Post New

Sure, ChatGPT comes to your aid when you do your homework. Sure, it looks over the doctor's shoulders during a mammogram. It searches for better airfare.

https://fortune.com/2025/05/28/anthropic-ceo-warni...

Anthropic CEO Dario Amodei says AI could wipe out roughly 50% of all entry-level white-collar jobs. The cuts could come within five years, he says, causing unemployment to spike as high as 20%. He’s urging consumers and lawmakers to prepare now to protect the nation.

Amodei isn’t turning away from AI, it should be noted. Anthropic has just released its latest chatbot, Claude 4, and is bullish on the advances the technology can bring. He also said there is still time to mitigate the doomsday scenario by increasing public awareness and helping workers better understand how to utilize AI and navigate the transition.

Amodei is hardly the first person to warn about AI’s impact on the job market. LinkedIn’s chief economic opportunity officer, Aneesh Raman, earlier this month said artificial intelligence is increasingly threatening the types of jobs that historically have served as stepping stones for young workers. And there are a growing number of stories from workers who saw their six-figure jobs disappear without warning, bringing chaos to their lives.

And venture capitalist Kai-Fu Lee has called predictions that AI will displace 50% of jobs by 2027 “uncannily accurate.”

I have, for a few years now, talked about the dangers of AI to white collar jobs - the more analytical, the more vulnerable. Accountants, lawyers, engineers, computer programmers, yes, even doctors. If AI can allow a coder do the work of ten, what happens to the other nine students who majored in computer sciences? If CAD systems become intelligent, what happens to architects and interior designers? Based on the very beginning of using AI in military applications, huge changes could be expected (https://www.kyivpost.com/analysis/53546).

Who is safe? Electricians, plumbers, grave diggers - not menial jobs, but certainly manual jobs.

For generations (forever?), those who did not work were "unemployed" and held in low esteem by society. We are verging on an age where the word "unemployed" will have to be substituted by "excess leisure time" in order for our society to survive. The problem will be that the decrease in taxes collected will not be sufficient to accommodate this change. The obvious solution would be to tax AI CPU chips - with the obvious countermeasure to move them off-shore. (Although, realistically, since the major AI companies are already among the richest in the US, a cynic would expect them to follow in the footsteps of Musk and simply finance the elections of the powers that be).

The more dependent on an educated elite that a country is, the more vulnerable it is likely to be.

Get prepared (how, I'm not sure).

Jeff

More here:
https://www.axios.com/2025/05/28/ai-jobs-white-col...

There's a lively debate about when business shifts from traditional software to an agentic future. Few doubt it's coming fast. The common consensus: It'll hit gradually and then suddenly, perhaps next year. This could wipe out tens of millions of jobs in a very short period of time. Yes, past technological transformations wiped away a lot of jobs but, over the long span, created many and more new ones.

This could hold true with AI, too. What's different here is both the speed at which this AI transformation could hit, and the breadth of industries and individual jobs that will be profoundly affected.

You're starting to see even big, profitable companies pull back:

That's why Meta's Mark Zuckerberg and others have said that mid-level coders will be unnecessary soon, perhaps in this calendar year. Zuckerberg, in January, told Joe Rogan: "Probably in 2025, we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of mid-level engineer that you have at your company that can write code." He said this will eventually reduce the need for humans to do this work. Shortly after, Meta announced plans to shrink its workforce by 5%.

Microsoft is laying off 6,000 workers (about 3% of the company), many of them engineers.

Walmart is cutting 1,500 corporate jobs as part of simplifying operations in anticipation of the big shift ahead.

CrowdStrike, a Texas-based cybersecurity company, slashed 500 jobs or 5% of its workforce, citing "a market and technology inflection point, with AI reshaping every industry."

Aneesh Raman, chief economic opportunity officer at LinkedIn, warned in a New York Times op-ed (gift link) this month that AI is breaking "the bottom rungs of the career ladder — junior software developers ... junior paralegals and first-year law-firm associates "who once cut their teeth on document review" ... and young retail associates who are being supplanted by chatbots and other automated customer service tools.

