Grand Parkway to Real Estate and Life Issues

[kalvado]

Do you have a good term to encompass all that A.I. is without including the word "intelligent"?

I am in no manner an expert on this topic, or even competently informed, but I call it artificial stupidity.  That isn't particularly helpful, so I might call it simulated thought through content mixing.  It finds patterns in words or visual information, without knowing what any of it means, and recombines that content based on other patterns it's observed.  Sometimes it goes sideways and makes stuff up, so you can't rely on it at all, but if you know enough about the topic to recognize a good answer, then you might see a good answer you recognize.  It's basically a substitute for asking someone who knows the answer, but if that someone isn't insane and reliably doesn't start rambling absurdities or mixing in self-destructive suggestions with otherwise good information, that someone is a better option.

I know someone who's impressed with its ability to answer questions.  If most of the answers on the Internet are correct, then that's the source of the correct answers the machine gives.  That, not any understanding of the question, or even what the words mean.  I'd compare it, reminding that I have no subject matter expertise, to the old problem of using Google to get the correct spelling of a word (I remember an article about that decades ago).  Google would not give you the correct spelling, but the most common one.  If the web was full of misspellings of a word, then there's bad source material.  Google only cared about what's common, not what was correct.  These days the site works differently and it may give you a dictionary spelling.

I see artificial stupidity, I mean whatever content mixing I called it, as another way to withdraw from the world, from other people, and from awareness of your own thoughts.  Rather than asking someone, or thinking about it, or being aware of the world so you can learn things yourself, you can ask a machine, just like you can draw up games, videos, and other entertainment.  It's another step toward endless distraction and the excision of the last vestiges of any real meaning in life.  That is my Al doomsday theory (I spell and pronounce it as AL, like Al Bundy, because I refuse to recognize the "i"), not force but surrender, not taking over the world not by overpowering us, but by making us stop wanting control of our lives.  That doesn't require any actual intelligence or will, just endless distraction.  It takes over only because we need something to run things, not because it took control from us.

How do I know that anyone writing in this thread is knowing what those words mean, not just addressing the stimulus patterns?

[wxfree]

Do you have a good term to encompass all that A.I. is without including the word "intelligent"?

I am in no manner an expert on this topic, or even competently informed, but I call it artificial stupidity.  That isn't particularly helpful, so I might call it simulated thought through content mixing.  It finds patterns in words or visual information, without knowing what any of it means, and recombines that content based on other patterns it's observed.  Sometimes it goes sideways and makes stuff up, so you can't rely on it at all, but if you know enough about the topic to recognize a good answer, then you might see a good answer you recognize.  It's basically a substitute for asking someone who knows the answer, but if that someone isn't insane and reliably doesn't start rambling absurdities or mixing in self-destructive suggestions with otherwise good information, that someone is a better option.

I know someone who's impressed with its ability to answer questions.  If most of the answers on the Internet are correct, then that's the source of the correct answers the machine gives.  That, not any understanding of the question, or even what the words mean.  I'd compare it, reminding that I have no subject matter expertise, to the old problem of using Google to get the correct spelling of a word (I remember an article about that decades ago).  Google would not give you the correct spelling, but the most common one.  If the web was full of misspellings of a word, then there's bad source material.  Google only cared about what's common, not what was correct.  These days the site works differently and it may give you a dictionary spelling.

I see artificial stupidity, I mean whatever content mixing I called it, as another way to withdraw from the world, from other people, and from awareness of your own thoughts.  Rather than asking someone, or thinking about it, or being aware of the world so you can learn things yourself, you can ask a machine, just like you can draw up games, videos, and other entertainment.  It's another step toward endless distraction and the excision of the last vestiges of any real meaning in life.  That is my Al doomsday theory (I spell and pronounce it as AL, like Al Bundy, because I refuse to recognize the "i"), not force but surrender, not taking over the world not by overpowering us, but by making us stop wanting control of our lives.  That doesn't require any actual intelligence or will, just endless distraction.  It takes over only because we need something to run things, not because it took control from us.

How do I know that anyone writing in this thread is knowing what those words mean, not just addressing the stimulus patterns?

