A conversation with AI, helped me out...

[psikeyhackr]

I do not engage in conversations with AIs.

I ask questions about specific subjects then check non-AI sources depending on the answers. I have had an AI give me contradicting answers on the same subject in a matter of minutes.

I have asked this of two different AIs and gotten similar answers:

Ask an AI when PhD economists should have figured out Planned Obsolescence in automobiles. An AI does not have an ego, or career, or reputation among colleagues to worry about.

Can the economists in the nation that put men on the Moon not figure out Planned Obsolescence in automobiles in decades?

Economists do not talk about the *NET* Domestic Product? They leave the Depreciation of Durable Consumer Goods, like Automobiles purchased by consumers, out of the equation.

So now the world must deal with the Results of Seven Decades of Defective Algebra!
===================

I was debating between Mechanical and Electrical Engineering when I was in high school. I went with electrical. I have figured Planned Obsolescence was going on in cars for decades. I have never owned a new car.

But here we are half-a-century after the Moon Landing and PhD economists do not talk about it. I have encountered people who have never heard of it. The term has come up in relation to smart phones in recent years.

Some people are prejudiced against AI use. The problem is with the technology being over hyped and overbuilt and proprietary intellectual property being used to train it not the technology itself.

I find it ironic for people to bitch about AI but not the Planned Obsolescence of Automobiles since at least the 1970s.
.

 

[Wizard of Whatever]

AI is usually Ok on objective data, but not really good on pattern and conclusion.

 

[DaveC426913]

I study subjects that I discuss with AI, many of the mistakes are just silly, easily seen.

Exactly.

Which means:
1. In areas where you know the answers, there's no reason to ask a chatbot.

2. In areas where you don't know the answers, you don't know when it gets it wrong, so you can't trust the answers.

The only remaining motive for using a chatbot is when you know the answer but want to pretend something else vindicates you. In other words: when you just want it to tell you only what you want to hear.

Because it helps me understand more than one view of something. Like OBE, I've been studying that for years, talked about it a bit yet still don't know as much as I wanted to learn about it.

No. As above: you don't know when it's making stuff up.

AI, just correlates masses of data and if you drill down with the answers, you'll get good information.

No. It is just telling you what you want to hear.

Treat AI as a friend, you'll ask better questions because you'll not be suspicious for 8 hours of the learning.

Yikes! "We got a live one here folks!"

I've had more sense out of it about this OBE subject in 8 hours than I have in 20 years of study, deep study that got me virtually no where.

I'll bet.

Telling you 'things you want to hear', and telling you 'things that are true' are orthogonal axes.

You are headed down a very destructive path.

 

[DaveC426913]

Because it's using the latest science

How do you know?

Are you aware that chatbots will make up citations?

 

[davewhite04]

I do not engage in conversations with AIs.

I ask questions about specific subjects then check non-AI sources depending on the answers. I have had an AI give me contradicting answers on the same subject in a matter of minutes.

I have asked this of two different AIs and gotten similar answers:

Ask an AI when PhD economists should have figured out Planned Obsolescence in automobiles. An AI does not have an ego, or career, or reputation among colleagues to worry about.

Can the economists in the nation that put men on the Moon not figure out Planned Obsolescence in automobiles in decades?

Economists do not talk about the *NET* Domestic Product? They leave the Depreciation of Durable Consumer Goods, like Automobiles purchased by consumers, out of the equation.

So now the world must deal with the Results of Seven Decades of Defective Algebra!
===================

I was debating between Mechanical and Electrical Engineering when I was in high school. I went with electrical. I have figured Planned Obsolescence was going on in cars for decades. I have never owned a new car.

But here we are half-a-century after the Moon Landing and PhD economists do not talk about it. I have encountered people who have never heard of it. The term has come up in relation to smart phones in recent years.

Some people are prejudiced against AI use. The problem is with the technology being over hyped and overbuilt and proprietary intellectual property being used to train it not the technology itself.

I find it ironic for people to bitch about AI but not the Planned Obsolescence of Automobiles since at least the 1970s.
.

