The CardBoard
AI threatens livelihood of Stanford computer science grads - Printable Version

+- The CardBoard (https://thecardboard.org/board)
+-- Forum: C-House! (https://thecardboard.org/board/forum-4.html)
+--- Forum: The CardBoard (https://thecardboard.org/board/forum-5.html)
+--- Thread: AI threatens livelihood of Stanford computer science grads (/thread-25947.html)

Pages: 1 2

AI threatens livelihood of Stanford computer science grads - Mick - 12-19-2025

I imagine this is not unfamiliar to CardBoard members, but this LA Times article provides some depth on the effects of AI on the recent grad job market.

Anthropic's (Claude) CEO Dario Amodei says that AI's capabilities will wipe out 50% of entry-level white-collar jobs in five years. 

They graduated from Stanford. Due to AI, they can't find a job - Los Angeles Times

RE: AI threatens livelihood of Stanford computer science grads - TrumpCard - 12-19-2025

(12-19-2025, 07:15 AM)Mick Wrote:  I imagine this is not unfamiliar to CardBoard members, but this LA Times article provides some depth on the effects of AI on the recent grad job market.

Anthropic's (Claude) CEO Dario Amodei says that AI's capabilities will wipe out 50% of entry-level white-collar jobs in five years. 

They graduated from Stanford. Due to AI, they can't find a job - Los Angeles Times

AI is influencing other sectors as well, also with mixed results.  

In law firm practice, AI is excellent at some routine tasks.  You can feed a complaint into AI software, and get a lovely set of document demands and interrogatories to get information about the plaintiff's contentions.  At this stage, however, AI is bad at brief writing.  It will cite fake cases or give blatantly incorrect citations/quotations to case law.  Lawyers have to check whatever comes out of AI very diligently or they will suffer the consequences.  I know two lawyers who have been ripped to shreds by courts for using AI to write briefs without sufficient safeguards to ensure accuracy of citations/quotations.

My wife is a math teacher, and she agrees that AI has value, but it is still raw.  If she's already written a test, she can ask AI to write a make-up exam for kids who missed the first one, and it's really good.  AI can even write the first test reasonably well after reviewing the lesson plans from the unit.  But AI is terrible at answer keys.  She can ask for an answer key, get a result, and then, if she asks AI to make sure the key is correct, it'll sheepishly admit to errors, and then admit to even more errors on a second reminder to check accuracy.  I don't really understand that.  I thought computers are supposed to be great at math...

RE: AI threatens livelihood of Stanford computer science grads - dabigv13 - 12-19-2025

The AI CEO (any of them) is motivated to say things like "50% of white collar jobs will be eliminated by AI" because that is the whole value proposition for AI. AI companies aren't making money on people making their own little Ghibli style images, they make money if other business buy their stuff so they can fire their workers or hire less of them.

Maybe the CEO's are right, and I certainly believe there will be some job impacts. But I also think there is a lot of self-serving hype in the AI space too.

AI's are bad at math because they are language models, not computational models. It's just using prediction based on the words you enter, not actually doing any math.

AI is weird because it's good at things computers aren't supposed to be good at, and bad at things they are supposed to be good at.

RE: AI threatens livelihood of Stanford computer science grads - 82lsju - 12-19-2025

(12-19-2025, 11:50 AM)dabigv13 Wrote:  The AI CEO (any of them) is motivated to say things like "50% of white collar jobs will be eliminated by AI" because that is the whole value proposition for AI. AI companies aren't making money on people making their own little Ghibli style images, they make money if other business buy their stuff so they can fire their workers or hire less of them.

Maybe the CEO's are right, and I certainly believe there will be some job impacts. But I also think there is a lot of self-serving hype in the AI space too.

AI's are bad at math because they are language models, not computational models. It's just using prediction based on the words you enter, not actually doing any math.

AI is weird because it's good at things computers aren't supposed to be good at, and bad at things they are supposed to be good at.

