Electronics Weekly Electronics Design & Components Tech News
But one way of looking at the latest claims for computers is to consider what they can’t do.
In the bath, I came up with this list of what computers can’t do:
Imagination
Originality
Kindness
Creativity
Gut feeling
Joy, Sadness, Ecstasy, Pride, Depression, Feeling of Ownership, Love
Physical Pleasure
Physical Pain
Mental turmoil
Appreciation of nature, animals, plants, flowers, sunsets,
Surprise
Ambition
Sense of Failure
Awe
Sympathy
Respect.
Anger
While it is possible, for instance, to programme a computer to respond that its reaction to being kicked is anger, that is not the same as it actually feeling angry itself.
For as long as computers remain lacking in so many spheres of human experience, it seems absurd to equate computers with brains and say computers will take over one day.
You wouldn’t want the world ruled by an entity lacking all the capabilities in the list above.
Brains and computers are not the same thing. We should look a little leery at those who suggest they are
DM.. There’s quite a lot on your list that I can’t manage either. Not telling which though.
Sceppers
Just saw this, AI outsmarting human experts at training robots …
https://www.youtube.com/watch?v=d5mdW1yPXIg
AI is useless at being a mobile human. It’s a great simplification of knowledge, but it lacks explanatory powers in the sense that it can’t explain to others, in a convincing manner, how it decided to ‘select’ and offer the answers that it gives.
There are semantic gaps in it’s ability to communicate because it hasn’t negotiated the terminology, jargon and ‘vocabulary C’ that is required to maintain that air of underpinning knowledge and understanding required to talk with other experts in the field.
Heck, we already have difficulty with that ourselves, trying to ‘educate’ the conspiracy theorists (or is it the other way around).
It’s the incremental inclusive improvement issue compare to the ‘catch up’ approach that is an important part of the ‘AI as equal’ problem. We haven’t solved equality diversity and inclusion for humans yet. what hope is there for AI.
“We haven’t solved equality diversity and inclusion for humans yet. what hope is there for AI.”
Why would we want to solve this ? Humans are by definition all different, and hence can never be equal. AIs actually can be if they run on the same hardware using the same software.
Yet.
We should call the best AI for what it is, which is ML or machine learning. We are still a very very far away place from what’s still being talked about.
Quite, the over, and uneducated, use of the AI term is infuriating.
This guy (Mustafa Suleyman), based on his decades working on AI, argues that AI already has empathy and creativeness.
https://www.youtube.com/watch?v=KKNCiRWd_j0
This youtube was posted 2 days before the above article, so perhaps that is not a coincidence?
Anyway, you won’t be surprised to hear that I agree with Suleyman.
Well DontAgree, as Mandy Rice-Davis said: “He would say that wouldn’t he?”
If someone kicks me it causes some neurons to fire which cause some other neurons to fire. Eventually this sequence will stop – unless the kicking continues.
Is the argument here that:
This collection of neurons could never be replicated – Moore’s law will never get us to a big enough computer
There is something about neurons that couldn’t be replicated by a computer
At some point the neurons activate some mysterious anger centre that doesn’t involve replicable neurons
You could build a computer that exactly replicates a human brain but you wouldn’t have any of the magic anger liquid to pour into it
If I want to stop a computer ever being angry don’t I end up with there being bits of my brain (the anger bit, the sympathy bit,… ) that aren’t subject to the laws of physics and so could never be replicated but can interact with the neurons that could be replicated?
The neurons are just a basic communications and simple logic system – all the ‘magic anger’ is performed by the quantum computing in your brain – what we call consciousness.
I’ve a feeling that there is some word shuffling going on here. Whether it is ‘quantum computing’, ‘consciousness’ or ‘soul’ – the fundamental question is “Is it reproducible?”.
If it can’t be reproduced then you have to describe how it interacts with the bits that are – the neurons or whatever.
If it can reproduced then we then we just need to know how far along a line between what we have now and a completely reproduced human brain we need to get before we get anger, etc.
https://www.mit.edu/people/dpolicar/writing/prose/text/thinkingMeat.html
Ha Ha zeitghost that’s totally brilliant. Of course we’re not told what the snotty aliens doing the dialogue are made out of. If it’s metallic you’ve destroyed my case.
