17 Mar 20
1 edit
https://phys.org/news/2020-03-qubits-room-temperature.html
They haven't yet made an actual useful quantum computer from it because its still in just the research stage but hopefully they will as that would be a possible way of not only making useful quantum computers but ones that can operate very cheaply!
17 Mar 20
@humy
It all sounds nice, but what can a quantum computer actually do to improve our everyday life, and can it do it better than a standard computer? (besides cracking passwords and stealing private stuff)
17 Mar 20
4 edits
@bunnyknight saidI once explained in another thread that nearly all of the often said kinds of tasks (such as cracking passwords etc) a quantum computer could do (which a classical computer cannot do, or at least no nearly so fast) would have extremely little if not absolutely no benefit to humanity but with one notable exception; Helping with computer simulations of quantum systems and quantum scale things, such as how protein molecules fold or how superconductors work etc. For those things, quantum computing would be able to massively boost the speed of research into better superconductors and better solar cells and better medicines etc and even better quantum computers themselves! And, in that way, very significantly help huminity. The whole of the rest of what is said what quantum computers can do for us is just worthless hype.
@humy
It all sounds nice, but what can a quantum computer actually do to improve our everyday life, and can it do it better than a standard computer? (besides cracking passwords and stealing private stuff)
17 Mar 20
@humy
So which would have a greater chance of becoming self-aware and self-evolving: a quantum-computer or a classical super-computer?
18 Mar 20
@bunnyknight
I think they will be hybrids, QC + classical super comp to make something better than either one alone.
18 Mar 20
5 edits
@bunnyknight saidQuite simply, a quantum computer, at least by itself, has absolutely NO chance of ever being what we vaguely call 'self-aware' or 'conscious' with 'mind' or even what we would normally call merely 'intelligent'! A quantum computer would basically be an extremely powerful but nevertheless extremely dumb calculator!
@humy
So which would have a greater chance of becoming self-aware and self-evolving: a quantum-computer or a classical super-computer?
So only a classical AI computer can have the attribute of true intelligence although such an AI may still find a quantum computer useful to speed up some of its important calculations.
Think of that analogies to:
If the AI is analogies to a human then the quantum computer is analogies to his pocket calculator; Yes, his pocket calculator can do certain useful calculations zillions of times faster than he can, but, its still completely stupid and only the human, not his calculator, has the real intelligence.
An ant would have more general intelligence than any quantum computer no matter how advanced at least providing that computer computer isn't hybridized with a classical computer.
I am far far less certain of the potential of possible quantum-classical computer hybrids for AI; I am afraid I can only make wild speculations about that.
18 Mar 20
@humy
So has any genius yet figured out how to reverse-engineer the human brain? Another words, how to simulate the brain processing unit using a digital computer?
And do they at least have any idea how the brain stores memory?
18 Mar 20
3 edits
@bunnyknight saidThe answer is no to all there.
@humy
So has any genius yet figured out how to reverse-engineer the human brain? Another words, how to simulate the brain processing unit using a digital computer?
And do they at least have any idea how the brain stores memory?
Although some partial understanding has been achieved.
It so happens I plan to do some serious research on advanced AI but first I must spend several more months finishing my book on the research I did into statistic analysis of data and then get that published before moving on to more interesting things.
18 Mar 20
@bunnyknight saidWith regard to memory they're making some progress, but to draw an analogy with cosmology they still haven't worked out whether the Earth is flat or not. Different types of memory are connected with different parts of the brain. For example, I'm typing this and my ability to type it is connected with procedural memory and the cerebellum, but autobiographical memory seems to be stored in the temporal lobes. Place memory seems to be stored in the hippocampus. The relevant Wikipedia page is:
@humy
So has any genius yet figured out how to reverse-engineer the human brain? Another words, how to simulate the brain processing unit using a digital computer?
And do they at least have any idea how the brain stores memory?
https://en.wikipedia.org/wiki/Neuroanatomy_of_memory
18 Mar 20
@DeepThought
I can think of another example: Using a standard computer we can now render a simulation of reality (light, shadows, refraction, gravity, physics of motion, deformation, ... etc) with such realism that it looks even nicer than reality.
Now simulating an organic brain could be a million times harder, but should still be possible.
19 Mar 20
1 edit
@sonhouse saidI don't see how such a report could possibly be correct. Here is why:
@humy
One report I read said when they get operational they will need thousands if not millions of Qbits to beat the coming Exoflop machines coming out next year.
https://www.scientificamerican.com/article/quantum-leaps-in-quantum-computing/
"...In principle, a 300-qubit quantum computer could perform more calculations at once than there are atoms in the observable universe...."
I know advances in classical computer processing power are being made all the time but no where near THAT processing power!
19 Mar 20
@bunnyknight saidThe reality simulations your talking about (probably) use finite elements to produce the simulations. This is fine and works really well, but each element is large compared with a molecule (or even a living cell).
@DeepThought
I can think of another example: Using a standard computer we can now render a simulation of reality (light, shadows, refraction, gravity, physics of motion, deformation, ... etc) with such realism that it looks even nicer than reality.
Now simulating an organic brain could be a million times harder, but should still be possible.
The problem is the brain simulation has to take place pretty much at the molecular level, unless there's some theory that allows one to average over this. Are memories connected with the strengthing of particular synapses, or are they stored inside the cells somehow? Either way a simulation of a human brain would have to contain 85 billion neurones and as many non-neuronal cells. Each neurone is pretty complex so we're going to run out of machine memory fast.