That's my kind of question.  I don't think we can know that, whether oneself actually understands or is just programmed, whether by technology or by biology, to believe that he understands.  In a sense, I doubt the existence of reality, being the physical and determinative reality we believe in.  To me, physical reality is an interpretation, our interpretation, possibly powered by the interpretation of something else.  I believe that the physical reality we observe and believe in is actually a projection of the most likely outcomes, within a range of possible outcomes.  Do I believe I know what this means?  Yes.  Do I actually know what it means?  I'm not actually certain that I even exist, so I don't know whether I know anything.

[kphoger]

I'm not actually certain that I even exist, so I don't know whether I know anything.

That darned Cartesian evil demon, at it again!

[Scott5114]

Oh, dear.  Does that mean it won't get substantially better than it is right now?  "Yeah, it kind of sucks right now, but just wait till it gets better" is really just "Yeah, it kind of sucks, and it always will"?

"It kind of sucks and it always will" is probably closer to the truth, yeah. The key thing to remember about the LLM approach is that it is probability based, and strings a sentence together by predicting what the most likely next word is. But just because the next word is most likely doesn't mean the next word is correct. It just leverages the fact that more sentences are available on the Internet which are true than which are false. There will always be some risk that a false sentence will be generated just because of the inherent rules of probability.

The current improvements the AI industry is focusing on are trying to mitigate this unfortunate tendency of probability, but unfortunately the downside of the LLM approach is that hole in it can never be eliminated, just mitigated. I saw someone on Reddit put it in way that I like—every time an AI produces a correct answer, it's a coincidence that what the AI generated just happens to match reality, so now the AI industry is in the business of trying to create as many coincidences as possible.

There is also the problem that, because of this approach, AI cannot recognize when it "doesn't know" something, because after all, there is always a "most likely answer", even if that answer is only 0.00001% likely to be correct. This is more likely to cause problems the more obscure the question you ask it is. I recently read of a case where a company fed all of their financials to an AI and had it generate suggestions on how to grow the company. It quickly spat out very impressive insights nobody in the company had put together, and the company began making decisions based on those insights. It wasn't until months later that someone noticed that some data it returned didn't look right. Come to find out the AI hadn't ingested all of the data properly for some reason, so it had just made something up—something way off base—for part of the data. So all of those suggestions the company was spending money to implement weren't remotely based on reality.

I think there are some uses for AI where the LLM approach is currently good enough—giving it a piece of text I've lazily written and having it gussy up the words seems to work fine. It'd be cool to use it to generate dialogue for video game characters so that you can continue to get novel dialogue out of them after whatever gameplay purpose they had has come to pass. But anyone who is trying to use it for life-changing stuff is setting themselves up to eventually look really stupid.

Do you have a good term to encompass all that A.I. is without including the word "intelligent"?

Can I use swear words?

[kphoger]

Do you have a good term to encompass all that A.I. is without including the word "intelligent"?

Can I use swear words?

Let me rephrase that...

In your opinion, what would be a better and more accurate term for everyone to use for A.I. instead of 'artificial intelligence'?

[Rothman]

What's weird to me is that I've seen people use AI to summarize videos on Facebook and the summaries are way off...even with the videos having transcripts or close captioning.  "AS" may very well be appropriate.

[kalvado]

Do you have a good term to encompass all that A.I. is without including the word "intelligent"?

Can I use swear words?

Let me rephrase that...

In your opinion, what would be a better and more accurate term for everyone to use for A.I. instead of 'artificial intelligence'?

LLM?
But frankly speaking there may be a lot to say about language and intelligence correlation in biological systems as well.

[wxfree]

I'm not actually certain that I even exist, so I don't know whether I know anything.

That darned Cartesian evil demon, at it again!

I thought, therefore, upon reflection, I probably was. That is the most likely outcome, but not certain.

[Scott5114]

In your opinion, what would be a better and more accurate term for everyone to use for A.I. instead of 'artificial intelligence'?

The problem is that there really isn't one, because text-based generation and image-based generation use rather different processes. I guess you could go with something like "probabilistic trained-model data generation", but most people aren't going to know what that means.

[GaryV]

So I asked it: "what is a better name than artificial intelligence"

Better names for "artificial intelligence" often focus on specific technical functions or collaborative roles rather than the misleading, broad term "intelligence." Common alternatives include machine learning (ML), algorithmic intelligence, augmented human intelligence (AHI), or simply advanced models. These terms better describe systems that learn, analyze data, and support, rather than replace, human intelligence.