How do you learn new things? Or have

Exactly.

Which means:
1. In areas where you know the answers, there's no reason to ask a chatbot.

It doesn't work like that, anyone honest who uses AI for answers check references, and don't stop learning.

2. In areas where you don't know the answers, you don't know when it gets it wrong, so you can't trust the answers.

Your visualisation of me using AI is steeped in ignorance. I have told you, I double, triple even quadrupal check info it reels back, in other words I study quicker, it's a tool not a monolith out of 2001.

The only remaining motive for using a chatbot is when you know the answer but want to pretend something else vindicates you. In other words: when you just want it to tell you only what you want to hear.

So I guess you're accusing me of this too?

No. As above: you don't know when it's making stuff up.

How do you know when a astronomy article of interest pops up in New Scientist, how do you approach it?

You are headed down a very destructive path.

? wtf are you talking about?

 

[exchemist]

How do you learn new things? Or have


It doesn't work like that, anyone honest who uses AI for answers check references, and don't stop learning.



Your visualisation of me using AI is steeped in ignorance. I have told you, I double, triple even quadrupal check info it reels back, in other words I study quicker, it's a tool not a monolith out of 2001.



So I guess you're accusing me of this too?



How do you know when a astronomy article of interest pops up in New Scientist, how do you approach it?



? wtf are you talking about?

New Scientist doesn’t make stuff up. It has an editor and a reputation to protect. A chatbot has neither. It is a stochastic parrot.

It is also designed to be addictive, by talking nicely to you and telling you what you want to hear.

 

[davewhite04]

New Scientist doesn’t make stuff up. It has an editor and a reputation to protect. A chatbot has neither. It is a stochastic parrot.

It is also designed to be addictive, by talking nicely to you and telling you what you want to hear.

Have you used ChatGPT?

 

[exchemist]

Have you used ChatGPT?

Certainly not. I have seen the results of people who have, and the effect it has on some of them, and that’s enough to convince me not to touch it with a bargepole. On the other, more serious, science forum I belong to, we are constantly bombarded with idiotic, badly written, prolix, pompous, erroneous shit written by chatbots. So no thanks.

I can see value in them as a souped-up search engine, provided they give you a list of the sources they have consulted in providing their answer, so that you can read those for yourself before relying on the output. But that’s about it. And the way they try to “engage” you is creepy and dangerous, in my opinion.

 

[DaveC426913]

davewhite04 himself has provided a textbook case study in how foolish it is to trust anything a chatbot says:

https://www.sciforums.com/threads/s...ry-is-not-a-leap-of-faith.167184/post-3777453

He started using this term "scientific fundie" to pidgeonhole and insult every "scientist" on this site.
When asked what he meant by that term, he engaged in more insults, wondering why no one seemed to know what the term meant.
And then we had to go look it up himself.
And he asked his chatbot.
And this is what his chatbot said:

AI:

what is a scientific fundie

A scientific fundie (short for scientific fundamentalist) is a slang, critical label for a person, not a scientific role.
It describes someone who is perceived to:

• Treat current science as absolute truth, rather than provisional knowledge
• Believe scientific consensus should not be questioned or debated
• Dismiss ideas outside mainstream science without engagement
• Confuse science (a method) with scientism (an ideology)

Great.

But here is what a different chatbot said:

Science fundie" isn't a formal term but generally refers to someone holding fundamentalist religious views (often Christian) who
either rejects mainstream scientific findings (like evolution or a billion-year-old Earth) in favor of a literal interpretation of sacred texts (Young Earth Creationism, Creation Science) or rigidly applies scientific methods to prove faith."

These two descriptions are diametrically opposed. The former says a fundie holds science as absolute truth while the latter says science fundies are science deniers.

That's because it is not a formal term - and his friend did not bother to share that nugget with him. He was fooled by a chatbot telling him what he wanted to hear.

