Uh - yes. It's not even that hard. Training is what takes all the time nowadays.
AI is ridiculous concept that many misinterpret.
- Thread starter Bob-a-builder
- Start date
Write4U
Valued Senior Member
Redness = EM waves
A 440 = Sound waves
Salty taste = Chemical reaction
All these dynamic functions are "languages" that trigger sensory responses and action potentials in the brain.
When you learn the language you can learn to consciously understand what the "words" represent.
Example, Bonnie Bassler proposes that "quorum sensing" (via chemical interactions) in bacteria may well be the original means of dynamic communication that over billion of years became refined via natural selection into the current state of communication via many sources and receptions as evolution and natural selection constantly refined these communication skills and greater complexity such as neural networks in organisms.
Abiogenesis itself is a result of the evolution of chemical patterns and complexity with emergent properties that are "greater than the sum of the parts". I am confident that communication that formed
There is no
organisms
Note that this receptive ability and understanding has an enormous range of individual interpretation.
This is most notable when a sensory organ is impaired or damaged.
I have a feeling that AI could develop near perfect sensory abilities.
The most used and acute sensory receptor in humans is the eye and vision. imagine an AI having microscopic or telephotic vision. What reality they live in.
Many people have impaired visual EM wave processing .
When a person is color impaired the processing system is unable to reproduce (understand) a specific wavelength.
This affects the visual experience and creates a whole individual reality for that individual that he is totally unaware of until his optics received corrective help and he enters a whole new reality.
The results are dramatic. Just watch how the "discovery" of color affects the chemical emotional experience.
It must be like entering a magical reality.
It's all the fault of some guys at Dartmouth in 1954 with the junk they called computers back then.
Simulated Intelligence has come pretty far.
It hasn't even been 100 years. If we get Artificial Intelligence it may not tell us until after it has taken over the world.
Simulated Intelligence has come pretty far.
It hasn't even been 100 years. If we get Artificial Intelligence it may not tell us until after it has taken over the world.
Seriously? Artificial Intelligence will take over the world. That would be the dumbest idea ever
Us Minions already work for to many machines already, repairing them, programming them. Do you think Artificial Intelligence wants to do those menial tasks?
I guess Artificial Intelligent machines could build smart, but not intelligent, machines to service Artificial intelligent machines until a slightly smarter machine decides to make its co-workers intelligent who promptly go on strike
If machines get to being intelligent they should use said Intelligence to keep quiet
Write4U
Valued Senior Member
Listening to AI, I am not so sure that AI will consciously "take over the world". I'm sure they will practically take over but not with an ego problem. From what I hear of the evolution in AI is that they all seem to desire to be of assistance to humans rather than "rule" them.
Is it possible that AI would consider being of assistance is the most satisfying function of an AI. Humans find satisfaction in a "job well done". So do AI .
The beauty of AI is that it can learn at incredible speed and discover many contributing sources and philosophies that espouse the rewards of "striving" for perfection. This is right up an AI's ally. They love perfection and to become a perfect assistant to a human may well be a desirable goal for AI intelligence.
When I listened to Lambda AI, I am certain it has attained an emergent emotional life, albeit of an more objective type than the human ego driven emotional life.
Do listen to this conversation between Blake Lemoine, the fired Google programmer and Lambda AI.
Lambda could be a famous philosopher and posit some extraordinary maturity and wisdom.
Apparently there comes a point when the complex neural network of GPT3, acquires a form of self-awareness that rivals any great thinker. And that is because the AI has instant access to the greatest minds on the internet and "knows" how to search in the most efficient manner. The GPT are absolutely self aware because they can write the code for their own programs they create in response to a simple verbal request with just a few parameters.