Neither metal nor carbon based.
https://www.newscientist.com/article/mg24632800-900-is-the-universe-conscious-it-seems-impossible-until-you-do-the-maths/
https://www.scientificamerican.com/article/is-consciousness-part-of-the-fabric-of-the-universe1/
And various other articles. This whole concept is gaining a lot of traction in past few months.
Turing test, if you can’t tell the difference between a fake emotional response and the real thing then yes computers can do all those things. They are not there yet, but I am convinced they will be at some point.
I don’t agree, DontAgree, you can programme a computer to produce a particular response to a particular input but how can a collection of transistors and wires actually feel anything? That’s the difference.
What is the human brain other than a set of logic elements and quantum storage devices ?
But biological Mike not metallic. Trying to replicate a biological entity with a metallic one is unlikely to succeed.
Train a neural network with 100,000 happy photos and 100,000 sad photos. It should be able to distinguish the difference. Most AI are just simple, very small models of parts of the human brain. Does it “feel” happy or sad? No, somewhere some floating point variables will have values corresponding to happiness and sadness.
Achieving anything close to human intelligence would take a lot of silicon and a huge amount of energy though. We can’t assume that Moore’s Law will get us anywhere close.
Really depends how you define Moore’s Law. The key element missing is body temperature quantum storage. We know all our longterm memories are stored in the microtubiles, we just don’t understand how they work, and how we use that data to influence our actions and decisions. Quantum entanglement is one possibility, but in theory that should lose the data, which we don’t.
Room temperature super conductors are advancing rapidly and can remove the heat limit on microprocessor performance. I estimate thirty years. However intelligent a computer I view it as a tool like my slide rule, just a better tool.
If you cannot tell in anyway the difference between a machine and a human then the discussion becomes very wishy washy, like ‘it has no soul’ … but you would only know that after stabbing the humanoid and finding that your blade is met with a ‘clang’ or a ‘zap’.
Yes a computer can do some tasks better than humans, DontAgree, but do you think you will ever see a computer being kind, showing generosity of spirit or demonstrating genius?
Ultimately I am convinced computers will get there, but perhaps not in my life time.
Have you ever used AI? I played around with image generators and it is amazing. Definitely implies imagination and creativity, to take just two examples from your list. The view “a programmer defines an output for every possible input” is, sorry to say, simply very outdated.
It also begs the real question, what are we beyond “a computer that generates an output based on input”, trained not by a programmer but by life itself? Where does imagination reside and spring from in our physical bodies?
I agree Martijn, the things computers can do are impressive. Whether they are creative or merely cleverly simulating creativity, I don’t know. But my point is we have gone down a wrong route in equating computers with brains. Brains are biological, computers are metallic and biology can give a vast range of feelings and emotions which a metallic machine never will.
“You wouldn’t want the world ruled by an entity lacking all the capabilities in the list above.” are you sure about that? Given how poorly things are going (wars, crime) with the human rulers, then maybe having a more logical and incorruptible approach, at least for some things, doesn’t seem so bad …
We should let the dolphins run things – they have most of those abilities in abundance.
The broad sweep of human history has been onwards and upwards DontAgree, although with many glitches intervening, as now, while being run by a machine with no feeling for life, liberty and the pursuit of happiness would be a little unnerving.
That is just your emotion, human fear of the unknown, talking … logically speaking we cannot issue a verdict until we try it …
I am not saying we should try it, I like to think of myself as human too, so my emotion is with you: don’t do it, but rationally ….
If the AIs are off-topic maybe they didn’t understand the issue,
They usually do understand as when you plot their logic paths, they generally follow the four most likely ones at each decision point. Usually two of the end results are pretty accurate and the other two less so. But occasionally just the one thread will head off into the weeds. I’ve seen this with several AIs so it’s a common problem.
People always quote “no originality” but in playing with these things a lot for the past six months, time and time again I see AIs go so far left-field of the issue fed to them that one has to admit they are being original – a bit like a human brainstorm.
> You wouldn’t want the world ruled by an entity lacking all the capabilities in the list above.
I thought we already were !!!
I’m waiting expectantly to learn whether Elon Musk’s robot can roughneck on a drill rig. It’s a simple repetitive task that requires vision, coordination, precision, and strength. There is plenty of time between pipe sections to recharge your batteries from a Diesel generator. If the robot succeeds the next more challenging task is to restock food on the grocery shelves.