Google Plans To Build a Commercial Quantum Computer By 2029 (engadget.com) 56
Google developers are confident they can build a commercial-grade quantum computer by 2029. Engadget reports: Google CEO Sundar Pichai announced the plan during today's I/O stream, and in a blog post, quantum AI lead engineer Erik Lucero further outlined the company's goal to "build a useful, error-corrected quantum computer" within the decade. Executives also revealed Google's new campus in Santa Barbara, California, which is dedicated to quantum AI. The campus has Google's first quantum data center, hardware research laboratories, and the company's very own quantum processor chip fabrication facilities.
"As we look 10 years into the future, many of the greatest global challenges, from climate change to handling the next pandemic, demand a new kind of computing," Lucero said. "To build better batteries (to lighten the load on the power grid), or to create fertilizer to feed the world without creating 2 percent of global carbon emissions (as nitrogen fixation does today), or to create more targeted medicines (to stop the next pandemic before it starts), we need to understand and design molecules better. That means simulating nature accurately. But you can't simulate molecules very well using classical computers."
"As we look 10 years into the future, many of the greatest global challenges, from climate change to handling the next pandemic, demand a new kind of computing," Lucero said. "To build better batteries (to lighten the load on the power grid), or to create fertilizer to feed the world without creating 2 percent of global carbon emissions (as nitrogen fixation does today), or to create more targeted medicines (to stop the next pandemic before it starts), we need to understand and design molecules better. That means simulating nature accurately. But you can't simulate molecules very well using classical computers."
Very nice (Score:2)
But will it run Linux?
Re:Very nice (Score:5, Funny)
Even better ! Quantum blockchain - a distributed ledger where all transactions co-exist with their reversals. Now we're talkin' about some awesome accounting.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
But will it run Linux?
Nope. QCs are completely useless for almost all types of computations regular computers can do.
Re: (Score:1)
Qunix, but you won't know if it's delivered or not until after it crashes and takes cats with it.
Re: (Score:2)
But will it run Linux?
No, but it will serve advertisements much faster.
Obligatory XKCD (Score:5, Insightful)
https://imgs.xkcd.com/comics/r... [xkcd.com]
So we are somewhere between "rest is just business" and "we haven't finished inventing it yet."
Re: (Score:2)
25+ years => It has not been conclusively proven impossible
"we're not really looking at applications right now" => "Il like being the only one with a hoverboard"
21st Century Dual Purpose (Score:2)
(Google) "...create fertilizer to feed the world without creating 2 percent of global carbon emissions (as nitrogen fixation does today), or to create more targeted medicines (to stop the next pandemic before it starts).."
(Government) "Yeah, speaking of fertilizer, enough of this bullshit. Can we skip to the slide where you tell us how many different kinds of encryption you can crack like an egg for us now? We'd appreciate it."
Re:21st Century Dual Purpose (Score:5, Interesting)
(Government) "Yeah, speaking of fertilizer, enough of this bullshit. Can we skip to the slide where you tell us how many different kinds of encryption you can crack like an egg for us now? We'd appreciate it."
RSA and elliptic curve signatures are going to be toast. Everything else is fine. Quantum computers are going to provide quadratic performance improvement for most algorithms, so cracking 128-bit AES will still require 2^63 steps on average. This is outside of the realm of possibility with quantum computers for a long time.
And simply switching to AES-256 moves this attack back into the realm of pure impossibility.
Re:21st Century Dual Purpose (Score:4, Interesting)
Also, those 2^63 steps are very much outside of what conventional computers can do these days without special measures. QCs are several orders of magnitude less reliable and losing coherence just once does invalidate the complete computation, you have to start from the beginning again. Nobody will see a QC that can break AES-128 in the next 100 years. It is completely unclear whether this is possible at all in this universe. Also remember this is for a known-plaintext attack, i.e. the very easiest attack imaginable.
Re: (Score:2)
Nobody will see a QC that can break AES-128 in the next 100 years.
This is almost certainly true. OTOH, anyone designing a new cryptographic protocol or system today should just use AES-256 and SHA-512 and eliminate the possibility entirely. And they should make sure they have plans in place to switch to quantum-resistant asymmetric algorithms when we collectively pick some winners from the candidates available. And plan for larger asymmetric keys. Probably 2-3 orders of magnitude larger.
Quantum computing will require some changes to the cryptography we use, but we can
Re: (Score:2)
I would say, "Quantum Computing may eventually require some changes", but completely eliminating the possibility for symmetric crypto is easy and should be done in the next 10...20 years or so, but no rush. I don't think crypto-hashes are really under threat though. Still, upgrading to SHA-512 is a good idea because it eliminates a risk from conventional attacks.