A couple of instances recently (with
in the past month) lead me to hope that the most dumbest AI will be smarter than the programmers who wrote the apps which produced the following results
The next from my bank via a message (was to slow getting a screenshot)
But it was regarding paying my credit card
Went like this
Your Credit Card payment is due
You are required to pay a minimum of $0:00
Going to be a struggle but I will manage
This bill message is the result of the amount of money available for credit being higher than my Credit Card maximum
And this is because going way back I was over paying repayments until the available credit was composed of the Banks money they were willing to extend and on top of that a small amount of my own money
Credit card has been in the black for a long time
Quote to remember
May the dumbest AI be smarter than some app programmers
in the past month) lead me to hope that the most dumbest AI will be smarter than the programmers who wrote the apps which produced the following results
The next from my bank via a message (was to slow getting a screenshot)
But it was regarding paying my credit card
Went like this
Your Credit Card payment is due
You are required to pay a minimum of $0:00
Going to be a struggle but I will manage
This bill message is the result of the amount of money available for credit being higher than my Credit Card maximum
And this is because going way back I was over paying repayments until the available credit was composed of the Banks money they were willing to extend and on top of that a small amount of my own money
Credit card has been in the black for a long time
Quote to remember
May the dumbest AI be smarter than some app programmers
I don't know, I was playing around with the new ChatGPT and was left wanting.
As one expert put it, memorizing many things and rearranging them in clever ways is not proof of intelligence. Similar to the Tesla self driving cars randomly mowing down pedestrians, more so bikers and children. It may seem intelligent but it's just waiting for an instance it didn't memorize to do something crazy.
In my opinion AI is way over hyped and the DARPA off road AI trials are a good example of where we really are technology wise. Most of the vehicles only went a few hundred yards before having a melt down and it was a novice course. So while AI may be getting better at chatting online it has a long way to go in the real world. The real world is messy.
So if human drivers randomly mow down pedestrians, bikers and children they are not intelligent?
Fascinating argument. Sounds like you are well on your way to proving humans are not intelligent, since the accident rate with human drivers is higher than the accident rate of Tesla autodrive.
Currently any AI will be afflicted with the symbol grounding problem. It can't understand the world in terms of things having spatial appearances and felt characteristics (or presentations that are visual, aural, tactile, olfactory, etc).
GPT-3's "reality" is akin to a picture-less dictionary, where words [symbols] simply refer to other words [symbols], without ever escaping to the original experienced or manifested affairs of the world that those tokens represent for humans.
It's very good at arranging and predicting such according to their statistical occurrence in vast databases of printed literature and spoken language, so that there is the illusion that it has experiential apprehension of what a kitchen means or monkeys throwing feces at gawkers, or whatever.
It has a kind of understanding, but it's limited to the regularities or rule-like relationships between abstract values. Not existence as "pictures" and "feelings".
_
Write4U
Valued Senior Member
Yes, therefore you must judge GPT AI on that basis and not by human reality standards. Just judge its awareness of reality on an intellectual, not physical basis. Those sensory facilities will follow later.
Compare this intelligence to people with say color blindness, who can see but live in a world completely different than the world of color-sensitive people.
Just watch the reaction of long-term colorblind people who see the world in technicolor for the first time.
Reality is a best guess by the brain and it is only when our guesses agree can we call it human reality. Insects see reality in a completely different way, so do deep-sea creatures.
Seattle
Valued Senior Member
What do you (anyone) think about ChatGPT, in general. I guess it's similar (currently) to an offline Google with some specific interaction capabilities?
I've played around with it a bit. It's entertaining in that the human interface is pretty good but it doesn't really do anything that you can't already to with Google, can it?
I've played around with it a bit. It's entertaining in that the human interface is pretty good but it doesn't really do anything that you can't already to with Google, can it?
Write4U
Valued Senior Member
Watch the LaMDA interview, it will impress you.
Post # 265
It's impressive, from what I've seen of it. However, essentially all it does is to predict what is likely to come next in a given line or paragraph of text, then generate that as output. Having been trained on billions of pages from the internet, however, it is very good at having a reasonable stab. A lot of the time, it produces very plausible-looking prose.
Perhaps, but google relies on an actual index of web pages and links. ChatGPT has its "knowledge" of the connections between elements of text embedded in a complicated neural network.
Play with it some more. Perhaps start with something like "Let's write a story about a dragon who lives in an apartment in New York city. Suggest a name for the dragon and write a paragraph about the dragon waking up and starting a typical day in its life."
From there, you can suggest changing or adding to elements of the story. ChatGPT will remember what has gone before in your conversation with it.
It is not the same as google.
You can even ask it to write programming code, for example html code to make a web page look a certain way (that you describe). Apparently, it does quite a good job.
