http://www.wbresearch.com/tradetecheurope/hftisdead.aspx If you though HFT bots were tough competition, here comes quantum computing bots that will be 1000s times more powerful and faster. I just don't see how any human trader has any chance of long term success in this business.
Quantum computing is still at an experimental stage and the useful algorithms are still limited to academic interest. An obvious problem with the BEC approach is that you have to cool the gas close to 0 K, which requires a huge volume of hardware.
Algorithms are not a problem. They can be worked out, but doing so is at this stage still makes little sense. It's the hardware that is the main issue, but the progress is being made and it's possible that quantum computing will eventually become a reality. Perhaps within the next 10-20 years.
The article itself is bullshit. * But the core statement is true. * If you have a need and enough cash, buy the chip from D-Wave, and you should be able to solve your optimization problems via quantum annealing. So, there is indeed great potential of QC for finance/trading in the near future ... (risk management for institutions).
Do your homework before making statements ... Lockhead Martin is applying (sort of) quantum computing (the D-wave ship) already for 2 years to solve optimization problems. (Im saying sort of, because there is still a bit of academic dispute if the chip is a true QC, ... but my latest info is that many of the academics now agree that the chip shows at least some quantum behavior ... )
Since you seem to know the most about this topic among participants on this thread, let me ask you. How many years we are away from the time when Goldman or Citadel or some other big-wig decide to make use of this D-Wave thing for their HFT decisions? ballpark - 5yrs, 20 yrs ?
I'm not an expert on these things. But I do not expect that the main interest will be for HFT. Rather in risk management, which is basically just an optimization problem. The big ones will start to use the D-Wave chip as soon as its cost (and speed for their specific optimization problems), makes economically more sense than running a farm of FPGAs.
you can have the biggest gun in the world, but if you have no targets to shoot at-then it's all fucking useless.