As to asymmetric crypto, do not go ECC is probably the best advice. Breaking RSA requires 3 times the QBITs it has bit-length. I do not see a 12k Q
Re:21st Century Dual Purpose (Score:4, Insightful)
Nobody will see a QC that can break AES-128 in the next 100 years.
Ten years ago we landed a craft on Mars. A little over twenty years ago, you were dialing up to the internet. Forty years before that, the internet wasn't even a twinkle in ARPAs eye. Sixty years ago we put men on the moon. Eighty years ago we split the atom and ultimately ended a second World War. And 100 years ago, we were recovering from the last pandemic driving around in steam-powered cars.
Since you like to make century-long predictions, let me know why you think your 1921 self, could have accurately predicted any of that shit.
Humans will always remain ignorant predictors of the future. Hell, we can't even learn from the past.
Re: (Score:2)
So? You seem to be ignorant as to the complexities involved.
Here is one: 75 years ago we invented digital computers. Today we are still using them because all efforts to replace them have completely failed. The problem with QCs is that we know too much about them and none of it good for scaling.
Re: (Score:2)
So? You seem to be ignorant as to the complexities involved.
Here is one: 75 years ago we invented digital computers. Today we are still using them because all efforts to replace them have completely failed.
75 years ago that "computer" was the size of your garage and ran on vacuum tubes. Then, we invented the microchip.
I think we can stop pretending we haven't done shit with that, for 75 years. The kids aren't exactly in the garage every weekend turning a wrench on blown tubes coding their homework for hours. And we've actually replaced a lot of computers with the infamous cloud. Most "computers" today serve as merely the interface.
The problem with QCs is that we know too much about them and none of it good for scaling.
The problem with humans, is ignorance. I find it rather incredible that ev
Re: (Score:2)
75 years ago that "computer" was the size of your garage and ran on vacuum tubes.
So? It used the same computing paradigm as today, just with less power and a clunkier interface.
A QC uses a fundamentally different computing paradigm.
No alternate computing paradigm has ever been successful, except for classical electronic or mechanical analog computers and they were used before digital computers were available and have been phased out mostly.
Side note: The first programmable computer was the Zuse Z3, and it worked electromechanically. The claim that the US had the first programmable compu
Re: (Score:3)
And yet we still haven't broken the light speed barrier, or invented a personal SSTO rocket backpack.
There are things that are difficult but possible, and things that are either impossible or nearly so. We're usually overly cynical about the former, and overly optimistic about the latter.
The fact that Google is talking about quantum chemistry suggests an outbreak of common sense. Quantum computers should be good at simulating quantum systems. The people still talking about encryption are living in fantasyla
Re: (Score:2)
And yet we still haven't broken the light speed barrier...
Yes, which is why we should stop using idiotic metrics to create the illusion of hope. Every time I hear of a new Earth discovered that's "only" xx light years away, I just laugh. Why are we even talking like that? The farthest man-made object in space, is still measured in light hours. That "only" took over four decades.
Ironically enough, we're back here on earth still measuring damn near every kind of engine using horses. We're stupid to the extreme in both directions.
...or invented a personal SSTO rocket backpack.
First off, we have invented rock
Re: (Score:2)
Ironically enough, we're back here on earth still measuring damn near every kind of engine using horses. We're stupid to the extreme in both directions.
That particular stupidity applies to the US and one more backward country I have forgotten, I think.
There are things We Know.
There are things We Know We Don't Know.
There are things We Don't Know We Don't Know.
We haven't even really created a good QC yet. I find it incredible that humans still feel everything we'll ever know about QCs, is guaranteed to fall into only those first two categories.
Just to clear that up, I never claimed we know most things about QCs or that we even know most things we do not know about them. We do know enough that, if they should be possible at all at larger scales (all observed "computations" so far are of complexity low enough they could be caused by some _other_effect), it will be extremely hard or practically impossible to get there. Also, at higher Qbit numbers the
Re: (Score:2)
To be fair, only a couple weird corners of the world actually do that.
SSTO stands for single stage to orbit. Rocket backpacks are possible, and easy enough you can build one in your garage these days. Single stage to orbit vehicles are probably possible on Earth, but not in backpack form. Well, maybe with metallic hydrogen.
Re: (Score:2)
There are things that are difficult but possible, and things that are either impossible or nearly so. We're usually overly cynical about the former, and overly optimistic about the latter.
Indeed. There are also things that are (likely) possible but make no sense. See, for example, flying cars or colonies on Mars (for a current one).
Re: (Score:2)
To be fair Jules Verne envisioned nuclear powered submarines during the civil war era.