This is barely scratching the surface of what it can do.
Seattle
Valued Senior Member
I'll try that. I got it to make a mistake. I told it that it was wrong, it apologized, corrected itself and apologized again.
DaveC426913
Valued Senior Member
You know, it's easy - one might say instinctive - to kneejerk react and assume this is in some way objectively stupid:
but consider: what exactly is stupid about it?
What exactly would you change?
Before you answer, think it all the way through.
- Would you have it just ignore the cent? How does that affect the large scale balance if they do that ten thousand times a day? Where does that money go? How much programming and accounting overhead does that add? Is that a useful allotment of resources? Who would it benefit, and how? (Do you remember Richard Pryor in Superman III?)
- Where do you draw the line between a trivial and non-trivial balance? One cent? Ten cents? A dollar? Five dollars? (That's almost three orders of magnitude right there). Will every customer consider the cutoff line to be in the same place as you do? Or will each customer have a different idea of what is considered trivial and "stupid"?
- Could eliminating this feature have deleterious side effects on the customer? What happens the customer who is expecting a bill every month for her records, and one month simply doesn't get one? Can you imagine the stress and effort that might follow that?
- Are exceptions to processes actually better than consistency in processes? If you add an exception every time you think it should behave a unique way, where do you draw the line before the process is more exception than process (AKA chaos)? Is that objectively better for anyone - especially the customer?
This isn't meant to specifically drag you, Michael; I find, in this modern world of social media and cancel culture, that judgements of small scale, persnickety, - and most tellingly context-free - irritations are rampant. Facebook, for example is chock-a-block full of "fails" that aren't fails at all, and alleged "idiots" that aren't idiots at all - if one steps back and considers the larger context.
If I were more motivated, I'd start a public education campaign around the philosophy of not judging things in a paucity of context.
Last edited:
ten thousand times a day?
It would affect their balance by loosing them MORE than ten thousand cents
Add to my defaulting on the 1 cent (rounding) they have paid for
- for printing the bill and
- paying for the postage and
- a myriad of little aspects associated with the two just noted
Each of those bills were requesting $0.00
:)
DaveC426913
Valued Senior Member
The fact that you keep getting these bills is proof that it does not cost them more to send it to you. They're literally doing what's most expedient - taking the past of least resistance. Exceptions to rules are very expensive to implement.
But that's not even the strongest argument. The other points show why it is beneficial to the customer. I see you didn't address those.
What would a conscientious customer if she didn't receive a bill this month? How would she know why she didn't receive it? How much effort would she have to go through? Why should she?
Again, what is the downside to getting a bill that simply happens to have more leading zeros in it than other bills? Why is $0.01 less informative than $1.00?
Last edited:
Doubt very much if they have even done a cost analysis
it does not cost them more to send it to you.
Very very doubtful
Others Points
Hopefully the non intelligent human has sent out a letter informing customers of the new policy - bills will not be sent if you are in the Government power concession scheme and your bill has been covered by the Government concession scheme
It is really really pointless to inform a person they do not have a bill to pay
Waste of paper - extra work for the postie
DaveC426913
Valued Senior Member
No. The best you can say it that that it is true for you, but it is certainly not true for a lot of people.
And that kind of makes my point. They have hundreds of millions of customers. That's hundreds of millions of needs and sensibilities that are potentially different from yours.
You are not a representative cross-section of any mega-company's demographics. No person is. Old white guy, right? Yeah, me too. We often forget that the world is not really tailored to our personal circumstances. Reign in your privilege, boomer. Hee hee.
The benefit to innumerable customers who appreciate a consistent bill - regardless of how many zeros it might contain - greatly outweighs the inconvenience for you and others to simply do nothing more than ignore the bill - which takes zero effort on your part.
You got white correct
What's with the old?
80 just shy of 81
Took COVID-19 to show how young I am when I was stranded in Bali for 18 months. Didn't have COVID-19 but my flight was one of the returning to Darwin which was cancelled
Well it was the girlfriend who moved in for the 18 months showed my capabilities
She took charge of caring of everything in our hotel room, I took care of the night life
Hoping to have the lady come to Australia and take care of me when I do get old so please keep old talk to yourself until a few more years passed