[Rothman]

In your opinion, what would be a better and more accurate term for everyone to use for A.I. instead of 'artificial intelligence'?

The problem is that there really isn't one, because text-based generation and image-based generation use rather different processes. I guess you could go with something like "probabilistic trained-model data generation", but most people aren't going to know what that means.

My brother works for Microsoft and, although I think Google searches alone show that AI suffers from a GIGO problem, he's related to me a few advances in the technology regarding what he calls AI agents' problem solving capabilities when it comes to task completion.  Agents figure out alternative tools to get a job done when tools mentioned by prompters are unavailable, for example.

In other words, although almost everyone that checks an agent's work will find issues half the time, the behaviors starting to be exhibited I think are concerning to the point where dismissing AI as incompetent and thinking that will be the case in the long term is naive.

[Scott5114]

The behaviors starting to be exhibited I think are concerning to the point where dismissing AI as incompetent and thinking that will be the case in the long term is naive.

AI has fundamental design problems incompatible with competence. You can probably get bridge piers to stand on a foundation of Swiss cheese for a little while.

Please note that because your brother works for Microsoft, he has probably been pressured by management to keep up the pretense that AI is inevitable, because the second AI has a "emperor-has-no-clothes" moment, Microsoft's stock tanks, and a lot of its business partners go bankrupt.

[Rothman]

The behaviors starting to be exhibited I think are concerning to the point where dismissing AI as incompetent and thinking that will be the case in the long term is naive.

AI has fundamental design problems incompatible with competence. You can probably get a skyscraper to stand on a foundation of Swiss cheese for a little while.

Please note that because your brother works for Microsoft, he has probably been pressured by management to keep up the pretense that AI is inevitable, because the second AI has a "emperor-has-no-clothes" moment, Microsoft's stock tanks, and a lot of its business partners go bankrupt.

Sure, but he's in the industry and closer to the actual developments than anyone on this forum, I'd imagine.

[Scott5114]

Sure, but he's in the industry and closer to the actual developments than anyone on this forum, I'd imagine.

For sure. But you don't have to take my word for it—Las Vegas tech writer Ed Zitron has looked into this fairly deeply. (A lot of his writings are paywalled but you can get a pretty good idea of the state of things, especially the shenanigans that are going on on the business side, from what he has available for free.) The information he discusses squares with my experience enough that I believe he's a credible source.

[kphoger]

In your opinion, what would be a better and more accurate term for everyone to use for A.I. instead of 'artificial intelligence'?

LLM?

But I don't like that, because it seems to limit (perhaps correctly) the scope of the term to word-based A.I.  It doesn't encompass music or art or anything like that.

I guess you could go with something like "probabilistic trained-model data generation", but most people aren't going to know what that means.

I'm also not about to adopt a 14-syllable alternative.  'Artificial intelligence' is long enough as it is.

So I asked it: "what is a better name than artificial intelligence"

Better names for "artificial intelligence" often focus on specific technical functions or collaborative roles rather than the misleading, broad term "intelligence." Common alternatives include machine learning (ML), algorithmic intelligence, augmented human intelligence (AHI), or simply advanced models. These terms better describe systems that learn, analyze data, and support, rather than replace, human intelligence.

I do kind of like the bolded two, or some combination of them.

[hotdogPi]

One person in another community I'm in jokingly calls LLMs "low level machines". The concept originally came from the completely unrelated LLVM, meaning "low-level virtual machine", where "low-level" means "assembly code". Of course, a "low-level machine" as in LLM is low-level as in stupid.

[Bobby5280]

For sure. But you don't have to take my word for it—Las Vegas tech writer Ed Zitron has looked into this fairly deeply. (A lot of his writings are paywalled but you can get a pretty good idea of the state of things, especially the shenanigans that are going on on the business side, from what he has available for free.) The information he discusses squares with my experience enough that I believe he's a credible source.

The business shenanigans (specifically all the circular dealing going on between companies like NVidia, OpenAI, Microsoft, etc.) should be the main thing triggering alarms in the short term.