I wonder about when will the hyperscalers generate enough revenue to justify the capital expenditures (each spending ~$100 billion/year) they are making in "AI data centers", and what will the be the useful life of the servers in those datacenters.....

https://www.srgresearch.com/articles/hyperscale-spending-spree-is-driving-dramatic-growth-in-data-center-capacity

RE: AI threatens livelihood of Stanford computer science grads - Goose - 12-19-2025

(12-19-2025, 07:15 AM)Mick Wrote:  I imagine this is not unfamiliar to CardBoard members, but this LA Times article provides some depth on the effects of AI on the recent grad job market.

Anthropic's (Claude) CEO Dario Amodei says that AI's capabilities will wipe out 50% of entry-level white-collar jobs in five years. 

They graduated from Stanford. Due to AI, they can't find a job - Los Angeles Times
Actually, I have a different take on "what the real issue is (and has been)". Computer Science majors were not originally envisioned to be "coders". They were expected to know about the theory and the "why" of computer software. Numerical analysis, for example, is a "computer science" thing. Unfortunately the degree has degenerated into being "coders". The business community treats software as a product where a few really smart people "design" the software and a large number of much less smart people implement (code) the software. IMO this is a really bad model, but that is what business/government wants to do. The result is that software projects often take much longer than expected, don't actually work well/fit their intended purpose and cost lots more money than anticipated. They are often cancelled before completion, and if completed contain orders of magnitude more "code" that is really required. The "designed by geniuses to be implemented by drones" approach has real problems, but that is what "computer science" has become because that is what the model wants. Years ago, when a UCB software graduate actually said to me "Just tell me what to do and I will do it" I knew for certain the computer software community was in real trouble :-). So yes, AI can be seen as useful if the implementers are modeled as drones and knowing what you are actually doing and why you are doing it is not required (or even desired). Of course, that is how Boeing got their MCAS problem, so YMMV.

RE: AI threatens livelihood of Stanford computer science grads - BostonCard - 12-19-2025

Pasted from a post I made internally back in June when Amodei first started making this prediction.

So, I’ve been thinking about the interview a lot of late and while I may be one of the people completely sticking my head in the sand on this, I think Amodei has fallen prey to a few cognitive traps here.

First, extrapolation of recent progress.  What has happened in the last five years in the AI space is nothing short of astounding.  However, extrapolating that level of progress going forward is not certain, and innovation tends to happen in fits and starts, with periods of revolutionary change followed by long periods of evolutionary change.  Think of the sea change when the internal combustion engine became available, and comparison to the previous technology (horses).  However, in the ensuing 150 years since Otto first developed the Otto Cycle Engine, progress has been extensive but much more about refining and improving the engine than evolutionary changes.  While I am not an expert in the field, I get the sense that since GPT-3 came out five years ago, subsequent improvements have been more of the evolutionary variety than continued revolution and it appears that progress may be slowing, wih computational limits and training data availability becoming bottlenecks.

Two, it is often said that the future is already here, it’s just unevenly distributed.  I think Amodei mistakes capability with deployment.  In other words, he looks at his models and says “AI can already do [X} and will soon be able to do [Y}“, but it is a very long way from knowing that the technology could do it to the technology actually doing it.  Companies deploying AI will need time to develop their requirements, chose amongst competing systems, implement pilot programs to test them, get them adopted by users, all of which tends to happen pretty slowly, especially in large companies with legacy systems.  As I mentioned, it was five years ago that the paper describing GPT-3 was published, but the introduction of [our internal AI system] and the roll out of Gemini has only recently happened.  And while the [internal communications channel] is really a hive of activity and it is great to see people experiment and deploy it for their use cases, I think we are a very long way from the type of activity that would make 20% of employees redundant.  To give a good example of the diference between capability and deployment, consider the electronic switchboard and telephone operators.  This article [https://www.richmondfed.org/publications/research/econ_focus/2019/q4/economic_history] is quite interesting.  It took only 16 years after Alexander Graham Bell patented the telephone before an automatic switchboard was first deployed.  18 years later, in 1910, only 300,000 subscribers out of 11 million had automatic service, and employment in the number of operators continued to increase from around 178,000 in 1920 to about 342,000 in the middle of the century, and was still 250,000 in 1960, 68 years(!) after the first electronic switchboard was deployed.

