I was trying to show how deeply reliant on technology we are. Ever since the invention of the wheel.

My point is that the problem is that we're forgetting the fundamentals, the basic principles of everything that surrounds us. And there will be a price to pay, eventually.

Yes I see what you mean. When I think about the cell phones it really makes me wonder. We depend very highly on our cell phones now. So much of our life goes on the phone, and we depend on the communications capability.

 

Yes I see what you mean. When I think about the cell phones it really makes me wonder. We depend very highly on our cell phones now. So much of our life goes on the phone, and we depend on the communications capability.

It was the invention of the press that skyrocketed culture throughout western civilization, and ended the Middle Ages. So yes, there's something to think about.

 

https://www.cbc.ca/news/canada/newf...cation-accord-nl-sources-dont-exist-1.7631364
N.L.'s 10-year education action plan cites sources that don't exist
MUN academics say fabricated citations may have been produced using AI

Yet another example of what I have been maintaining all along -- and about far more than AI. If a tool can be abused by people that want to use it to replace their thinking for them, it will invariably be abused but LOTS of people and very likely even most people, eventually, for that purpose.

Any technology runs this risk -- even that slide rule. But the less having a grasp of the fundamentals is necessary to successfully use a technology, the worse its impact is. Unfortunately, the need to develop tools, of all kinds, that enable this result is the price of having the tools needed to push progress every further forward. It's a double-edged sword. The solution requires that we humans accept that we have to still insist that people learn the fundamentals. But too many people insist that the fundamentals are anr't needed any more precisely because they don't need to know them in order to successfully use the tools.

It really is a case of the old joke wherein the race to create better foolproof technology is being lost to the universe's ability to create better fools.

EDIT: Fix typo.

 

ability to create better fools.

One of man’s greatest achievements.

Pre-civilization, the fools didn’t live long enough to reproduce.

 

https://english.elpais.com/technolo...t-time-i-had-written-something-by-myself.html
ChatGPT dissidents, the students who refuse to use AI: ‘I couldn’t remember the last time I had written something by myself’
Some college students are beginning to limit their use of artificial intelligence, so as not to hinder their own creativity, discipline and critical thinking

This student’s view isn’t the most common, but it’s not exceptional, either. More and more students are ceasing to use artificial intelligence (AI) in their assignments. They feel that, with this technology, they’re becoming lazier and less creative, losing the ability to think for themselves.

“I’ve stopped using artificial intelligence to do my university work because it doesn’t do me any good. Last year, I felt less creative. This year, I’m barely using it,” says Macarena Paz Guerrero, a third-year journalism student at Ramon Llull University in Barcelona. “At university, we should encourage experimentation, learning and critical thinking, instead of copying and pasting questions into a machine without even reading them,” she adds.

The workers who are most critical of AI are those who are most demanding of themselves. In other words, the more confident a person is – and the more confidence they have in the tasks they perform – the less they resort to technology. “We’re talking about overqualified [individuals]... that is, students or workers who stand out for their high abilities and encounter limitations when using AI,” says Francisco Javier González Castaño, a professor at the University of Vigo, in Spain.

Exactly. The more you know, the less you trust them.

 

https://english.elpais.com/technolo...t-time-i-had-written-something-by-myself.html
ChatGPT dissidents, the students who refuse to use AI: ‘I couldn’t remember the last time I had written something by myself’
Some college students are beginning to limit their use of artificial intelligence, so as not to hinder their own creativity, discipline and critical thinking



Exactly. The more you know, the less you trust them.

What's sad (and telling) is that this is newsworthy. After all, "news" is primarily the realm of what is out-of-the-ordinary.

 

https://techcrunch.com/2025/09/14/v...nto-ai-babysitters-but-they-say-its-worth-it/

These experienced coders have discovered issues with AI-generated code ranging from hallucinating package names to deleting important information and security risks. Left unchecked, AI code can leave a product far more buggy than what humans would produce.

Working with AI-generated code has become such a problem that it’s given rise to a new corporate coding job known as “vibe code cleanup specialist.”
...
Meanwhile, Rover finds that AI “runs into a wall” when data conflicts with what it was hard-coded to do. “It can offer misleading advice, leave out key elements that are vital, or insert itself into a thought pathway you’re developing,” she said.

She also found that rather than admit to making errors, it will manufacture results.

She shared another example with TechCrunch, where she questioned the results an AI model initially gave her. The model started to give a detailed explanation pretending it used the data she uploaded. Only when she called it out did the AI model confess.

“It freaked me out because it sounded like a toxic co-worker,” she said.

‘you’re absolutely right’

 

OpenAI Has a Fix For Hallucinations, But You Really Won't Like It

​

The paper provides the most rigorous mathematical explanation yet for why these models confidently state falsehoods. It demonstrates that these aren't just an unfortunate side effect of the way that AIs are currently trained, but are mathematically inevitable.

 

OpenAI Has a Fix For Hallucinations, But You Really Won't Like It
View attachment 355988​

I would say provide a confidence level from 0 to 100 percent. That's the way we deal with things that have merit but are only a sort of hint. An example would be when collecting clues to a crime. Some clues would have more merit so would be assigned a higher value like 60 percent, while others would be even higher 85 percent, and some lower 25 percent.

