AI Fails
- Thread starter ISULibrarian
- Start date
I asked about the date. It said that AI had made a speculative assessment of a future game.
(Hot damn, Chat GPT should run for office)
Here's what it said about the score...
I couldn’t find any completed game results for a December 21, 2025 Iowa State Cyclones women’s basketball game from reliable sources on the web — it appears the Dec 21 matchup vs. Kansas had not yet been finalized or published in the official box scores/news releases at the time of the searches. ESPN.com
However, here are the most recent confirmed Iowa State women’s basketball scores near that date:
- Dec 14, 2025 — Iowa State defeated UNI 81–53. Iowa State Athletics
- Dec 10, 2025 — Iowa State beat Iowa 74–69. Iowa State Athletics
If you’re referring to a specific Dec 21 result you saw reported by another system or AI, it may have been incorrect or speculative — because no validated score for Dec 21, 2025 is available in the official game recaps or box scores at this time. Let me know if you want me to check again later or look for the live/finished score once it’s published.
The bubble popping (which I expect it will) won't make AI go away. Similar to how the Dot Com bubble popping didn't make the internet go away. It just means that a lot of people's retirement funds will suck for quite a while. AI absolutely is a transformational technology, but also right now it's crazy over-valued.
Last edited:
Like all tech capabilities that have come before it, AI can be absolutely brilliant and absolutely stupid at the same time.
Some of the disasters above are the result of really stupid humans. You can’t just “set it and forget it” with these programs. Designers and operators need to set boundaries and parameters. There are degrees of autonomy that can be considered.
I’m not an AI fan boy and have a healthy amount of skepticism. I have used it successfully in mortgage lending operations without incident. There are also other use cases I wouldn’t even consider it. And many use cases in between that I’d definitely want a human in the loop (I.e. more at the “assisted” or “augmented” end of the autonomy scale).
Like many things, it’s not all good or all bad. I do fear that it’s going to kill a lot of entry to mid level white collar jobs that are formulaic and repetitive. I could take my former team of 400 mortgage operations professionals and probably run it with 100-150 people with AI capabilities available today.
Would there be some risks? Sure. But you’re willing to accept some risks if you’re saving tens of millions a year on labor. That’s the calculus that will take place.
Some of the disasters above are the result of really stupid humans. You can’t just “set it and forget it” with these programs. Designers and operators need to set boundaries and parameters. There are degrees of autonomy that can be considered.
I’m not an AI fan boy and have a healthy amount of skepticism. I have used it successfully in mortgage lending operations without incident. There are also other use cases I wouldn’t even consider it. And many use cases in between that I’d definitely want a human in the loop (I.e. more at the “assisted” or “augmented” end of the autonomy scale).
Like many things, it’s not all good or all bad. I do fear that it’s going to kill a lot of entry to mid level white collar jobs that are formulaic and repetitive. I could take my former team of 400 mortgage operations professionals and probably run it with 100-150 people with AI capabilities available today.
Would there be some risks? Sure. But you’re willing to accept some risks if you’re saving tens of millions a year on labor. That’s the calculus that will take place.
IMO you don't replace people with AI, you use AI to make people more efficient. It has accelerated my productivity and this is just early stages. It can't replace me as it needs human oversight. You shouldn't be using it to replace people, at least not yet.
Sports will always be one of AI's greatest weak spots (among others). That's because AI's overarching weakness is that it is incapable of quickly adjusting to new information. This is because all it's capabilities are based on determining predictive weights based on massive averages, and new information is just a drop in that bucket.
It's the same reason Tesla "full self driving" failed in so many edge cases: it didn't have enough examples of extremely rare cases to mathematically determine the proper weights in that context.
Oh, also it isn't determining these weights based on fact or objective truth. Everything is adjusted based entirely on user feedback.
It's the same reason Tesla "full self driving" failed in so many edge cases: it didn't have enough examples of extremely rare cases to mathematically determine the proper weights in that context.
Oh, also it isn't determining these weights based on fact or objective truth. Everything is adjusted based entirely on user feedback.