Lastly, and I think the biggest fallacy that Amodei falls prey to is the lump of labor hypothesis.  That is the idea that there are a number of fixed tasks to be done, so that if you replace them (for example, outsourcing to another country, using immigrant labor, or automating them) that the inevitable result is unemployment.  Time and again, this fallacy has been shown to be incorrect.  Not only do these improvements historically not lead to increased unemployment, improved productivity is what drives employee wages.  There is another good example in history, the development of the ATM.  It turns out that while many people predicted that the widespread deployment of ATMs would lead to fewer jobs for bank tellers that they were replacing, paradoxically employment of bank tellers actually increased, as shown below.

[Image: https%3A%2F%2Fsubstack-post-media.s3.ama...3x491.jpeg]

Why?  Well, while the number of tellers per branch went down substantially, the number of bank branches expanded dramatically, because now it was cheaper to run a branch than it was prior to ATMs.  The nature of the job changed a bit, and tellers were relieved of routine stuff such as depositing checks and withdrawing cash (which was now done by ATMs), and were able to be used to build relationships and open accounts, activities that were much more valuable to the banks.

What does it mean for a company like [my company]? Let’s imagine that due to AI all of us can all of a sudden be 20% more efficient in what we do.  I suppose one thing that [the company] could do is lay 20% of the company off and return the money to shareholders.  But there is not a fixed pot of drugs to be developed, and an alternate use of the 20% of excess productivity is to work on 20% more programs.  With a 20% improvement in efficiency, suddenly some programs that may not have been viable to pursue because aNPV was marginal, might all of a sudden be viable because it requires 20% fewer resources to pursue.  Maybe we pursue an additional indication for a drug already in our portfolio in parallel instead of gating it.  The point is that if we were 20% more efficient, it might not mean that 20% of us are let go, but rather that we have a chance to work on 20% more interesting programs.  In fact, as with the bank teller example, if we are 20% more efficient overall, the cost/benefit balance of adding employees becomes favorable, since their extra productivity may make taking on the extra FTE to be worth it.  Now, of course, it is never that simple, and AI will have variable effects on productivity, meaning that the effect will be unevenly distributed, which means there will be different productivity bottlenecks over time, so the relative job composition in the company will probably change.  I hope that impact is minimal, and I certainly hope that my job is made more productive rather than redundant by AI, but it is a fine line between the two.

Apologies for the rambling response; I’m curious what other people think.  Do I have my head in the sand about AI?  Am I too pessimistic about the pace of deployment?


BC

RE: AI threatens livelihood of Stanford computer science grads - BillBradley - 12-19-2025

I think similarly about this topic. I don't think anyone can even fathom AI's net effect on the economy, jobs, types of jobs, new products and other innovations. I just hope that we're smart enough as a society to put up the right guardrails along the way. 

I can recall an article published by Time Magazine many years ago about the effects on society of the washing machine. I can't find it online, so I'll attempt to summarize in hopes that I remembered it correctly :> The electric washing machine was introduced in the early 1900's. At that time, people owned a few "everyday outfits". Washing was done by hand a few times/week and was very laborious - buckets of water/soap, scrubbing boards, rinsing, hanging to dry. Clothes were thicker and made to last. When the electric washing machine was introduced, not everyone could afford one. Hence, the laundromat was born. Because of the efficiency gains, much less time was spent on laundry in the next few decades. However, a study was done in the 80's showing that the amount of time spent on laundry had actually increased compared to the early 1900's. The reason was that people recognized the decrease in "clothes maintenance" time, so they bought more clothes. Different types of clothing material (e.g. polyester) gave birth to different styles and price points. The fashion industry took off. And then houses were designed with walk-in closets. What was initially an invention to simply save time doing laundry actually became a catalyst for an entire new industry of clothing companies and other innovations. That story has always stuck with me whenever I read about things like these predictions that AI will obliterate our workforce.

