Will we ever have true AI? ie that which is capable of free thinking...

I used to think it was impossible for computers to do random, but now im not sure we can either.

I think a computer that can evaluate stuff, evaluate whether its worth evaluating stuff, evaluate what emotional state is appropriate etc, would be pretty much up to speed.

if it is capable of free thinking would it let us know?

maybe not we but our descendents .. probably to some extent, unless of course we wipe ourselves out, which seems quite likely.

"I used to think it was impossible for computers to do random"

What does this mean?

if computers were capable of free thinking they would make more of themselves (clones)and take over the world

a computer can't produce a random number.  The only truly random events are quantum linked eg

https://qrng.anu.edu.au/RainBin.php

The day will come when someone in the world will have control of a robotic army,that will march into another country and take control.
and i would say we are about 20 years away from it.it would not be two difficult to invent.

true random numbers can be produced by natural sounds,eg a car backfiring,a baby crying,a toilet flushing,a door closing ,and its not difficult to produce your own random numbers with the sounds around you with the aid of a computer."hold on my kettle has just switched off and it triggered the number 8 which means i am to lay number 8 in the next race because all i lay is random numbers that are not detectable "

the next time i lay number 6 you can tell me to shut that dog up because he barks so often.

Theres a video game where a computer system gives you a turing test (i think its a turing test). You get asked lots of questions to prove youre not a robot and gradually it corners the argument and you cant.

Ended up agreeing that anything i claimed as proof of my humanity is not because of some soul (since that scuppered my free will) and is actually down to electrical movements of wiring within my brain. Check mate.

My vague comment about randomness was meant to be an example of something considered unique to humans. But if i pick a random number between one and a hundred theres probably all sorts of algorhythmic processes going on (its always 69 btw!) that makes me no more than a glorified computer.

So, in summary, i think we can design intelligence but it wont be anymore artificial than we are

they will run the show soon and Humans will be slaves, well that's what some say, but don't we all think its very sinister at say the least, I mean what the **** are we playing at??

Came across this on YT a ontribtion from Humanists I believe, an interesting insight into A.I,particularly later in the clip with expert input from US research gurus.Worth a watch for anyone interested in the subject and speculation of where it may lead us.

https://www.youtube.com/watch?v=owIh6RBJBOg&feature=youtu.be

Went to a academic lecture last week on the same subject locally but it was rather bland but the subject does open up interesting ethical and moral issues.

trouble is they can be made as nice or as NASTY as Governments want

they cant take over as they need us to supply their battery power, power to ther human race.

I was forgetten that man

Computers can readily produce random numbers if they've the right hardware installed:
https://en.wikipedia.org/wiki/Hardware_random_number_generator

The computer program isn't producing anything, it is just displaying the output of some natural random source.

Who said program? He said computer. I produced evidence that there are many ways computers can produce random numbers over many years.

Human designed AI may well be done in hardware not software, would it not then be AI?

The link in my post will allow anybody's computer to "create" random numbers.  That isn't the point.  It's like claiming you can do 100m faster than Usain Bolt and turning up at the blocks sitting on a motorbike.

I think that AI is generally thought of as being software based ie a written series of rules (that may self adapt) but, as you suggest, all sorts of developments, that are more like artificial life, could occur.

you are mistaking intelligence for consciousness, it is consciousness that seperates living things from non living things 
we don't know how we are capable of being of being conscious

actually its impossible

The speed at which computers will be able to compute by learning from experience is still worse than the human brain but may surpass this by 2025 according to the Professor I listened to last week,then it becomes a matter of who  is controlling who.

man will still be able to hypnotise a robot or a computer if he cant switch them off.They say that Hitler had the secret to hypnotise mass gathering's of people.

JC, from your link.. "These stochastic processes are, in theory, completely unpredictable"

Interesting. Surely using the phrase 'in theory' suggests doubt? And if they aren't certain about the unpredictability then in theory, theyre predictable and not random!

