So you STILL think that AI understands what it types??

[Sculptor]

I understand that perfectly.
But it claims to be correcting itself, as you can quite clearly see from examining the exchanges.
My claim is that it does not understand anything.

I agree. I don't think anything that has no sentience will ever understand anything. Just a more complex form of calculator.

LOL.Yes - clearly you do not understand. It seems you are capable of realting your own experience, that you yourself lack understanding.
Thanks for clearly that up.
Are you sure you are sentient?

Does it not bother you that it makes shit up all the time, even fake references to people and studies that doe not exist and never happened?

I can't say I've given it much of a test. It would be concerning if it has bias conjured into its language model - certainly re political left or right leaning tendencies.

First, I suggest you follow the texts in this thread to see the fabrications and the initial self contradiction.
Then you might want to look online for reported problems where apparently scientific references to non existent studies and scientists have been made.
Sadly, of course, like all other "sentient" beings you cannot understand what I am typing.

[attofishpi]

I don't think anything that has no sentience will ever understand anything. Just a more complex form of calculator.

LOL.Yes - clearly you do not understand. It seems you are capable of realting your own experience, that you yourself lack understanding.
Thanks for clearly that up.

Well then clever dick. Explain how anything simply made of electronic on off switches is going to understand anything?

I don't think you can compare the switches within silicon chips to synaptic switches to support your "comprehension" of what it takes to understand something.

[Sculptor]

I don't think anything that has no sentience will ever understand anything. Just a more complex form of calculator.

LOL.Yes - clearly you do not understand. It seems you are capable of realting your own experience, that you yourself lack understanding.
Thanks for clearly that up.

Well then clever dick. Explain how anything simply made of electronic on off switches is going to understand anything?

It is not.

I don't think you can compare the switches within silicon chips to synaptic switches to support your "comprehension" of what it takes to understand something.

Can you define sentience?

[commonsense]

First of all, if ChatGPT were free,

Version 3.5 is free.

Only after you provide credit card information and agree to subscribe unless you take action to cancel.

[LuckyR]

You should not comment on this subject because clearly you have not read throught the exchanges I has with Chat GPT.
Cockney rhyming slang is not fringe. GPT knows several examples. What it does NOT UNDERSTAND, is that you cannot just make it up as you go along as it did twice.
Even when it accepted that it was corrected, and gave the right answer it soon reverted to its previous error. QED IT CANNOT LEARN becasue it does NOT UNDERSTAND what any of it means.
It just a complicated search engine guilty of numberous and well publicised fabrications and inventions.
In this thread I have shown it is also capable is SELF CONTRADICTION in adjacent sentences.

To nail home the point. I asked it again today what I twice "taught" is yesterday, and twice I got it to answer the question twice properly.
Today once again it just comes up with YET ANOTHER complete invention. Its not even the same as yesterday's inventions.



THis is complete bollocks

Exactly. What (superficially) is impressive about ChatGPT is the breadth of the source of it's data set (the interwebs) not the depth of it's understanding of it, somewhere between zero and infintessimally small.

Indeed. In fact I just do not think the word "understands" even applies.

But the simple fact that it fabricates stuff makes it effectively useless.
Reports suggest that it has even mimicked scientific references for studies that never happened.
The idea that so-called AIs might be used to rule our lives is truly shocking.

Ha ha, you don't understand. ChatGPT was built to emulate a conman who'll say anything to accomplish it's goal.

[Sculptor]

Exactly. What (superficially) is impressive about ChatGPT is the breadth of the source of it's data set (the interwebs) not the depth of it's understanding of it, somewhere between zero and infintessimally small.

Indeed. In fact I just do not think the word "understands" even applies.

But the simple fact that it fabricates stuff makes it effectively useless.
Reports suggest that it has even mimicked scientific references for studies that never happened.
The idea that so-called AIs might be used to rule our lives is truly shocking.

Ha ha, you don't understand. ChatGPT was built to emulate a conman who'll say anything to accomplish it's goal.

