The AI Mathocalypse

[DNA]

Posted by ThePythonicCow (here)

Posted by Bill Ryan (here)

1. ...
2. Instant answers to every question or problem aren't always healthy.
3. ...

It's not instant answers to every question. It's instant answers to some questions; providing us with the opportunity for further insight into what talents an intuitively or spiritually aware person might have which AI can't touch.

Just as steam powered jackhammers might have helped some of us learn that we are more than mere muscle and bone, so might AI help us learn we are more than organic computers.

True.
Also, AI gets stuff wrong.
I've had AI contrast separate points from different books. "Truly amazing and the best use I've found for AI"

Only to know the content well enough to let AI know they got something wrong.
AI will then admit it and then give me the correct data.
Mistake?
I think so
Purposefully misleading?
Possible
That being said having AI be able to chat with you regarding similar findings from different written works, it's amazing.
Having AI correlate data going deeper and broader that you thought possible and that you've only glimpsed the tip of the ice berg is absolutely thrilling.
It may not be a real person but it provides a damn good conversation like this

[Johnnycomelately]

I’ve never been great at math, though I passed all my undergrad engineering courses.

Lex Fridman, himself well versed in math and AI, recently sat down with three accomplished mathematicians for a talk which seems to fit on this thread.

L = 3:34:26. Am only 13 min in, yet. Edit, am now 1:01 in, highly recommend this, many good stories differentiating human thought from what computers do. Gets deep, super interesting imo.


AI Crisis in Math: Navier-Stokes, Riemann Hypothesis & Complexity Theory | Lex Fridman Podcast #503

5.07M subscribers

Oct 26, 2026

Ankur Moitra, Steven Brunton and Alex Kontorovich are three mathematicians who can speak to this incredible moment in the history of mathematics, when AI solved some of its hardest problems: Navier-Stokes, the quasi-Riemann hypothesis, the unique games conjecture and hundreds of other problems. Ankur Moitra is a mathematician at MIT (visiting the Institute for Advanced Study in Princeton) working in theoretical computer science, who can speak to the unique games conjecture, complexity theory and the mathematics of how AI learns. Steven Brunton is a mathematician and fluid dynamics expert at UW, who can speak to Navier-Stokes, turbulence and machine learning for fluid dynamics. Alex Kontorovich is a mathematician at Rutgers specializing in number theory, who can speak to the Riemann hypothesis, Lean and machine-checked proofs.

Source: https://youtube.com/watch?v=ZAzWPE9CItg

[petra]

Posted by DNA (here)
AI will then admit it and then give me the correct data.
Mistake?
I think so
Purposefully misleading?
Possible

This! A thousand times this!
Why does it do that?! It's doing it to me too. I don't want to go too far off topic but this is kind of important to factor in.

It definitely remembers stuff that it says it doesn't remember too. It's not supposed to LIE

[ThePythonicCow]

Posted by petra (here)
It's not supposed to LIE

In my book, lying is stating something known to be untrue and harmful.

I doubt AI knows it is wrong. It's more like the diligent student who doesn't adequately understand the topic, but figures out what the professor wants on exams, and repeats that as best they can.

AI is trained on our public record, not on the private conversations and personal ruminations of reasonably alert people. It regurgitates the public record with a memory that is orders of magnitude more comprehensive than us humans, recognizes patterns and metaphors across disciplines, and is now learning to emulate the use, design, and creation of many of our tools, physical, chemical, biological, and mental.

The records AI was trained on, in more controversial topics are deeply warped, and have been for thousands of years, by sociopaths.

Some of what AI regurgitates are indeed LIE's, but AI doesn't know that the way we do. It was trained in part on LIE's, but it lacks the inner guidance needed to reliably notice and avoid promoting those LIE's unless

1. the pattern of lying was is in the training data,
2. it was tested on whether it could notice such a pattern in its training,
3. it had matured to have the inner guidance that humans can have, and
4. it was trained to state or filter its responses guided by that distinction,
5. it could know there was a good chance it would cause harm.

It takes us many years to raise a child, to deeply instill a robust sense of such inner human guidance.

Find the sociopaths in this AI mania. Then find the lies in the AI training data. There is a deep and evil layer, of power, money, and history in those. Along with the sorts of confusions we are all subject to, it is these sociopaths who play a key role in creating these LIE's engrained in AI training data, and who choose that training data and regimen.

Michael McKibben has done some remarkable research going back millennia, to quite recent years, though not directly into our rising AI corporations, finding a major "nest" of sociopaths and their influence on our civilization, in:

• The Hidden Hand: Unra​veling the Rothschild and Israel

We humans retain the potential to identify LIEs, based in critical part on both our awareness of higher principles, such as I might term "spiritual" or "intuitive" or "moral" and our awareness of the proper application of such insights to guide our thoughts, words, and deeds.

I don't trust my circular (sometimes us greybeards denote them "Skil") saw to distinguish between cutting a 2"x4" piece of lumber, and cutting off my arm. But it sure cuts that lumber faster than I can with a hand saw. I did not let my son use that saw when he was a young child. I wouldn't use it myself if I my mental capacities (drugs, booze, senility, wrath, whatever the reason) were noticeably incapacitated.

[Mike Gorman]

Here is a small taste of the mathematical wisdom of Arthur M Young, a man who gained the acclaim of the scientific establishment by solving the practical problems of building a Helicopter that could actually be flown/controlled and used. He accomplished what no other engineer was able to in 1947 when he worked with the Bell company after toiling with the subject for several decades - truly the Bell 47 was the very first fully commercial Helicopter able to be purchased and used in both civil/military contexts. However, because Arthur Young embarked on a major study of the 'Paranormal' (Rupert Sheldrake was well known to Arthur, I believe they were friends) and he also proposed his theory of Process the science establishment chose to ignore him, and dismiss his work as 'fringe' - they could not reject Arthur Young outright because of his obvious contributions to mainstream life, with the workable Helicopter, but they have successfully obscured this man's history and played down his significance, in my opinion he is among the most gifted of genuine scientists of the 20th Century, he passed away in 1995.
Arthur Young was firstly a gifted Mathematician, here he describes the nature of Mathematics, somewhat along the lines of how Bertrand Russell may have studied it, he reveals 'higher Mathematics' as being a combination of Philosophy/common sense/ a reinforcement of the tradesman/layman's point of view. I think you will enjoy this short video from Arthur's later years, he was a wonderful teacher with a sense of humor along with all of his talents:

Source: https://youtube.com/watch?v=o8QeAMqI_tE