Nvidia's Huang Says 'Very Useful' Quantum Computers Likely Decades Away (yahoo.com) 17
Nvidia founder and chief executive Jensen Huang believes "very useful" quantum computers are likely decades away, tempering expectations for the emerging technology. "If you kind of said 15 years for very useful quantum computers, that would probably be on the early side. If you said 30, it's probably on the late side," Huang said during Nvidia's analyst day. "If you picked 20, I think a whole bunch of us would believe it."
Quantum state? (Score:2)
That is far too optimistic (Score:2)
With the glacial scaling we have observed the last 40 years or so, we may eventually get QCs that can do useful things, but > 100 years is a more realistic estimate. We may also never get them, conventional computers may stay ahead (the claims of "quantum supremacy" are just lies), or we may get them as demos, but far too expensive per computation to ever be used.
Re:That is far too optimistic (Score:5, Insightful)
Think any proposed specific timeline is impossible to support or dismiss. The nature of this is that it will be advanced not by any currently known means, but by breakthroughs. The requisite breakthroughs might happen within 10 years, might never happen, might be fundamentally impossible.
Re: (Score:3)
Guy who makes "AI" chips disses quantum competitors in front of stock analysts. His stock price immediately shoots, quantum computer stocks dive.
Doesn't take a CUDA core to figure out this one.
My tech is red hot, your tech is doodley-squat. Indeed quantum computing gets real, his stock will nosedive with velocity never before seen.
Re: (Score:2)
or we may get them as demos, but far too expensive per computation to ever be used.
At the point where some QC publically demonstrates quantum supremacy, it will be used almost regardless of cost. For certain organizations, the word "budget" doesn't really exist in the dictionary.
Re: (Score:3)
This reminds me of a series of talks Google hosted on fusion power. They had at least two experts in the field give presentations on the state of fusion for energy, one of which was Dr. Robert Bussard. Dr. Bussard died in 2007, and this presentation he did was shortly before his death, so it must be something like 20 years ago now.
One "complaint" (for lack of a better term) Dr. Bussard had in his presentation was funding for his research. He had some limited funds from the US Navy as the Navy had interes
Re: (Score:2)
There is a point of diminishing returns though, more money doesn't always make technological progress move faster.
There are examples that show that quite frequently throwing money at a problem slows down progress.
Re: (Score:2)
Re: (Score:2)
You misunderstand. I am not talking a few billions here. I am talking effort so high that the human race cannot do it for more than a brief stunt.
shades of AI (Score:2)
10 years ago a useful version of AI wasn't expected for 20 years but then there were some breakthroughs and a huge amount of progress. I expect quantum will be much the same. Some key physical barriers will be overcome that results in vastly more qbits and fewer errors. Then there will be a merger between AI and quantum, and away we go into the singularity.
Obvious VC play (Score:2)
No need to worry about newer tech than ours (Score:2)
Just buy Nvida products.
If at all (Score:1)
Jensen shorted the stocks (Score:1)
Define useful and usecase (Score:2)
We just don't need quantum computing for most use cases now.
QC use cases
- pure satisfiability problems
- proof verification
- factoring (finding all factors)
- precise decryption (ml will be able to get close enough and also circumvent qc safe encryption)
QC use cases already obsolete
- estimated decryption (good enough to read is good e