4 ms·

We are not limited to run AMD64 cores. We can also have GPU cores and when some problems are really important to solve, we can run them on FPGA or create an ASI

We are not limited to run AMD64 cores. We can also have GPU cores and when some problems are really important to solve, we can run them on FPGA or create an ASIC for it. Also we can expect that there would be computing centers what resources the mere mortals can access as a service as they can access the resources of the computer clusters today. Very likely universities fill have their own quantum computers. Bigger ones first, then smaller ones. I think that the situation is not that depressing but it definitely will create some disparity in the beginning.

akerro 11y ago
In 20years we will have AMD128 and some ASICs for personal usage, Google will have something what will be 100M faster than quantum computer, it will be 100M * 100M faster than quantum computer. It will contain more computing power than rest of the world together. Every next generation of computers will be produces in shorter time than the last one, and will be more powerful than the last one. Cost of producing next generation computer will be cheaper then producing ASICs for us. Are bitcoin miners the more popular ASIC? >Very likely universities I don't mean no one else won't have or rent any, sure they will, there are a lot projects that will need it, we people won't have any. I think I presented my opinion too negatively :<
- eveningcoffee 11y ago
  I don't mean no one else won't have or rent any, sure they will, there are a lot projects that will need it, we people won't have any. If we look at the history of computing, then yes, one could be so pessimistic, but would it be really true? If there will be a breakthrough in the QC, then there will be sudden economic motivation for people (companies) to use quantum computers. This will generate the need for people who could work with such computers. This will again generate motivation for universities to train such people and the need to access quantum computers. I think that this will happen much much faster than it happened with the classical computing. Of course I also believe that there would be no personal QC any time soon if this is what you had in mind. Edit: I also do not understand why you are down voted. I think that it is an important perspective.
  - TheOtherHobbes 11y ago
    Not a downvoter, but I think there's a practical limit to how useful the data collected by Google is - and the company is probably close to that limit already. I'm not sure QC can change that. The only way to get more value would be total 24/7 Orwellian surveillance, and I don't think that's going to be a popular option. QC for crypto is a no-brainer. QC for anything else, including data mining/ML, is a much fuzzier prospect. I'm not sure anyone really understands what the practical applications could be, never mind how to use QC to make them possible. Any suggestion that you can take a warehouse full of web logs and tracking stats, give it a quantum shake, and have a few million pre-qualified addicted customers fall out is likely nonsense.
    - eveningcoffee 11y ago
      QC for crypto is a no-brainer. QC for anything else, including data mining/ML Problems in ML can be mostly proposed as an optimization task and if QC would works out, ML would be likely its main application. It would likely make an huge difference in solving a classification problems as you could train your ML model with much more data much faster. I t would also make the clustering problem much faster too as it also can be expressed as an optimization problem. Also there are many other optimization problems that would benefit a lot from the speed up (regardless if it is huge constant speed up or an asymptotic one).

akerro 11y ago

In 20years we will have AMD128 and some ASICs for personal usage, Google will have something what will be 100M faster than quantum computer, it will be 100M * 100M faster than quantum computer. It will contain more computing power than rest of the world together. Every next generation of computers will be produces in shorter time than the last one, and will be more powerful than the last one. Cost of producing next generation computer will be cheaper then producing ASICs for us. Are bitcoin miners the more popular ASIC? >Very likely universities I don't mean no one else won't have or rent any, sure they will, there are a lot projects that will need it, we people won't have any. I think I presented my opinion too negatively :<

eveningcoffee 11y ago
I don't mean no one else won't have or rent any, sure they will, there are a lot projects that will need it, we people won't have any. If we look at the history of computing, then yes, one could be so pessimistic, but would it be really true? If there will be a breakthrough in the QC, then there will be sudden economic motivation for people (companies) to use quantum computers. This will generate the need for people who could work with such computers. This will again generate motivation for universities to train such people and the need to access quantum computers. I think that this will happen much much faster than it happened with the classical computing. Of course I also believe that there would be no personal QC any time soon if this is what you had in mind. Edit: I also do not understand why you are down voted. I think that it is an important perspective.
- TheOtherHobbes 11y ago
  Not a downvoter, but I think there's a practical limit to how useful the data collected by Google is - and the company is probably close to that limit already. I'm not sure QC can change that. The only way to get more value would be total 24/7 Orwellian surveillance, and I don't think that's going to be a popular option. QC for crypto is a no-brainer. QC for anything else, including data mining/ML, is a much fuzzier prospect. I'm not sure anyone really understands what the practical applications could be, never mind how to use QC to make them possible. Any suggestion that you can take a warehouse full of web logs and tracking stats, give it a quantum shake, and have a few million pre-qualified addicted customers fall out is likely nonsense.
  - eveningcoffee 11y ago
    QC for crypto is a no-brainer. QC for anything else, including data mining/ML Problems in ML can be mostly proposed as an optimization task and if QC would works out, ML would be likely its main application. It would likely make an huge difference in solving a classification problems as you could train your ML model with much more data much faster. I t would also make the clustering problem much faster too as it also can be expressed as an optimization problem. Also there are many other optimization problems that would benefit a lot from the speed up (regardless if it is huge constant speed up or an asymptotic one).