One of the things I find troubling is that an 'ai' bot sometimes is not able to understand its own mistake ever, and therefore never be able to correct it even if told multiple times it is not right. That means the whole conversation comes to an end without a resolution. It is kind of nutty that with some of the simplest formulas it can't get a reasonable result, but with some more complicated formulas it does very well.

I have read that it is being said that one of the Chat Gpt bots passed the Turing test with a score above 70 percent, but then the Turning test is now being questioned as not being a good enough test

 

I would say provide a confidence level from 0 to 100 percent. That's the way we deal with things that have merit but are only a sort of hint. An example would be when collecting clues to a crime. Some clues would have more merit so would be assigned a higher value like 60 percent, while others would be even higher 85 percent, and some lower 25 percent.

One of the things I find troubling is that an 'ai' bot sometimes is not able to understand its own mistake ever, and therefore never be able to correct it even if told multiple times it is not right. That means the whole conversation comes to an end without a resolution. It is kind of nutty that with some of the simplest formulas it can't get a reasonable result, but with some more complicated formulas it does very well.

I have read that it is being said that one of the Chat Gpt bots passed the Turing test with a score above 70 percent, but then the Turning test is now being questioned as not being a good enough test

The Turing test only proves the ability of an algorithm to deceive a human being into thinking that the machine is conscious. It has nothing to do with actually proving that it is, indeed, conscious.

 

OpenAI Has a Fix For Hallucinations, But You Really Won't Like It
View attachment 355988​

In short, the OpenAI paper inadvertently highlights an uncomfortable truth: the business incentives driving consumer AI development remain fundamentally misaligned with reducing hallucinations.

Until these incentives change, hallucinations will persist.

It's cheaper just to lie.

 

an 'ai' bot sometimes is not able to understand its own mistake ever, and therefore never be able to correct it even if told multiple times it is not right.

I asked ChatGPT if it corrects its database when a questioner points out an error in its response. It replied that it does not; it relies only on the training data it already has. But is that reply accurate?

 

I asked ChatGPT if it corrects its database when a questioner points out an error in its response. It replied that it does not; it relies only on the training data it already has. But is that reply accurate?

The questioner might be in error or deliberately lying. Correcting its database from random people or bots is an easy way to poison the training database.

 

an easy way to poison the training database.

..... which is already polluted by misinformation.

 

..... which is already polluted by misinformation.

Even if it (the database) were perfect, most of the same errors would happen per the OpenAI research paper. It's not the data that's the main source of errors, it's the LLM process of guessing the next word.

 

Even if it (the database) were perfect, most of the same errors would happen per the OpenAI research paper. It's not the data that's the main source of errors, it's the LLM process of guessing the next word.

In other words, it's the limitations of the technology itself ... right?

 

In other words, it's the limitations of the technology itself ... right?

Exactly. Some people seem to expect a wind-up toy plane to eventually get to supersonic speeds by using bigger springs or rubber bands, it's not happening.

 

https://mikelovesrobots.substack.com/p/wheres-the-shovelware-why-ai-coding

It turns out, though, and I’ve collected a lot of data on this, it doesn’t just not work for me, it doesn’t work for anyone, and I’m going to prove that.

But first, let’s examine how extreme and widespread these productivity claims are. Cursor’s tagline is “Built to make you extraordinarily productive.” Claude Code’s is “Build Better Software Faster.” GitHub Copilot’s is “Delegate like a boss.” Google says their LLMs make their developers 25% faster. OpenAI makes their own bombastic claims about their coding efficiencies and studies2. And my fellow developers themselves are no better, with 14% claiming they’re seeing a 10x increase in output due to AI.

 

The Turing test only proves the ability of an algorithm to deceive a human being into thinking that the machine is conscious. It has nothing to do with actually proving that it is, indeed, conscious.

Hi,

Yeah, it's not a definitive test and I think everybody knows that now. Back in the day it was sufficient but like everything else, everything in the past that was once fact becomes trumped by new fact. It seems to be one of the rules of life, most likely because we as humans are not perfect, we're just like 'ai' in that sense except we know enough to recurse while 'ai' does not seem to be able to do that very well yet.

 

I asked ChatGPT if it corrects its database when a questioner points out an error in its response. It replied that it does not; it relies only on the training data it already has. But is that reply accurate?

Hi,

That's a great observation. It's like a version of incompleteness unfolding before our very eyes. A system is never capable of examining itself in its entirety.
It also is not always able to give the right result I think because logic is more fundamental than reason:
"Prove that this statement is false",
if it is false then it is true, but if it is true then it is false, which is a contradiction.
or:
"This statement cannot be proved".
If it proves it then the statement is false because the statement claims it cannot be proved, but if it does not prove it then that statement is unprovable.

I guess this is just another part of reality where it depends on other things as well as the main target. I guess we can say that questions like this are multidimensional, but then that sort of disproves logic itself: there is no such thing as a single logical statement that stands alone.
This reminds me of why we use hyphens in language. If we do not show more than one thing as belonging to a structure in and of itself, there could be an ambiguity in the entire meaning which could lead to disastrous results. We also commonly have to use more than one statement to be clear about something.