover fire. "All of the tech companies are doing this cool new thing! Don't you want to be the next unicorn?", which itself seems to be a variation of "if you drink this beer, you might score with attractive people, like in this beer commercial".

To be fair, LLMs do deserve a bit of a "holy ****!" moment. OTOH, LLMs are somewhat boring in the sense that they're just a program, bound by the limitations of any other Turing machine; unlike quantum computing, which actually moves the needle on what classes of problems can be solved.

Musk says in 25 years we’ll only work if we want to. AI and robotics will do everything, we’ll be on a form of UBI and cost of living will be a fraction of what it is now.

Yep, like we got full self driving by 2017 and Mars missions by 2026. I am skeptical of such "too cheap to meter" optimism; they (companies) will defend their rice bowl (people dependent upon their services). See also how utility companies are suddenly concerned about safety when it comes to low-cost plug-in solar ;)

When I was a lad and the internet was new, I was filled with the optimism of a world abstracted beyond borders or races; it seemed inevitable that a world of pure information, devoted to Truth and Reason and Progress and Freedom lie just around the corner. Look how that turned out ;) I think AI will turn out to be dual use the same way. I also think that in 25 years, we'll still be faxing medical records between doctors' offices with incompatible EMR platforms. Incompatibility makes money. Division makes money. Unhappiness and dissatisfaction makes money.

I think we're nearing the AI rug pull, where the addicts customers are on the hook, the prices go up, and the models get dumbed down to an 8th grade level like newspapers for all but the real paying customers.

tl;dr the smart will use AI as a tool when appropriate, the dumb will use AI to improve their Dunning-Kruger, and the tech companies will laugh all the way to the bank as it fuels the next couple decades of user engagement/intelligence/surveillance/disinformation/advertising/pornography that make the world go round. It should also provide some future political wedge issues like AI rights and AI marriage. It should also fuel software bloat as the kabuki security theater cat and mouse game continues to be played, and attack surfaces continue to grow instead of shrink, because it's good for business.

What's not to like?
 
If I remember the Foundation Series by Asimov, it foretold the result of what might happen with AI.
Fortunately I won't be around for the run-up to the Butlerian Jihad. I hope.
Stopped at a Wendy's today and successfully placed my order talking to a machine that was able to decipher what I was trying to say first time.
I had that experience at Taco John's the other day. The chipper little AI bot only took two tries to get it right. If only the resulting food had resembled something edible. They could automate the whole place and improve it substantially.
 
AI can do the data analytics but it can't fix a toilet or run wire through your attic.
Yet.

One of the things not being talked about is that as technology advances, there will be less of a need for skilled workers in technology. With advancments in robotics, it is probably within 20 years before they become competition to manual labor too, such as fixing toilets and run wires in your attic. And they'll be better at it than you and I are.

The math on UBI doesn't math. Fortunately, they will never replace pilots.
 
Cro-Magnon man used a rock and a pine wood fire. Early man used shaped hand tools, oak/maple hot wood fire. Bronze man used - well, bronze and a air-blown fire. And so on. Modern man uses an Oscilloscope and logic analyzer. Future man will use AI, and an advanced iPad, maybe with a visor/video glasses.

It's a tool, like every tool used since we started walking upright. This BS about world changing paradigm shifts because of a better mouse trap are all hogwash. I can think of 20 jobs right now that AI will never touch. If I gave it more thought, I could come up with plenty more. Long ago it was qualified as the 'displacement - replacement' syndrome.

How many buggy whips were made in 1900, vs 1920? Yabut - how many pneumatic tires, were made in the same timeframe? AI is just going to be an enhanced at-large multi-pronged Wiki tool.

Making predictions on labor market by using Meta as a reference is like measuring a rod journal with a tailor's soft tape.
 
Agree with just about all above, but I'd change the analogy for the leads of the tech companies a bit. I don't think they're salesman, I think they're the carnival barkers at the circus, in front of the "see the two headed whatever" trailer. And the people lining up are the investors. Because none of these tech stocks, at least as I see them, are based on performance but instead what someone thinks might happen X months from now.

The funny part to me is that AI is perhaps best suited to replace sales people. It speaks authoritatively about things that it can't comprehend, is prone to stretch the truth that it accidentally gets right, and the way things are going it's completely unaccountable for anything. The only thing holding it back that I see is that some tech people actually seem to find listening to a salesperson enjoyable if they're cute.
 
The funny part to me is that AI is perhaps best suited to replace sales people. It speaks authoritatively about things that it can't comprehend, is prone to stretch the truth that it accidentally gets right, and the way things are going it's completely unaccountable for anything. The only thing holding it back that I see is that some tech people actually seem to find listening to a salesperson enjoyable if they're cute.
Definitely sending that quote to my youngest son… a sales rep…
 
Because none of these tech stocks, at least as I see them, are based on performance but instead what someone thinks might happen X months from now.
I mean hasn’t speculation always been a pretty core part of the market? I agree that the stock performance is way outside the actual performance of the companies (none more so than the Musk companies IMO), but it’s not like speculation is a new feature. Correct me if I’m wrong, I don’t pay much attention to the financial industry.
 
The only thing holding it back that I see is that some tech people actually seem to find listening to a salesperson enjoyable if they're cute.

well, some tech people are such geeks/nerds that they could very well become enthralled by an AI avatar or fake persona
 
The funny part to me is that AI is perhaps best suited to replace sales people. It speaks authoritatively about things that it can't comprehend, is prone to stretch the truth that it accidentally gets right, and the way things are going it's completely unaccountable for anything. The only thing holding it back that I see is that some tech people actually seem to find listening to a salesperson enjoyable if they're cute.
The not-so-funny thing to me is that you just described politicians.
 
Musk says in 25 years we’ll only work if we want to. AI and robotics will do everything, we’ll be on a form of UBI and cost of living will be a fraction of what it is now. One thing is for sure, we need to get on top of this and have guardrails in place to protect those displaced from the workforce.
We already have a head start on UBI. There are government benefits for almost any of the necessary (and many not-so-necessary) things in life available for people that can't or won't work. I only see that as expanding to more benefits for more people as the need for manual labor shrinks and the production of food and electronics becomes more automated. I would modify Elon's statement to say that in 25 years only people that want more porridge will work for it. The number of people willing to accept "basic" living will accept UBI and be happy about it. Of course though, the standard of living will be more like a 3rd world country than what we have come to know. athough the gulf between the haves and have-nots will make today's society look like a homogeneous utopia.
 
Last edited:
Musk says in 25 years we’ll only work if we want to. AI and robotics will do everything, we’ll be on a form of UBI and cost of living will be a fraction of what it is now. One thing is for sure, we need to get on top of this and have guardrails in place to protect those displaced from the workforce.

Ask the joke goes, AI can't lay bricks, or unclog a drain.
Yet.
 
Back
Top