Chatting with AI

General discussion for topics related to the FreeBASIC project or its community.
BasicCoder2
Posts: 3814
Joined: Jan 01, 2009 7:03
Location: Australia

Re: Chatting with AI

Post by BasicCoder2 »

grindstone:

Yes I noticed not looking directly at the image made the faces easier to discern.

I think what was being said was that this skill correlated with other methods of defining/measuring an IQ value.
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

BasicCoder2 wrote:I think what was being said was that this skill correlated with other methods of defining/measuring an IQ value.
That is precisely what is being said.

If I take my reading glasses off and squint a little, then the patterns lose their influence and the women's faces are obviously seen. They are different women, and image two has a Mona Lisa Smile. Image four is no longer an issue and has been cropped, showing no ears.

To increase my IQ then all I need to do is “take my reading glasses off and squint a little”. I don't think so.

We need to be careful with correlation.

Folk who have had a few bevvies may see the women's faces more clearly than folk who partook in water only. To increase my IQ then all I have to do is have a drink or two. Nope on that one as well. :)
BasicCoder2
Posts: 3814
Joined: Jan 01, 2009 7:03
Location: Australia

Re: Chatting with AI

Post by BasicCoder2 »

deltarho[1859]

Got it :)
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

I had a quick chat yesterday morning but failed to connect later.

I got through this morning.
BasicCoder2 wrote:I would ask it big world questions like how to bring world peace :)
I got this:
Bringing world peace is a complex and multifaceted goal that requires the efforts of individuals, communities, and governments at all levels. Some steps that can be taken to promote peace include:

Encouraging and facilitating dialogue and communication between nations and groups in conflict

Addressing the underlying economic, political, and social issues that contribute to conflict

Promoting and protecting human rights

Providing education and resources to promote understanding and tolerance

Supporting and building strong, inclusive communities

Encouraging and participating in peaceful conflict resolution and mediation.

It's important to note that peace-building is a long-term process and it is unlikely that world peace can be achieved overnight.
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

aurelVZAB wrote:How we get to Mars with Mercury-Quantum engine ?
70 days if it works. :o

Meanwhile, see this.

The USS Enterprise (Galaxy class: NCC-1701-E) can reach Warp 9.8 flat out and get to Mars in less than one second. Not a good idea. “Mr Scott, can you give me Warp 9.8?”. “Aye I can that captain.”. “Make it so. Stop!” :)

With the Impulse engines at half throttle (half light speed) Mars is just 10 minutes away, giving the crew enough time for a brew. :D
BasicCoder2
Posts: 3814
Joined: Jan 01, 2009 7:03
Location: Australia

Re: Chatting with AI

Post by BasicCoder2 »

deltarho[1859]:

My impression of that response is the program has just collected human statements about how to work toward a more peaceful world without any new insights. It seems to me a sophisticated data base query program using natural language to input and output data. That data base is probably the internet. Essentially it might have typed into the google search engine "how to bring about world peace" and used the resulting links to generate a set of sentences to form a reply without itself having any understanding of the question or the answers.
Last edited by BasicCoder2 on Jan 25, 2023 15:26, edited 1 time in total.
dodicat
Posts: 7698
Joined: Jan 10, 2006 20:30
Location: Scotland

Re: Chatting with AI

Post by dodicat »

The Enterprise bends space time just enough to bring Mars really close to it, then jumps the gap then restores space time.
Photons do this. The Enterprise is just modelled on a photon, with a Scottish engineer to make sure all operations run smoothly, he is a hero in his own space time. He avoids any inter galactic politics, tries to avoid going out and about with the Captain and crew, knowing full well they always kick up trouble. He is happy in his own engine room.
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

@BasicCoder2

I agree. Currently, it is a pseudo AI as with quantum computers, both of which have some way to go but are not that far away from being fully operational or near to it. I'm guessing that 2035/2040 should see both of them being worthwhile additions to our technology. Quantum computers on the desktop? They are well into the future.

@dodicat

I agree. Scotty used to make me chuckle when he always overestimated how long it would take to do a job only to be told to do it in half the time, knowing full well he could do just that. Great character. :)

By the way, Voyager is faster than the Enterprise and can reach a speed of warp 9.975, but only for short periods otherwise the engines overheat.
thrive4
Posts: 42
Joined: Jun 25, 2021 15:32

Re: Chatting with AI

Post by thrive4 »

dodicat wrote: Jan 11, 2023 12:39 I think as age creeps up in us both, we have taken diametrically opposite paths.
If I saw a drone pressing my doorbell I would throw a net over it, capture it, strip it and sell the spare parts.
Same frequency here, Jungian or not... :?
'They say the future... it's on a microchip
Don't you know we're all on the same wooden ship'

From the aptly named song 'information overload' by
Living color:
https://www.youtube.com/watch?v=phcNq-K81R4
aurelVZAB
Posts: 646
Joined: Jul 02, 2008 14:55
Contact:

Re: Chatting with AI

Post by aurelVZAB »

Yes i know that band called Living Colour ..on or two songs

hmm Impulse engine ...it is not clear to me what that might be ?
like Hall engine or high oscillation microwave to create trust
ok to much for my small brain.... :roll:
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

I put this to my preferred search engine, DuckDuckGo [Privacy-Focused]: “AI is not Artificial Intelligence”.

There are many people 'out there' who reckon that is true. I read a few and considered my take. I found one close to mine: Artificial Intelligence can Never be Truly Intelligent

I am going to mention Jungian Typology, again. We have four functions, Thinking, Feeling, Sensation, and Intuition. The first and second are polar opposites, with the third and fourth being polar opposites. Graphically they are usually shown depicting their polar opposition. However, I draw them differently:

Code: Select all

         Thinking
Feeling             Intuition
         Sensation
I do this to allow the two rational functions in the y-axis and the two irrational functions in the x-axis. What we have then is vertical thinking and lateral thinking; an overlap of Carl Jung and Edward de Bono.