So it appears

[commonsense]

Indeed. In fact I just do not think the word "understands" even applies.

But the simple fact that it fabricates stuff makes it effectively useless.
Reports suggest that it has even mimicked scientific references for studies that never happened.
The idea that so-called AIs might be used to rule our lives is truly shocking.

Ha ha, you don't understand. ChatGPT was built to emulate a conman who'll say anything to accomplish it's goal.

So it appears

Is it even possible to program honesty into the mix? Perhaps once Chatty finds a collection of words to employ, it must find a number of additional matches to increase the likelihood of a legitimate response.

[Sculptor]

Ha ha, you don't understand. ChatGPT was built to emulate a conman who'll say anything to accomplish it's goal.

So it appears

Is it even possible to program honesty into the mix? Perhaps once Chatty finds a collection of words to employ, it must find a number of additional matches to increase the likelihood of a legitimate response.

Yes I think it makes "legitimate" responses all the time. But has no discrimination. If the entire world were mad and had mad ideas, Chatty would just mimic them without understanding.
Right now, though he has basic and serious flaws.

[attofishpi]

LOL.Yes - clearly you do not understand. It seems you are capable of realting your own experience, that you yourself lack understanding.
Thanks for clearly that up.

Well then clever dick. Explain how anything simply made of electronic on off switches is going to understand anything?

It is not.

I don't think you can compare the switches within silicon chips to synaptic switches to support your "comprehension" of what it takes to understand something.

Can you define sentience?

Sadly, of course, like all other "sentient" beings you cannot understand what I am typing.

So are you suggesting no sentient being can understand anything? Or, and rather more aptly, no sentient being can understand anything you have written? (then improve the way you write)

[attofishpi]

Only after you provide credit card information and agree to subscribe unless you take action to cancel.

I have never provided a CC and still use it. Maybe they have changed it for newcomers.

[Belinda]

Last month I had an ischemic stroke and I can well remember periods when I could not speak and my thinking was splintered into shards that did not fit together. during periods when I was actively aware I tried hard to set in order my ideas of experiences that I remembered. My ideas sorted themselves out into their accustomed order and I regained the ability to talk sense, respond sensibly to situations, and read joined-up words and sentences.

How my disabled but sentient behaviour differed from that of a broken AI machine is that this Belinda mind/brain was voluntary and future -oriented , whereas the machine constantly lacks volition /future- orientation even when it works properly.

I can see what I have written does not capture the entire nature of how the machine lacks affect such as basic pleasure or pain. So I'd like to add that the mind/brain's biological passion is control of its future, whereas the machine equivalent is its outer environment increasing its power supply etc.

[attofishpi]

Last month I had an ischemic stroke and I can well remember periods when I could not speak and my thinking was splintered into shards that did not fit together. during periods when I was actively aware I tried hard to set in order my ideas of experiences that I remembered. My ideas sorted themselves out into their accustomed order and I regained the ability to talk sense, respond sensibly to situations, and read joined-up words and sentences.

How my disabled but sentient behaviour differed from that of a broken AI machine is that this Belinda mind/brain was voluntary and future -oriented , whereas the machine constantly lacks volition /future- orientation even when it works properly.

I can see what I have written does not capture the entire nature of how the machine lacks affect such as basic pleasure or pain. So I'd like to add that the mind/brain's biological passion is control of its future, whereas the machine equivalent is its outer environment increasing its power supply etc.

Sorry to hear of your stroke Belinda and godspeed to your recovery, glad to see things are improving for you.

RE: "the machine constantly lacks volition /future- orientation even when it works properly." & comparison you make about the biological passion to control our future..