Totally fair point and I don’t disagree. I do, however, think that if you spread those productivity gain across a team of 10-20 people doing similar jobs you will see companies eliminate some positions. This is obviously not full on “replacement” and some of those positions may get shifted into an oversight/orchestrator kind of role.
This is obviously going to vary by industry and by role. I don’t see anyway there aren’t negative effects on the labor market.
Labor market adapts if they know what they are doing, like moving people to delivery drivers for fast food or having people fill orders for online grocery orders.
"Adapt or die" - Chris Williams
I think current positions will be eliminated and replaced by new ones. AI is new but tech isn't, the number of employed Americans by percent hardly fluctuates with new advancements. Back in the day, typewriting used to be a solid gig. Now it's expected before middle school.
You are more optimistic than I. I don’t think it’s the doomsday scenario (at least not yet) that some are predicting. In my field (all are unique in their own right, of course) I could see significant reductions. It will not be complete loss and - to your point - it will create some new ones in its place. But still see a net loss that is significant. Financial institutions will be happy to book these savings and will do little repurposing of employees.
I agree with a lot of this, but your last point may not be true for all use cases. To my earlier point about the intelligence and foresight of humans being a critical factor in performance, you can control what information you feed to it. There is a strong “garbage in, garbage out” component to it like so many other technology applications.
I don't know your field but I'm a software engineer which is my basis for my thoughts, feel like what I do will become the next typewriter type job as time goes on. Will just become something most people can at least do at a minimal level and be devalued, but I'm glad to live in and work in the time of the boom. Love what I do but also understand that a lot of the nature of AI is to replace what I do, done by the people that do what I do. But that's why I say embrace the tech, because early adopters are the ones that are ahead when the time comes that it takes over. AI is one of those times.
Software engineering is tough right now from what I hear. I can’t help but see the irony in the fact that software engineers amongst the first group of professionals to be seriously impacted.
I work in the mortgage industry where my teams process and underwrite loan documents. So may policies and regulations are algebraic in nature. Basically a web of conditional if-then statements. People will tell you it’s really nuanced but there is a difference between something being “complex” and truly requiring human judgment. Memorizing a bunch of underwriting guidelines based on the loan type, low purpose, loan amount, FICO score, loan to value ratio, etc is exactly what machines are good at and they don’t make mistakes. You can always kick your corner cases and exceptions out to a human.
Problem historically in our business is not the logic as much as it is the underlying data. Some is digitized and standardized, but it’s still humans reading through a ton of PDF documents trying to find a small number of data points they need to make a decision.
I can see generative AI largely solving this problem and then agentic AI taking over from there. Again, not in all cases but in most.
People that say “well the lenders will be able to pass those cost savings on to borrowers” are just naive. Consumers will be charge the same and financial institutions will just print more money.
I do think there will be other advantages for those that adopt - primarily in the amount of time it takes to process your loan. If I can approve you in 3 days and someone else is taking 3 weeks there will be winners that emerge quickly. So agree with your point 100% about the need to adapt.
I didn't mean to imply software engineering jobs are in a bad spot, just in the future the skills will be more commonplace. I don't know a single unemployed software engineer here in Iowa. If anyone out there is, let me know and I can help you out. I just wanted to hear perspectives from other industries, always looking to learn. Thanks for the mortgage perspective.
I work in accounting/finance and IMO that industry will be hit very hard likely within the next 10 years.
And there are some things that are simply not replaceable. We will likely need much fewer radiologists in the very near future. Average salary in America for radiologists is 500k. There just aren’t going to be thousands of 500k+ salary jobs out there they can just roll into.
Your last point is particularly important, in my opinion.
Replacing people at the counter in a fast food restaurant and repurposing them as delivery drivers is one thing. When you start talking about your example (the radiologist making $500k) or even mine (the underwriter making $125k) is where you're going to start seeing economic disruption.
I'll be careful what I say here because I don't want this to go to the cave, but there are strong "who does this really benefit - wealthy executives and shareholders or the middle class" vibes to it.
Just like people?