Did you say "Friend?" Yes: friend. His words.

davewhite04 self-described his chatbot as a "friend" four distinct times in this very thread:
"AI was better than any friend"
"AI can be a friend"
"It is like a friend"
and the kicker:
"Treat AI as a friend, you'll ask better questions because you'll not be suspicious"


And he trusted his chatbot so utterly that he turned off his own brain and let it speak for him. He made a fool of himself by:
- thinking a term exists that doesn't exist formally
- pretending to be astonished the rest of us did not know (this made up term)
- not knowing what it meant himself
- deciding that, because his AI friend made up a meaning, he was warranted in laying on the insults even more
- getting it wrong
- sticking his fingers in his ears and ignoring the messenger.




Just say no, kids.

 

[exchemist]

davewhite04 himself has provided a textbook case study in how foolish it is to trust anything a chatbot says:

https://www.sciforums.com/threads/s...ry-is-not-a-leap-of-faith.167184/post-3777453

He started using this term "scientific fundie" to pidgeonhole and insult every "scientist" on this site.
When asked what he meant by that term, he engaged in more insults, wondering why no one seemed to know what the term meant.
And then we had to go look it up himself.
And he asked his chatbot.
And this is what his chatbot said:


Great.

But here is what a different chatbot said:



These two descriptions are diametrically opposed. The former says a fundie holds science as absolute truth while the latter says science fundies are science deniers.

That's because it is not a formal term. He was fooled by a chatbot telling him what he wanted to hear.

davewhite04 self-described his chatbot as a "friend" four distinct times in this very thread:
"AI was better than any friend"
"AI can be a friend"
"It is like a friend"
and the kicker:
"Treat AI as a friend, you'll ask better questions because you'll not be suspicious"


And he trusted his chatbot so utterly that he turned off his own brain and let it speak for him. He made a fool of himself by:
- thinking a term exists that doesn't exist formally
- pretending to be astonished the rest of us did not know (this made up term)
- not knowing what it meant himself
- deciding that, because his AI friend made up a meaning, he was warranted in laying on the insults even more
- getting it wrong
- sticking his fingers in his ears and ignoring the messenger.


View attachment 7233

Just say no, kids.

To be fair, the first chatbot (Dave’s) made a decent stab at guessing what the term “scientific fundie” could be expected to mean, even though, so far as I know*it is not a generally recognised term. So the error in that case was to pretend the term exists, when it doesn’t really. A truly intelligent responder would have pointed out in the reply that the term has no recognised meaning, but a meaning could be guessed, along the lines that follow.

The second one just ignored the qualifier “scientific” and churned out an answer based on the term “fundie”alone, which is slang for religious fundamentalist, so utterly wrong.

*A conventional web search for “scientific fundie” brings up nothing, just a load of references to religious fndamentalists.

 

[davewhite04]

To be fair, the first chatbot (Dave’s) made a decent stab at guessing what the term “scientific fundie” could be expected to mean, even though, so far as I know*it is not a generally recognised term. So the error in that case was to pretend the term exists, when it doesn’t really. A truly intelligent responder would have pointed out in the reply that the term has no recognised meaning, but a meaning could be guessed, along the lines that follow.

The second one just ignored the qualifier “scientific” and churned out an answer based on the term “fundie”alone, which is slang for religious fundamentalist, so utterly wrong.

*A conventional web search for “scientific fundie” brings up nothing, just a load of references to religious fndamentalists.

You're simply dishonest like Dave, James and a few others. Where are your sources?

 

[davewhite04]

Certainly not. I have seen the results of people who have, and the effect it has on some of them, and that’s enough to convince me not to touch it with a bargepole. On the other, more serious, science forum I belong to, we are constantly bombarded with idiotic, badly written, prolix, pompous, erroneous shit written by chatbots. So no thanks.

I can see value in them as a souped-up search engine, provided they give you a list of the sources they have consulted in providing their answer, so that you can read those for yourself before relying on the output. But that’s about it. And the way they try to “engage” you is creepy and dangerous, in my opinion.

Fair enough.

Welcome to the creepy future.

 

[exchemist]

You're simply dishonest like Dave, James and a few others. Where are your sources?

Here: https://en.wikipedia.org/wiki/Scientism

There are other similar ones, too.

