7 ms·

There seems to be some disenchantment about quantum computing in the comments already, but I'll add my take: This sounds similar to the AI craze that took off

by parsimo2010 7y ago

There seems to be some disenchantment about quantum computing in the comments already, but I'll add my take: This sounds similar to the AI craze that took off in the 80s (or was it the 70s?) where we were making such great progress that general AI was just around the corner. Tons of people got on the hype train and then we just never delivered. Investment dollars fell off, and we got the "AI winter." But then a few advances were made such as deep learning, and another craze has taken off with AI (now called machine learning) with slightly different expectations. It sounds like quantum computing promises a lot, and we have mathematical proofs that show that a proper quantum computer could do great things. Right now we are making tons of progress, and money is flowing in. But eventually I think that progress will slow down considerably before we reach the level of every corporation having a general quantum computer in their server room, and investment dollars will fall off. Then a decade or two later another breakthrough will happen and the hype train will start rolling again. Edit: I guess what I wanted to get at is that you probably shouldn't call people stupid for investing in promising technologies. The bubble will probably burst, but a venture capitalist probably shouldn't hold back their investment money because they think the "real" breakthrough will happen next decade. If real breakthrough is happening now, but the VC waits, they may have missed out on one of the most profitable opportunities of their lifetime.

jerf 7y ago
Quantum computing may someday be amazing, but at the current rate of progress, in 5 years it'll still probably be a side show. There's a small-but-reasonable chance of some sort of massive breakthrough that will make them suddenly jump in power, but I wouldn't call it a good bet from a financial perspective. (Bear in mind, for instance, that while Google may have demonstrated a 53-qubit machine, it's still deeply limited and they had to construct a problem for it to demonstrate quantum supremacy on. It isn't even an arbitrary 53-qubit machine yet.) At the moment, I see no clear path towards monetization for these startups yet, in any time scale that makes sense for a startup. AIUI it's not like there's even a ton of algorithms that you'd want to run on a quantum computer yet that are known to be clear quantum wins. For instance, breaking encryption requires a qubit per bit of encryption, IIRC; it's not just that "quantum computers break encryption" but "4096-arbitrary-qubit computers break 4096-bit encryption". 4096-qubit machines seem a long shot on the 5-year time frame, even interpreting current progress as exponential, let alone a 4096-qubit machine that can actually run Shor's algorithm.
- turk73 7y ago
  Five years? I was involved in a graduate research program in quantum computing in 1996! It really hasn't gotten too far in all that time and I'm pretty happy that I didn't make that my dissertation topic because it was just not timely or possible to make decent progress on any of it back then. We're more than 20 years on and it's still dubious as hell.

jerf 7y ago

Quantum computing may someday be amazing, but at the current rate of progress, in 5 years it'll still probably be a side show. There's a small-but-reasonable chance of some sort of massive breakthrough that will make them suddenly jump in power, but I wouldn't call it a good bet from a financial perspective. (Bear in mind, for instance, that while Google may have demonstrated a 53-qubit machine, it's still deeply limited and they had to construct a problem for it to demonstrate quantum supremacy on. It isn't even an arbitrary 53-qubit machine yet.) At the moment, I see no clear path towards monetization for these startups yet, in any time scale that makes sense for a startup. AIUI it's not like there's even a ton of algorithms that you'd want to run on a quantum computer yet that are known to be clear quantum wins. For instance, breaking encryption requires a qubit per bit of encryption, IIRC; it's not just that "quantum computers break encryption" but "4096-arbitrary-qubit computers break 4096-bit encryption". 4096-qubit machines seem a long shot on the 5-year time frame, even interpreting current progress as exponential, let alone a 4096-qubit machine that can actually run Shor's algorithm.

turk73 7y ago
Five years? I was involved in a graduate research program in quantum computing in 1996! It really hasn't gotten too far in all that time and I'm pretty happy that I didn't make that my dissertation topic because it was just not timely or possible to make decent progress on any of it back then. We're more than 20 years on and it's still dubious as hell.