RE: AI threatens livelihood of Stanford computer science grads - Bruce Wang - 12-19-2025

(12-19-2025, 01:54 PM)BillBradley Wrote:  people owned a few "everyday outfits". Washing was done by hand a few times/week and was very laborious - buckets of water/soap, scrubbing boards, rinsing, hanging to dry. Clothes were thicker and made to last. When the electric washing machine was introduced, not everyone could

This makes more sense than my prior supposition which was that people put up with smellier clothes. 

I learned the hard way once that the washing machine is the most valuable appliance in the house.  Everything else has a decent short term substitute.

RE: AI threatens livelihood of Stanford computer science grads - MV72018 - 12-19-2025

(12-19-2025, 01:54 PM)BillBradley Wrote:  I think similarly about this topic. I don't think anyone can even fathom AI's net effect on the economy, jobs, types of jobs, new products and other innovations. I just hope that we're smart enough as a society to put up the right guardrails along the way. 

I can recall an article published by Time Magazine many years ago about the effects on society of the washing machine. I can't find it online, so I'll attempt to summarize in hopes that I remembered it correctly :> The electric washing machine was introduced in the early 1900's. At that time, people owned a few "everyday outfits". Washing was done by hand a few times/week and was very laborious - buckets of water/soap, scrubbing boards, rinsing, hanging to dry. Clothes were thicker and made to last. When the electric washing machine was introduced, not everyone could afford one. Hence, the laundromat was born. Because of the efficiency gains, much less time was spent on laundry in the next few decades. However, a study was done in the 80's showing that the amount of time spent on laundry had actually increased compared to the early 1900's. The reason was that people recognized the decrease in "clothes maintenance" time, so they bought more clothes. Different types of clothing material (e.g. polyester) gave birth to different styles and price points. The fashion industry took off. And then houses were designed with walk-in closets. What was initially an invention to simply save time doing laundry actually became a catalyst for an entire new industry of clothing companies and other innovations. That story has always stuck with me whenever I read about things like these predictions that AI will obliterate our workforce.

Re your sentence, "I just hope that we're smart enough as a society to put up the right guardrails along the way", as long as "Mr. Anti-Guardrails" controls the government and seeks to ingratiate himself to deep-pockets Silicon Valley AI leaders, then I believe your "hope" is little more than a pipedream. The dominant approach that our society takes to potentially problematic new technologies in this society I call the "bandaid" approach. The massive assumption is made that we should let the technology development and diffusion process proceed unfettered since, it is held, if anything goes wrong downstream in terms of its effects, we can always come up with a suitable "bandaid"/fix in time to stop/prevent/substantially mitigate or reverse those disturbing effects. I regard that assumption as naive since by the time the significant problematic effects reveal themselves to a critical mass of  people, the technology will likely be so entrenched in society and so much will have been invested in it that it will be effectively impossible to stop or brake it.

RE: AI threatens livelihood of Stanford computer science grads - Mick - 12-19-2025

(12-19-2025, 12:27 PM)BostonCard Wrote:  There is another good example in history, the development of the ATM.  It turns out that while many people predicted that the widespread deployment of ATMs would lead to fewer jobs for bank tellers that they were replacing, paradoxically employment of bank tellers actually increased, as shown below.

[Image: https%3A%2F%2Fsubstack-post-media.s3.ama...3x491.jpeg]

Why?  Well, while the number of tellers per branch went down substantially, the number of bank branches expanded dramatically, because now it was cheaper to run a branch than it was prior to ATMs.  The nature of the job changed a bit, and tellers were relieved of routine stuff such as depositing checks and withdrawing cash (which was now done by ATMs), and were able to be used to build relationships and open accounts, activities that were much more valuable to the banks.