(You could have looked this up yourself before rashly accusing me of “dishonesty”.)

 

[gmilam]

In the field of artificial intelligence (AI), a hallucination or artificial hallucination (also called bullshitting,[1][2] confabulation,[3] or delusion[4]) is a response generated by AI that contains false or misleading information presented as fact.

Hallucination (artificial intelligence) - Wikipedia

en.wikipedia.org

 

[davewhite04]

Hallucination (artificial intelligence) - Wikipedia

en.wikipedia.org

A book spews bullshit if you can't understand it.

 

[gmilam]

A book spews bullshit if you can't understand it.

Great analogy for AI! Because AI doesn't understand what it's reading, it's just putting words together.

How Does ChatGPT Work Technically? Unraveling the Technical Wizardry - Aitechtonic

Curious about how ChatGPT works technically? This comprehensive article dives deep into the inner workings of ChatGPT, explaining its technical aspects, algorithms, and underlying architecture. Discover the fascinating world of AI language models and understand how ChatGPT creates human-like...

aitechtonic.com

 

[James R]

You're simply dishonest like Dave, James and a few others.

Er... what?

You really need to stop accusing people of dishonesty, unless you can - at a minimum - demonstrate that they have been dishonest about something.

I might add that this is something I often have to consider in moderating here. We have a posting guideline that basically says "Don't knowingly tell lies". Now, if I were to slap a warning on every post in which I suspect a person might be deliberately telling a lie while in fact being fully aware of what the truth is, then more than one of the regular posters here would be long gone from this site. But one has to be fair, and as a moderator it is right to document the reasons why you are moderating somebody. It is the documentation that makes you accountable for your actions, you see. People who are concerned can go and check the objective evidence for themselves, which is deliberately put on the record. If evidence for the claim is not provided, then an accusation of dishonesty risks looking like an empty personal snipe, which is reputation damaging at the least and damning (especially for a moderator) at worst.

I suggest that, going forward, the next time you feel like throwing out an accusation that somebody is "simply dishonest", you should make sure that you can document one or more cases of deliberate dishonest things they have said. At a bare minimum, you must document at least ONE instance.

Also, beware of rash generalisations. If a person is wrong about something, don't assume they are deliberately lying about it. Consider that they might honestly not know the truth. And if a person lies about ONE thing, that doesn't necessarily make them "simply dishonest" in all things. Everybody tells lies, but most people are basically honest.

If you continually throw generalised character assessments around, based on nothing but your own desire to lash out at people, what happens is that you lose the respect of people around you. It ends up being you who looks like the liar, because that is exactly what you are if you're just making stuff up to accuse people.

On another matter, davewhite04: why did you call me out in this thread? Before now, I have posted nothing in the thread. Chip on your shoulder about something else? Got a grudge?

Time to grow up, davewhite04.

 

[James R]

Science fundie" isn't a formal term but generally refers to someone holding fundamentalist religious views (often Christian) who either rejects mainstream scientific findings (like evolution or a billion-year-old Earth) in favor of a literal interpretation of sacred texts (Young Earth Creationism, Creation Science) or rigidly applies scientific methods to prove faith."

To be fair, it looks like this AI probably interpreted the query as referring to religious fundamentalists when they attempt to talk about the failings of science. That is, the "science fundie" in this scenario is not the person who accepts science, but the person who denies it for fundamentalist religious reasons.

This just goes to show that one needs to be careful in framing queries to AI chatbots. If you don't give enough context, they can misinterpret and/or hallucinate.

I'd be interested to know which two AIs are quoted here. The first one looks like Gemini or ChatGPT, but I could be wrong. What about the second one?

 

[billvon]

Have you used ChatGPT?

I have. That's how I know how bad it is.

 

[DaveC426913]

I'd be interested to know which two AIs are quoted here. The first one looks like Gemini or ChatGPT, but I could be wrong. What about the second one?

DaveWhite is a fan of ChatGPT - the one that claimed 'science fundie' means pro-science.
The one I used was Google's Gemini - the one that claimed 'science fundie' means anti-science.