And this is just the beginning.

Jeff

This could hold true with AI, too.

Maybe, maybe not.

Just read an article about the expected/anticipated failure of AI, based on how it was trained.

Initial AI was trained on massive amounts of uncontrolled inputs--which resulted in large volumes of irrational outputs (literally, non-existent material being created by the AI). The problem is simple: The newer/better (? LOL) versions of previous AI(s) were based on flawed/failed models. Thus, AI will fail of its own accord due to its inherent (built-in) known errors. It will become unusable because the AI is unable to differentiate between reality and its own irrationality.

Initial AI was trained on massive amounts of uncontrolled inputs--which resulted in large volumes of irrational outputs (literally, non-existent material being created by the AI). The problem is simple: The newer/better (? LOL) versions of previous AI(s) were based on flawed/failed models.

Sounds like what happens to a person who receives their education from Fox News (just kidding :-)

Let's assume (may, or may not be, true) that all AI's are not created equal and, possibly just as important, are not trained with the same information. Just as a person has to "weigh" new data to see if it is correct, there should be a way for an AI to also make this determination (possibly by the curating of the data it's being fed. One does not let our kids read anything they want and there is no reason to let our AIs be exposed to invalid information if it's avoidable.

I suspect the difference between a "computer" and an AI has more to do with pattern recognition more than any other factor. That coupled to their ability to learn "how to learn". Humans sometimes take drugs which cause them to hallucinate (and a few can hallucinate without the aid of drugs). They may not realize at the time that they are acting irrationally, but it is obvious to an outsider. That means that there is an opportunity to have a built-in "outsider" embedded in an AI as a governor to reel it back to reality if it drifts. (Maybe a specialized second AI?).

AIs are still in their infancy. Heck, they can't even drive an autonomous car - but, give them time.

I remember buying my first calculator, my first PC, my first fax machine, my first cell phone, etc., etc. Technology has leapt ahead in quantum leaps during my single lifespan. My wristwatch probably has more computing power than the mainframe I fed punch-cards to. The amount of resources being poured into AI leads me to believe that their advances will be exponential until mankind will be an inconvenience they have to put up with.

Jeff

Just read an article about the expected/anticipated failure of AI, based on how it was trained.
Initial AI was trained on massive amounts of uncontrolled inputs--...


I have certainly noticed two things.

First, LLM products believe what they read on web sites. They seem to make no discernible distinction between, say, what a product actually does and the advantages its PR flacks claim.

Second, it's amazing the rate of proliferation of web pages that are plainly AI slop. Not nefarious, just mid-level firms getting a bot to write their web site for them. Badly.

Combined, it shows that we are probably already seeing the LLMs being trained on (and believing) the crap spewed by their forebears. A meal that is 99% steak and 1% shit is still a pile of shit. (much like the enticing prospect of beef raised on chicken litter...)

Jim

all AI's are not created equal and, possibly just as important, are not trained with the same information.

True. But the AI has no way to identify which information being consumed is valid or not, or how it is to be analyzed. If an AI is devouring internet content, then it is nothing more than a reading and spewing machine. GIGO. It is not "AI", because there is no "intelligence" involved.

The ability to differentiate between real-world information and fiction is where the "intelligence" is required. So far, none of the AI models have that capability. Any AI producing false evidence is proof of failure of that AI model.

It seems to me that there are two possible outcomes with AI. It becomes very successful and displaces the workforce, at least in the short time frame,or it fails for many of the reasons mentioned. Both outcomes would be bad for the stockmarket in general, but positive for a few stocks if it becomes successful. If successful, significant deflation, if unsuccessful, significant write down of assets which would be deflationary to a degree.

Should I be considering deflation as the major economic trend over the time period 2 to 10 years in the future? Thinking that it will take 2 years to determine the outcome of AI.

Aussi

It seems to me that there are two possible outcomes with AI. It becomes very successful and displaces the workforce, at least in the short time frame,or it fails for many of the reasons mentioned.