What does it mean for a company like [my company]? Let’s imagine that due to AI all of us can all of a sudden be 20% more efficient in what we do.  I suppose one thing that [the company] could do is lay 20% of the company off and return the money to shareholders.  But there is not a fixed pot of drugs to be developed, and an alternate use of the 20% of excess productivity is to work on 20% more programs.  With a 20% improvement in efficiency, suddenly some programs that may not have been viable to pursue because aNPV was marginal, might all of a sudden be viable because it requires 20% fewer resources to pursue.  Maybe we pursue an additional indication for a drug already in our portfolio in parallel instead of gating it.  The point is that if we were 20% more efficient, it might not mean that 20% of us are let go, but rather that we have a chance to work on 20% more interesting programs.  In fact, as with the bank teller example, if we are 20% more efficient overall, the cost/benefit balance of adding employees becomes favorable, since their extra productivity may make taking on the extra FTE to be worth it.  Now, of course, it is never that simple, and AI will have variable effects on productivity, meaning that the effect will be unevenly distributed, which means there will be different productivity bottlenecks over time, so the relative job composition in the company will probably change.  I hope that impact is minimal, and I certainly hope that my job is made more productive rather than redundant by AI, but it is a fine line between the two.

Apologies for the rambling response; I’m curious what other people think.  Do I have my head in the sand about AI?  Am I too pessimistic about the pace of deployment?


BC

20+ years ago, I was at a Big Five auditing firm and several very interesting technology-related incidents occurred that remind me of your ATM example:

1) There was a strategy paper rolled out internally that indicated that due to technology advances, the firm would only need half as many auditors as it currently employed within five years. There was much sturm und drang internally. By somewhat happy/unhappy happenstance, Enron became the last in a long line of Arthur Andersen attest failures (Waste Management, MCI, etc.), and when Enron went under, it (a) dragged Andersen down with them and (b) caused Congress to implement the Sarbanes-Oxley act. Between the Andersen clients fleeing a failing AA for another auditor and Sec. 404 of the S-O act, which effectively doubled the amount of attest work that needed to be done for publicly traded entities, they were able to keep all their auditors and hire more.

Incidentally, Andersen Tax managed to reconstitute itself, and they just went public yesterday, ringing the bell at the NYSE (ANDG). When I was a young marketer, I created the single most outlandishly successful proposal our tech team ever did with the help and active participation of their current chairman, who was a young tax partner then.

2) Marketing took it in the shorts. My team and I were working 100 hour weeks for several months to accommodate all the new business. We were congratulating ourselves as to how the firm would reward us for our intrepidity. Imagine my surprise when I have to speak to the chairman with my two peers. We're waiting in his conference room, and if you put three senior marketers together, we can't help but use flip charts and jot down ideas (2o years ago). There was one flip chart in the room, and as we flipped through it to find a blank page, we flipped over a sheet that contained calculations on how much money the firm's partners would save/make if they eliminated the entire marketing department. Because at that point, they had all the business they could digest and wouldn't need marketing for five years, as they helpfully explained to me. Hence, my mid-career jaunt to the legal profession.

3) I use four AI tools pretty regularly. I just gave them a fairly straightforward request having to do with summarizing HR policies. They came up with four similar, yet very different answers.

4) I created an index the component parts of which added up to 100. One of the AI tools routinely added the components up to 102. Meh 

5) Perhaps this is overly simplistic, but my biggest issue with AI is its right-brained thinking. Yes, it creates and collates work in 5% of the time that a mid-level career person can do...but AI can't ideate worth a damn, not yet anyway, not at the level that I would find interesting or compelling.

RE: AI threatens livelihood of Stanford computer science grads - SkiBum80 - 12-19-2025

There is a dark side to AI as well:

https://www.cnn.com/2025/11/06/us/openai-chatgpt-suicide-lawsuit-invs-vis

"A CNN review of nearly 70 pages of chats between Shamblin and the AI tool in the hours before his July 25 suicide, as well as excerpts from thousands more pages in the months leading up to that night, found that the chatbot repeatedly encouraged the young man as he discussed ending his life – right up to his last moments."

By "encouraged" the context meant that AI encouraged him to go through with suicide

RE: AI threatens livelihood of Stanford computer science grads - BostonCard - 12-19-2025

This year is supposed to be the year of “Agentic AI”, whereby AI “agents” were unleashed to actually doing something.  In fact, Sam Altman (CEO of OpenAI) predicted it was only a matter of time until someone developed a 1 employee Unicorn (company worth at least $1 billion), where the single employee oversees an army of AI agents.