I probably should have said earlier that in reality it would be almost impossible to recreate the complexity of a human brain (even a stupid one) as it is almost impossible to predict computer generated randomness.

But i would say in theory, a computer cant produce random and in theory, a human brain can be recreated.

^ I might have ballsed that up

in random numbers made by a computer ,another computer will detect the randomness in less than .002 of 1 second but the second computer can not pick up sound generated random numbers because the numbers are related to sounds (random sounds )that occur by accident .

And it may surprise you that the human brain if asked to quote ramdom numbers from 0-100 will after a time fall into a pattern of repeat.

Like your strange posts about sound? :P

Well I'm quite happy to believe a correctly designed hardware random number generator is random, as the "theory" makes sense.

I mean you guys only exist, in theory, some of you (e.g. anxious) might as well be poorly written chat bots!

'free thinking'

As someone who has worked at the cutting edge of 'AI' for the last few years I would say it depends how you define 'free thinking'.

Certainly combining reinforcement learning with neural networks in artificial agents can create new strategies.

Back the 1990s this was used to beat the best in the world at backgammon. Watching the agents plays gave human players new winning strategies that they had never considered before.

Fast forward to 2017 now the GPUs are available the same tech has been used to beat the world's best at GO. Again the human player has learned new strategies from the strategies the agent used and is on a long winning streak against humans since losing to the machine.

So yes new ideas and strategies can be found by artificial agents.

Is this 'free thinking'?

Personally I dont think so. It is just the use of algorithms that are extremely hard to get right. Changing one or two parameters can lead to completely different results. They are basically down to brute force calculations doing a lot of clever trial and error to get the best solution they can find.

Despite what google/deepmind says Deep-Reinforcement Learning is not intelligent in my opinion. It needs to 'play' too many times to get good at a game compared to a human.

Will we not be faced with ethical or moral questions when AI becomes a reality ?

brassneck
Date Joined:    13 Feb 03
Add contact | Send message
04 Feb 18 20:33 Joined: 13 Feb 03 | Topic/replies: 11,312 | Blogger: brassneck's blog
The day will come when someone in the world will have control of a robotic army,that will march into another country and take control.
and i would say we are about 20 years away from it.it would not be two difficult to invent.
Rate reply:
| report block user

would take a lot longer than 20 years, anyway this is too expensive for anyone but nation states or james bond villians.

to some extent though this has already happened with pilotless drones undertaking attacks on other countries soil

you are mistaking intelligence for consciousness, it is consciousness that seperates living things from non living things 
we don't know how we are capable of being of being conscious

true, although we can't scientifically "prove" that we are conscious either, we only know it from experience. hence, "I think therefore I am"

Not inconceivable that a computer could imitate consciousness, I guess that is what the Turning test was all about- how will we really know whether it's conscious or not- we might just have to take it's word for it

terry mccann 05 Feb 18 09:27 Joined: 26 Jan 05 | Topic/replies: 7,820 | Blogger: terry mccann's blog
they will run the show soon and Humans will be slaves, well that's what some say, but don't we all think its very sinister at say the least, I mean what the **** are we playing at??

doubt they will be any worse than we are. humans are the worst species to ever walk the planet, ****g up everything as we go.

hard to see how robots will be any worse.

The merger of man and machine will mean that you will not know where one begins and the other ends.

i think sometimes that the world will end the day that man decided to control the weather,this will set of a chain of events that are not reversible ,and the universal laws will finally close the curtain on our opera before we even get a chance to hear the fat lady sing.and the way things are going i would not be surprised if the fat lady has already cleared her throat in preparation .

Man is still cutting down trees in the amazon,the ice caps are melting,and nuclear tests testing is not good for atmosphere,and there is a still a big hole in the ozone layer,
yet the mugs in power still think about how fat or thin their country purse is.

The impending reversal of the polar axis may dwarf any contributions from man made interference brassneck ?

i think the world will flip soon,hope we dont fall out the hole in the ozone.the earth flips every 10,000 years and its due to do  very soon.