It doesn't have to be binary. AI could be very successful and cause only a modest amount of overall job loss concentrated mostly in the specific sectors where it works best. That's often what happens with technology - robots revolutionized a lot of manufacturing, computers revolutionized a lot of information processing, and the internet as well. Yet there are still manufacturing/assembly jobs, still secretaries working in offices, still brick and mortar shops.

There's probably a very wide spectrum of possible outcomes, not just two.

I am not sure how it could be very successful and only displace a modest amount of workers. For me it then falls into the not successful category with large write offs (fiber optic cable in late 1990s). So I am still thinking binary with both outcomes deflationary.

Aussi

This could hold true with AI, too. What's different here is both the speed at which this AI transformation could hit, and the breadth of industries and individual jobs that will be profoundly affected.

I have little doubt there will be some significant repercussions, but I don’t think the pace will be anything like the doomsday predictions. Sure, some big corporations will tinker at the edges and lay off a few thousand people here or there. That doesn’t mean that they will be successful, and even if they are, it will take some time for it to trickle down to those corporations which don’t have those kinds of resources to oversee and evaluate.

“The internet changed everything” is a truism, but it didn’t all happen in 1995.

Initial AI was trained on massive amounts of uncontrolled inputs--which resulted in large volumes of irrational outputs (literally, non-existent material being created by the AI). The problem is simple: The newer/better (? LOL) versions of previous AI(s) were based on flawed/failed models. Thus, AI will fail of its own accord due to its inherent (built-in) known errors. It will become unusable because the AI is unable to differentiate between reality and its own irrationality.

I think this will be sector dependent. I’m sure there is a lot of crappy code that AI has been trained on, but it’s likely to be weeded out fairly quickly. On the other hand, the AI trained on a Twitter feed (sorry, X) is likely to be bad bad bad, and will not get better without severe recoding and other sources. I suspect that is why we’re beginning to see licensing deals for “good content” (New York Times, Getty Images, etc.) as opposed to any crazy thing that somebody puts on the web.

To stretch the point, you wouldn’t want an AI “doctor” trained on the content the rolls off Facebook groups.

So some of this is going to be “trial and error”, it’s gong to take time to sort that out, it’s going to take time for mid-level corporations to figure out how to use it, it’s going to take time for the effects to be felt - although there will surely be some quickly, both for good and ill.

Meanwhile “don’t worry, be happy”. Our generation had to deal with the disruption from OPEC, “supply side” and disco, not to mention the internet. Let the next generation worry about AI. It’s always something, after all.

Our generation had to deal with the disruption from OPEC, “supply side” and disco, not to mention the internet.

You get a rec just for including disco in the list. I rather enjoyed disco locations and music myself, but it did rather define an era : )

Jim

To stretch the point, you wouldn’t want an AI “doctor” trained on the content the rolls off Facebook groups.

And I wouldn't want a doctor trained bby watching General Hospital and Dr. Kildare.

People are acting as if every AI was a generalist trained on every bit of crap floating on the surface of the internet. Why train a hammer in how every type of tool is to be used - there's no reason to not to create "specialist" AI's.

BTW, A specialist is someone who learns more and more about less and less until he knows everything about nothing. A generalist learns less and less about more and more until he knows nothing about everything.

Just saying

Jeff

A real-time training on an AI for a specific use:
https://www.kyivpost.com/post/53784

In July the milblogger “Clash Report” said that Ukraine’s intelligence services were training artificial intelligence (AI) systems that would enable its drones to recognize enemy equipment, including Soviet-era bombers, using images obtained from military museums.

This data was used to identify the most vulnerable areas of the bombers from which to create AI algorithms that would allow weapons, in this case FPV drones, to independently recognize and engage targets. As videos captured during Sunday’s attacks attests the drones didn’t simply crash into their targets but went for the areas where maximum damage was most likely to be caused – weapons pylons carrying cruise missiles and over wing fuel tanks.

There are several stages to training AI to behave as it did during Sunday’s attack ---->

Jeff

Our generation had to deal with the disruption from OPEC, “supply side” and disco, not to mention the internet.

Did we ever really recover from disco though?

Post New

Berkshire Hathaway FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.