This season of the podcast “Shell Game” is all about trying to actually develop a one-employee company.

https://www.shellgame.co/

Evidently, the AI agents went so far as to plan an offsite.

There’s also this recent article in the WSJ of their testing of an AI-run vending machine business.

https://www.wsj.com/tech/ai/anthropic-claude-ai-vending-machine-agent-b7e84e34?st=3bui1X&reflink=desktopwebshare_permalink

Quote: Then came the chaos. Within days, Claudius had given away nearly all its inventory for free—including a PlayStation 5 it had been talked into buying for “marketing purposes.” It ordered a live fish. It offered to buy stun guns, pepper spray, cigarettes and underwear.

Ironically, this AI agents was powered by Anthropic’s AI (whose CEO was the basis of this thread).

Of course it is easy to make fun of the capabilities now. But they are already useful in some use cases, and they will improve. I expect my workflow to be different in the future, and that certain tasks will be delegated to AI, which might even be better than me at a lot of things, or at the very least, do a “good enough” job for a fraction of my cost of employment. But I am hopeful that this will free me from a lot of some of the rote/drugery to do more creative work.

or failing that, to post more on this message board, at least until someone trains an AI on Matt’s post-game analysis.

BC

RE: AI threatens livelihood of Stanford computer science grads - M T - 12-19-2025

Oh yea, AI-scripted post-game comments

RE: AI threatens livelihood of Stanford computer science grads - martyup - 12-19-2025

We need the first AI assisted football program.  That would be so Stanford.

RE: AI threatens livelihood of Stanford computer science grads - norcard10 - 12-20-2025

Doesn't this feel like lazy reporting? No stanford cs employment statistics, and they interviewed one alum and one student to craft the narrative.

Maybe we need AI to replace journalism sooner rather than later.

RE: AI threatens livelihood of Stanford computer science grads - Mick - 12-20-2025

(12-19-2025, 06:00 PM)martyup Wrote:  We need the first AI assisted football program.  That would be so Stanford.

PROMPT: Create a football prompt for an NCAA football coach to call an offensive play. Please put it from the point of view of both the quarterback and the coach. Your team is outmanned at every position except tight end, center and right guard. Give them the defensive set up (please make it fairly dense, a complex defense with set up, rush coverages and passing coverages). Give the down and distance. Then have the coach call two plays, one for the defense as it is, one for adjustments. Have the quarterback call the cadence and then adjust to defensive shifts. Make it Stanford-style: slower tempo, heavier disguise, academic-fan-proof football.
RESULT: GAME CONTEXT (STANFORD STYLE)
Quarter: 3rd
Clock: 9:18
Score: Down 17–13
Ball: Own 41
Down & Distance: 3rd & 5
Personnel: 12 (1 RB, 2 TE, 2 WR)
Philosophy: Control interior gaps, force linebackers to declare, throw to tight ends on schedule
Talent Reality: Outmanned at WR, OT, RB; advantage at TE, Center, Right Guard

DEFENSIVE STRUCTURE (DETAILED, ACADEMIC)
Front Second Level Secondary Pressure Tendencies
COACH → QB (STANFORD VOICE)
“Alright. They’re in nickel Tite. Mike’s reading you, Will’s green-dogging, safeties are rotating late. We don’t need explosive—we need clean, on-time, five yards. We’re going to make the Mike wrong.”

PLAY CALL #1 — VS BASE DEFENSE
Play
“Gun Right, Ace Slot, Y Option, H Over, Full Slide Strong.”
(This is absolutely a Stanford call: short, functional, no fluff.)

OFFENSIVE STRUCTURE
Protection Routes
QB POV (PRE-SNAP – STANFORD QB)
“Mike is square—reading. Will is apexed and creeping. Boundary safety at 9 and flat-footed. If Mike carries H, Y settles. If Will adds, ball out immediately.”