My take is simply that AI presented to us is based upon vertical thinking and seems to be bereft of lateral thinking and is therefore not intelligent. Intelligence requires an irrational aspect. The above link mentions a lack of feeling, but does not mention a lack of intuition.

Can a machine have an irrational aspect?

Many years ago, some of us were beavering away writing 'machine code'; a phrase rarely used nowadays. We could tot up the cycles of our mnemonics to determine how long a piece of code would take to execute. Those days are well gone. Modern CPUs predict what is expected next and are allowed to get it wrong. Too much leniency will slow things down. No leniency will slow things down. There is an optimum level which gets most out of our CPUs. So, we are now talking optimization and that may vary generating a level of uncertainty; giving us an irrational aspect. Optimization is then intelligence. Modern car engines also use optimization, and it would be hard to determine what our engines are up to at a given time.

So, there is some intelligence in AI, but we are only scratching the surface, and it may be quite a while yet before AI is actually Artificial Intelligence. If an AI result is got via an irrational aspect, how can we prove that is more or less intelligent than a result got by a human via an irrational aspect? Perhaps we should call it Heisenberg Intelligence? Nah, lateral thinking will do. :)

I disagree then with the link and many others reckoning that it will fall short. We just need uncertainty in addition to certainty. With uncertainty, we open the door to imagination. “Imagination is more important than knowledge.” (Albert Einstein supposedly)

Will we ever see a Mr Data? He was found by Starfleet in 2338. We will have a Mr Data long before that. :D
BasicCoder2
Posts: 3814
Joined: Jan 01, 2009 7:03
Location: Australia

Re: Chatting with AI

Post by BasicCoder2 »

The Chinese room argument confuses the system with the individual acting as a component in that system. The system can be said to understand Chinese even if the human acting as a simple mechanism within the system does not.

I would suggest that when some one thinks about an intelligent entity they also imagine it is sentient as a requirement for its behavior. A description of sentience (being conscious, feeling like a non physical entity behind the eyes) has not yet been made in a way that everyone would say "Of course, that is what it is."

Ultimately we can only talk about behaviors and one thing a program can be is a theory of what produces a behavior called "thinking", "feeling", "intuition", "sensation" and all the other words used to describe our inner mental world. With the output of the program can also see where it fails to behave as if it has those mental attributes. So for example thanks to fast powerful computers sophisticated artificial neural nets can act as powerful pattern recognizers. But that is all they are, pattern recognizers. And they don't recognize patterns the way we do and thus can fail catastrophically in a way a human recognition system would never fail. Thus we can confidently say as a theory expressed as a program that in some way they fail to explain how the human recognition system works.

https://newatlas.com/good-thinking/faci ... n-clothes/

https://spectrum.ieee.org/slight-street ... algorithms

If you program a machine to make a sighing sound when patted and a wail when hit does it really feel pleasure and pain. I would say not. Pleasure and pain is a computation of some kind independent of sighing or wailing.
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

BasicCoder2 wrote:in a way a human recognition system would never fail.
Here is the classical representation of Jungian Typology:

Code: Select all

          Sensation
Thinking             Feeling
          Intuition
The y-axis is Perception and the x-axis is Judgement.

What amazes me is how an event can be perceived very differently by two people, indicating the human recognition system can lead us 'up the garden path' before we employ any judgement via thinking or feeling. I am not talking about optical illusions here, but every day events.

AI's perception is far from perfect, so it needs fixing. Fixing human perception may never happen.

Judgement should be an AI choice: “Do you want a thinking judgement or a feeling judgement?”
BasicCoder2
Posts: 3814
Joined: Jan 01, 2009 7:03
Location: Australia

Re: Chatting with AI

Post by BasicCoder2 »

Although Jungian typology may be a way to categorize personality types from observing human reactions to given situations it doesn't really explain the mechanisms of thought which is what you need to know to program an AI to react the same way to the same situations. Describing and categorizing behaviors is not providing any information on the mechanisms that produce the behaviors.

By thinking judgement I assume the steps are clearly seen to be true as apposed to a feeling judgement which just feels right but the reasoning can't be spelt out step by step. Intuition is a bit like a neural net which gives the answer that "feels" right but it can't tell you the reasoning behind it only that it is similar to some previous experience that had some good or bad outcome.
deltarho[1859]
Posts: 3686
Joined: Jan 02, 2017 0:34
Location: UK
Contact:

Re: Chatting with AI

Post by deltarho[1859] »

@BasicCoder2

I agree, for the most part, with your last post. In Typology, feeling should not be confused with emotion, “but is solely concerned with the setting up of a subjective criterion of acceptance or rejection.”

I am not proposing we use Jungian Typology as a tool for AI. What I am suggesting is that we add lateral thinking to vertical thinking before we can say AI behaves similarly to how we behave. We need to add a level of uncertainty so that the “reasoning can't be spelt out step by step”; your words. PRNGs are mechanistic and can be spelt out step by step, but CPRNGs, being organic, cannot. AI, currently, is rational so falls short of Artificial Intelligence.

What we need is guys like Werner Heisenberg, Paul Dirac among others to get involved in AI as they did with Quantum Mechanics and not just Systems Analysts. Talking of quantum, I doubt that 'real' AI will take off until we have 'real' quantum computers. I have no doubt that 'real' AI will happen – the question is when. My current life expectancy is 87 years, but there is a 1 in 10 chance of 96, so I may see 'real' AI; provided I keep taking the pills. :)

Added: What is Fuzzy Logic in AI and What are its Applications?
Post Reply