The DJIA may have hit "$50,000" recently, but that's all from the "magnificent 7" tech companies having their valuations extremely inflated from this AI gold rush. Much of it is all bullshit. OpenAI would need to bring in more than $2 trillion in revenue within the next 3 or so years to justify its valuation; last year it brought in $60 billion. It may take another year or two before financial reality finally sets in, but when it happens the stock market could get hammered really bad. The AI bubble isn't the only thing that can put downward pressure on the markets either. The housing industry bubble, consumer debt setting new record highs and a worsening jobs market are serious factors.

The thing that concerns me personally about AI is how businesses will use it to get rid of employees. With all the flaws of AI duly noted, employers really don't care about work being done perfectly. They are more than willing to settle for completed work that is just good enough to get by -especially if it is done very quickly.

[wxfree]

It's another step toward endless distraction and the excision of the last vestiges of any real meaning in life.  That is my Al doomsday theory (I spell and pronounce it as AL, like Al Bundy, because I refuse to recognize the "i"), not force but surrender, not taking over the world not by overpowering us, but by making us stop wanting control of our lives.  That doesn't require any actual intelligence or will, just endless distraction.  It takes over only because we need something to run things, not because it took control from us.

Big Nate's doomsday scenario
https://www.gocomics.com/bignate/2026/02/16

https://www.gocomics.com/bignate/2026/02/17
He should say "I hope he's hit rock bottom."

[Scott5114]

Much of it is all bullshit. OpenAI would need to bring in more than $2 trillion in revenue within the next 3 or so years...

Which is a huge stinking problem for OpenAI when every ChatGPT query loses them money!

The thing that concerns me personally about AI is how businesses will use it to get rid of employees. With all the flaws of AI duly noted, employers really don't care about work being done perfectly. They are more than willing to settle for completed work that is just good enough to get by -especially if it is done very quickly.

The problem for companies is that AI is very good at producing output that looks like it is good enough to get by, but isn't. It can produce a document that has the form of a legal brief, but when you check the citations to case law none of the cases actually exist. Or it can produce a document that has the form of a financial statement, but all of the numbers are made up. Or it can produce marketing copy which—yeah okay marketing copy is useless so AI is perfect for that.

[Bobby5280]

The problem for companies is that AI is very good at producing output that looks like it is good enough to get by, but isn't. It can produce a document that has the form of a legal brief, but when you check the citations to case law none of the cases actually exist. Or it can produce a document that has the form of a financial statement, but all of the numbers are made up. Or it can produce marketing copy which—yeah okay marketing copy is useless so AI is perfect for that.

This is why some computing industry experts are calling the current situation the "centaur" phase. What that means is a human worker combined with an AI "agent." A company will have a lower number of human employees using AI tools to supposedly do more work and get it done faster. Instead of having a human employee write something like a legal brief an AI "agent" will do the grunt work and the human has to proofread it and correct errors. Software companies are doing more and more of this with coding work. The problem is all the error correcting can be a real time vampire, not to mention very unrewarding.

I know I don't get any enjoyment out of having to fix the odd glitches in some customer provided "logo" AND get it converted into clean, vector form. I could be doing something more productive rather than wasting time on that shit.

[TheCatalyst31]

I love how this thread got a name with three topics in it, yet as soon as it was split the discussion pivoted to a fourth topic.

[Max Rockatansky]

I love how this thread got a name with three topics in it, yet as soon as it was split the discussion pivoted to a fourth topic.

When you live on a toll island you have the time to ponder the mysteries of life.

[Scott5114]

I love how this thread got a name with three topics in it, yet as soon as it was split the discussion pivoted to a fourth topic.

ma'am this is AARoads, I would have been surprised if it hadn't

[kalvado]

I love how this thread got a name with three topics in it, yet as soon as it was split the discussion pivoted to a fourth topic.

ma'am this is AARoads, I would have been surprised if it hadn't

Unfortunately my efforts in pushing biology as a fifth one failed...

[kphoger]

I know I don't get any enjoyment out of having to fix the odd glitches in some customer provided "logo" AND get it converted into clean, vector form. I could be doing something more productive rather than wasting time on that shit.

I have to imagine that it's more fulfilling to spend time cleaning up a human-produced logo than a machine-A.I.-produced one.