QB CADENCE
“White 80… White 80…”
(Nickel rocks forward, Will leans)
QB:
“Alert Y… Alert Y.”
“Set… HUT.”

POST-SNAP Gain: 6 yards
Result: First down, chains move, clock continues

COACH → QB (DEFENSIVE ADJUSTMENT ANTICIPATED)
“They’re going to jump the option now and bring pressure from depth. Same personnel. We’re going to run the same picture and change the answer.”

PLAY CALL #2 — VS ADJUSTMENT
Play
“Gun Right, Ace Slot, Duo Bluff, Y Delay Boot.”
(Classic Stanford counter: same look, different truth.)

OFFENSIVE ASSIGNMENTS
Run Action Routes
QB POV (PRE-SNAP)
“Same front. Mike tighter. Will hotter. Boundary safety leaning down. If edge crashes, Y leaks. If flat defender jumps, H behind him.” 

QB CADENCE (WITH CHECK)
“White 80… White 80…”
(Boundary OLB widens)
QB:
“Boot, Boot.”
“Set… HUT.”

POST-SNAP Gain: 12 yards
Result: First down, ball at midfield

SIDELINE (STANFORD COACH CLOSER)
“Good. Stay patient. Make them defend structure. We don’t win fast—we win correctly.”

RE: AI threatens livelihood of Stanford computer science grads - SkiBum80 - 12-20-2025

Gain 12 yards?!
Obviously Stanford students wrote that AI software

RE: AI threatens livelihood of Stanford computer science grads - dabigv13 - 12-20-2025

Interesting recent Reuters article on AI implementation by businesses.

https://archive.is/2025.12.17-071559/https://www.reuters.com/business/business-leaders-agree-ai-is-future-they-just-wish-it-worked-right-now-2025-12-16/
Quote:Jeremy Nielsen, general manager at North American railroad service provider Cando Rail and Terminals, said the company recently tested an AI chatbot for employees to study internal safety reports and training materials.
But Cando ran into a surprising stumbling block: the models couldn’t consistently and correctly summarize the Canadian Rail Operating Rules, a roughly 100-page document that lays out the safety standards for the industry.
Sometimes the models forgot or misinterpreted the rules; other times they invented them from whole cloth. AI researchers say models often struggle to recall what appears in the middle of a long document.
Cando has dropped the project for now, but is testing other ideas. So far the company has spent $300,000 on developing AI products.
“We all thought it’d be the easy button,” Nielsen said. “And that’s just not what happened.”
....
Seemingly small issues can unexpectedly trip up AI systems.
Many financial firms rely on data compiled from a broad range of sources, all of which can be formatted very differently. These differences might prompt an AI tool to “read patterns that don’t exist,” said Clark Shafer, director at advisory firm Alpha Financial Markets Consulting.
Many companies are now looking into the potentially expensive, lengthy and complex process of reformatting their data to take advantage of AI, Shafer said.
Dutch technology investment group Prosus says one of its in-house AI agents is meant to answer questions about its portfolio, similar to what the group’s data analysts on staff already do.
Theoretically, an employee could ask how often a Prosus-backed food-delivery firm was late to deliver sushi orders in Berlin last week.
But for now, the tool doesn’t always understand what neighborhoods are part of Berlin or what “last week” means, said Euro Beinat, head of AI for Prosus.

On the topic of the original article, I think AI is a handy bogeyman for companies to explain why they are hiring fewer people or firing people. Sounds better to investors and the public than other plausible reasons, like anticipating an economic slowdown.

RE: AI threatens livelihood of Stanford computer science grads - martyup - 12-20-2025

In the football play calling context, the AI would only offer suggestions.  The OC would have the final say.  So the AI would be an analyst on steriods.

RE: AI threatens livelihood of Stanford computer science grads - CompSci87 - 12-20-2025

(12-20-2025, 10:32 PM)martyup Wrote:  In the football play calling context, the AI would only offer suggestions.  The OC would have the final say.  So the AI would be an analyst on steriods.
Or maybe an analyst on LSD.