This is something many many years ago I contemplated when I mused over creating my own AI. I thought first I would want to simulate emotions and this "emotion" would then be a factor when the next step in the process would take place where the AI would analyse pre-coded character traits.
What I envisioned was something like a 360 degree 'pie' chart with emotions making up portions of the chart. On boot, the AI would have a random selection of mood from the emotion chart. So consider, today it is unhappy. It would then look at the pre-coded character traits - what does AI instance, in this case for example "Boris" what does Boris enjoy to increase his mood level. This would then drive the AI to research areas that Boris finds interesting and eventually the probability for a better mood swing on the chart would increase. Thus the pie size for each mood would continually fluctuate. (not sure if any of that makes sense!)

[Belinda]

Last month I had an ischemic stroke and I can well remember periods when I could not speak and my thinking was splintered into shards that did not fit together. during periods when I was actively aware I tried hard to set in order my ideas of experiences that I remembered. My ideas sorted themselves out into their accustomed order and I regained the ability to talk sense, respond sensibly to situations, and read joined-up words and sentences.

How my disabled but sentient behaviour differed from that of a broken AI machine is that this Belinda mind/brain was voluntary and future -oriented , whereas the machine constantly lacks volition /future- orientation even when it works properly.

I can see what I have written does not capture the entire nature of how the machine lacks affect such as basic pleasure or pain. So I'd like to add that the mind/brain's biological passion is control of its future, whereas the machine equivalent is its outer environment increasing its power supply etc.

Sorry to hear of your stroke Belinda and godspeed to your recovery, glad to see things are improving for you.

RE: "the machine constantly lacks volition /future- orientation even when it works properly." & comparison you make about the biological passion to control our future..

This is something many many years ago I contemplated when I mused over creating my own AI. I thought first I would want to simulate emotions and this "emotion" would then be a factor when the next step in the process would take place where the AI would analyse pre-coded character traits.
What I envisioned was something like a 360 degree 'pie' chart with emotions making up portions of the chart. On boot, the AI would have a random selection of mood from the emotion chart. So consider, today it is unhappy. It would then look at the pre-coded character traits - what does AI instance, in this case for example "Boris" what does Boris enjoy to increase his mood level. This would then drive the AI to research areas that Boris finds interesting and eventually the probability for a better mood swing on the chart would increase. Thus the pie size for each mood would continually fluctuate. (not sure if any of that makes sense!)

Thanks Attofishpie

I like Boris and his pie chart. Unfortunately perhaps from one problem in philosophy to another?

What is human nature ? is an old problem . Zoology tells what wild dog and wild mouse nature is, but human nature is determined by a mixture of nature(biology) and nurture(culture), and nobody knows how his pie chart would be apportioned. Culture is sapiens' own creation, and it seems to me Boris is sapiens' creation thus a part of his culture and here would be the problem of Boris-nature.

[Sculptor]

Well then clever dick. Explain how anything simply made of electronic on off switches is going to understand anything?

It is not.

I don't think you can compare the switches within silicon chips to synaptic switches to support your "comprehension" of what it takes to understand something.

Can you define sentience?

Sadly, of course, like all other "sentient" beings you cannot understand what I am typing.

So are you suggesting no sentient being can understand anything? Or, and rather more aptly, no sentient being can understand anything you have written? (then improve the way you write)

No I am suggesting that you may not be sentient

[Sculptor]

Last month I had an ischemic stroke and I can well remember periods when I could not speak and my thinking was splintered into shards that did not fit together. during periods when I was actively aware I tried hard to set in order my ideas of experiences that I remembered. My ideas sorted themselves out into their accustomed order and I regained the ability to talk sense, respond sensibly to situations, and read joined-up words and sentences.

How my disabled but sentient behaviour differed from that of a broken AI machine is that this Belinda mind/brain was voluntary and future -oriented , whereas the machine constantly lacks volition /future- orientation even when it works properly.

I can see what I have written does not capture the entire nature of how the machine lacks affect such as basic pleasure or pain. So I'd like to add that the mind/brain's biological passion is control of its future, whereas the machine equivalent is its outer environment increasing its power supply etc.

Sorry to hear that.
This post is crystal clear. So looks like you might have made a complete recovery or are on the way to it.
What have the doctors said about how to carry on, and improve?