Re: (Score:2)
Re: (Score:2)
RSA and elliptic curve signatures are going to be toast.
Which means Bitcoin is going to have a massive re-balancing.
Re: (Score:2)
Re: (Score:2)
So you'll have a transaction that converts from wallets from elliptic signatures to the new scheme.
How can you do that in a way that prevents wallet stealing?
Re: (Score:2)
How can you do that in a way that prevents wallet stealing?
The first transaction wins. Basically, it'll look like transfer from a wallet with old signature to a wallet with a new signature. It can only be done once, so knowing the private key of an empty wallet won't provide you anything.
Re: (Score:2)
Oh yeah, then I'll have to still assert my original statement. Bitcoin is going to have a massive re-balancing. People who are fast will have the money, not people who originally had it.
Re: (Score:2)
Let's look on the bright side, though. At least Craig Wright "Satoshi" will be able to get his lost tokens back.
Re: (Score:2)
Re: (Score:3)
(Google) "...create fertilizer to feed the world without creating 2 percent of global carbon emissions (as nitrogen fixation does today), or to create more targeted medicines (to stop the next pandemic before it starts).."
(Government) "Yeah, speaking of fertilizer, enough of this bullshit. Can we skip to the slide where you tell us how many different kinds of encryption you can crack like an egg for us now? We'd appreciate it."
Won't happen.
The government has a split personality here. On the one hand, it's really important for national security that both the government and important elements of the economy have access to strong cryptographic security. On the other hand, spies and law enforcement really want to be able to read everything. The NSA actually has completely separate divisions, one focused on increasing security, the other focused on breaking it.
NIST is leading the charge [nist.gov] in the effort to select and standardize new
Re: (Score:2)
(Note: "crypto" is short for "cryptography", NOT "cryptocurrency")
After that explanation, I appreciate the subtle touch of humor here.
Your passage practically defines and exudes confidence with conviction. I'm not trying to necessarily question it...OK, scratch that. Maybe I am. I'd like to go back in time to review just how convicted we were about our strong crypto of yesteryear, as we stand over those graves of hashes previously heralded as Won't happen.
Remember when older crypto came with government
Re: (Score:2)
I'd like to go back in time to review just how convicted we were about our strong crypto of yesteryear, as we stand over those graves of hashes previously heralded as Won't happen.
To what yesteryear are you referring? Outside of hashes, our crypto of the last half century has held up very nicely. DES is still about as secure as it was when it was invented (after the NSA fixed the S boxes to made DES resistant to differential cryptanalysis, a technique not yet discovered by academic cryptographers). The key size is too small because computers have gotten faster, but the base algorithm is still just fine -- and 3DES is secure in practice today and for the foreseeable future. RSA is st
Great, more bullshit (Score:2)
On the other hand, it is just a "plan", and in this case a shameless lie. QCs, if they ever work in any useful way (which is far from certain today) will not play any role in the next few decades. Maybe in 50 or 100 years there will be some breakthrough, but there may not be one as well.
Re: (Score:1)
Google will abandon the plan at least twice before 2028. Then, of course, they'll abandon it again several years later.
Re: (Score:2)
Indeed. I forgot that that is their MO. Makes for much simpler planning though, just announce whatever you like bit make sure there is ample time.
Gimmick (Score:5, Insightful)
It's pretty easy to semi-promise something 10 years out. By 10 years, most will have forgotten about the promise, or you will be promoted or removed such that it's not your problem anymore. It's a sleazy way to jack up stock price without having to do anything for real.
Re: (Score:2)
Exactly, add to it Google's failed "Quantum Supremacy" PR stunt from two years ago and it's clear it's going nowhere.
Re: (Score:1)
Is Google already washed up? They are acting like Oracle.
Re: (Score:2)
It's also necessary when you plan to invest large amounts of money. Can't just tell people "we are spending tens of millions on this secret project which we don't want to announce in case it doesn't pan out", at least not if you want to consider that investment and R&D an asset for accounting purposes.
It's normal for companies to announce major R&D projects. No promises here, just a plan with a clear roadmap to a stated goal, the kind of thing engineers are asked for all the time.
Re: (Score:2)
Re: (Score:1)
People stopped listening to IBM a long time ago. They gutted themselves in a profits-over-quality push and killed their reputation.
So now they can (Score:4)
I'm taking bets (Score:2)
Re: (Score:2)
Commercials (Score:2)
Re: (Score:2)
Or maybe not. (Score:2)
Re: (Score:2)
They will do it (Score:2)
Google... (Score:2)
Look for the follow up announcement where they are dropping support for it in 2031...
Singularity (Score:2)
Gotta hand it to Kurzweil on his date predictions.