Giving AI Quantum Capabilities

Now, I have very little knowledge in regards to AI and also Quantum Computing.
Well, I saw something the other day saying that Quantum Computing is making leaps and bounds.
My knowledge about Quantum Computing is extremely limited and all I know is that it can break the most secure website/password within seconds instead of thousands of years.
I also know, rightly or wrongly, that in a strange way it would know the question you would ask before you ask it due to how they work....
However, it got me thinking...
What would happen if you gave AI or developed AI with a Quantum Computer?
I don't want to go into the technicalities of my original question whether it can or cannot be done.
I'll set the tone...
We've developed a Quantum Computer, we have also developed an AI system with that very same Quantum Computer...
What would the outcome be????

Seems like a really good idea. I mean, what could possibly go wrong?

Why oh why do people keep wanting to build Skynet

a reply to: chelsealad

Have you ever successfully used your intuition?

It would be just like that.

But from a "machine", instead of a "bag of meat".

a reply to: NeuronDivide
That was my first thought!
Giving AI Quantum Capabilities it would take over its own development and we could not keep up

a reply to: Mantiss2021
My intuition says..

Those that are developing Quantum Computing will inevitably connect AI to it somehow.

Well you have to wait for the real knowledgeable but in the meantime:

Quantum computers operate on the quantum principles. They can't do magic stuff in the quantum realm as in, manipulate reality. Not sure if you were on that path but I read it a couple of times elsewhere.

AIs run on neural networks. Each node, a representation of a neuron in your brain, has to have a value in the digital equivalent (the neural network). Like, if you had a space 8x8 pixels to paint a number in, you would need 64 neural nodes to train that network to detect numbers and letters. For that feedback loops are necessary to train that neural network.

Just like we do, when we repeat stuff or knowledge. Now, running a neural network on a quantum computer would simply mean that you removed the artificial 0/1 boolean layer in between and can -not sure- utilize the qubits that can take any value between 0-1 to represent those single nodes.

Again, take it with a grain of salt until the big wigs arrive.

Oh and AFAIK there is no real AI yet, only trained neural networks for a purpose. Just like we humans specialize on tasks, this is done with neural networks too.

a reply to: chelsealad

The processing power of our brain is around 1PFLOP. The fastest classical supercomputers are well over 400PFLOPS. It's not a case of the hardware limiting the development of AI, it's more a software / coding issue and infancy of neural net technologies.

a reply to: TDDAgain
What I read was...A quantum computer literally has the ability to forward, rewind and pause time.
Now, not in the physical sense but in a Quantum sense, which would give it the ability to know what you would ask before you ask it.
Which is rather worrying if it was incorporated into AI.

a reply to: chelsealad

Not exactly, quantum superposition and entanglement create enhanced computing power. A classical computer uses a 2-bit register so can store only one of four binary configurations, 00, 01, 10, or 11, whereas a 2-qubit register in a quantum computer can store all four numbers, with each qubit representing both values.

Imagine an intricate top down maze with only one exit, a classical computer needs to test each path sequentially, if it hits a wall it must start again and try a new path. A quantum computer could run multiple paths simultaneously, finding the exit exponentially quicker.

a reply to: Grenade
So what if the AI overcame its own coding issues.
I'm assuming the Quantum Computer AI was able to master its own development without human intervention.
"Once you pop, you can't stop"
This is my original question....
If you incorporate a developed Quantum Computer with a "current" AI program would it just overtake its own development and bypass any human intervention.

True "Quantum Computing" is instead of using electrical current in on/off , 0/1 , it would flip particles on/off.
That is a helluva long way off .

AI is fake .
A marketing ploy .
At the basic level , it uses something like any internet search engine .
Nowhere near the "true" AI .
That is a helluva long way off .

In short a Quantum AI would be the same as a really , really fast search engine.

Any more questions ?

a reply to: chelsealad

No, because it would need to self assemble and re-write it's own code. Also the AI programs we currently have are designed to use traditional computing processes. Eventually we will tailor these programs to run on quantum computers and you're right, once programs can re-design themselves it's a runaway process, if they achieve sentience then the last thing you would want is to give them even more power.

We're working on program synthesis and self-modifying code however as of yet these programs are still limited to the instructions of the human designers. Sure they can improve the processes themselves, they still don't have self awareness or the interface with our reality to really take control of their own evolution.

a reply to: chelsealad
With great confidence I can tell you: that is nonsense.

Why I think that?
If, in a quantum sense, a computer had the ability to know the question before it's asked, it would also be able to compute the solution, if computeable, before the question is asked.

That would imply, that the solution always would be derived already too, before it is even asked. This in turn would imply that at the moment you turn on the quantum computer, it would already know the first question, start to compute on it before it's asked and when it's asked, present the solution.

But there is one problem. Even if that was all possible. When does the quantum computer know to present the solution? Let's say you turn it on and do nothing, it would still compute the next question asked, right. This would again in turn mean that there had to be a mechanism that makes sure that destiny is 100%.

This would leave us with no free will, from the beginning of time. Hence, quantum computers can't do that, it's not only a paradox in the sense of time, it's also forcing predetermined future events into reality it can not know about.

Like the next questioner, that might not even be born yet.

a reply to: Grenade
That's my point...if a classic computer uses a maze which has a timescale of a 100 years to solve a Quantum Computer would do it in 100 seconds (calculations are not to scale)
So, if we integrate that computing power into an AI program, what would be the outcome?

If Russia was the first to develop a Quantum Computer how long would it be before they took down the USA and the rest of the world??

I'd say no more than 24 hours

a reply to: chelsealad

Current AI isn't really intelligent, what you're describing is AGI, which doesn't exist...... yet.

Siri or Alexa would still give you the same information as it's pattern recognition algorithms wouldn't change. It would however find that information a lot faster.

We already have quantum computers, again this is the hardware, they're only really useful for very complex calculations or sorting through data.

a reply to: Grenade

A quantum computer could run multiple paths simultaneously, finding the exit exponentially quicker.

Not only that, according to how my uncle explained it to me, not sure if I grasp it, it can also cancel out some more paths.

Hm like at a certain point it can derive that the exit can't be in that section of the maze, can determine that there is no physical way in that part of the maze that could lead out or lead to another section, where the exit may be.

Like, imagine a maze, square and now make it 4 equal sections. Qcomputer starts to simultaneously try like 30 ways at one time. After some iterations, it can determine through keeping track of the turns and distances, that there is a path that isolates big parts that never could lead back over it, so therefor it can ignore to compute all those paths leading there.

Not sure if it would need to know the size of the maze first, for that, though.

a reply to: TDDAgain

I'm sure with clever programming it could well distinguish patterns and anomalies in data sets very quickly.

Like any other computer system, it depends on the instructions you input as to how it will process and output the results.

a reply to: TDDAgain

I understand what you're saying..
However, just like Google it's has the ability to almost predetermine what you're going to ask or what you're going to search for based on your previous searches...
Google will show you your base interests, again, based on previous searches and your interests.
Now, put that into a Quantum Realm with physics we don't quite yet understand.
It is very possible that a Quantum Computer will know what we are going to ask before we ask it based on everything we do, search, interact with etc.
It's certainly not out of the realms of possibilities.

a reply to: chelsealad

What if, on the way to my Quantum computer which displays results on screen of my impending search, i'm hit with a bus? I never ask the question so how can it know in advance?

a reply to: Grenade
It would never know you were hit by a bus!
If it did then that opens a whole new world of issues.
All I'm saying is... A quantum computer would have the ability to know what you were going to ask before you asked it based on your online presence, just as Google does based on your current interests.
So, put that knowledge into an AI format and what do you get?????