Photonic Quantum Advantage? | Hacker News Reader

Photonic Quantum Advantage?

gilkalai.wordpress.com

43 pointsnsoonhui30 comments

Looks like Scott Aaronson acknowledges this as a potential real issue in an update to his previous blog post (https://www.scottaaronson.com/blog/?p=5122). I'll be curious to see where this goes...

I can't intuit whether "advantage" has some specific meaning here, but an advantage, in the dictionary definition sense, is the fact that photonic quantum systems easily operate at room temperature, unlike, say, electron-based systems.

"Quantum Advantage" and the synonymous "Quantum Supremacy" (considered somewhat inappropriate because of other uses of the same word in racist orgs) are recently established "terms of art" in this field. They refer to "an experiment that unambiguously demonstrates computational capabilities unreachable by a classical supercomputer". Importantly, it is not really expected that the computation in question would be something useful, rather just something classically difficult as a proof of concept. I would say "useful quantum supremacy/advantage" experiments are not yet demonstrated and will take a few more years (maybe almost a decade). Still, the progress since conceiving this field in the early 90s is astronomical.

> considered somewhat inappropriate because of other uses of the same word in racist orgs

are we really in a time where "double plus good" is the only acceptable way to indicate that something is better or do we have to point out that racists probably also used each of these words in problematic contexts?

Apparently, yes. Lots of organizations felt the need to rename the master branch, so, what other proof do you need?

As a side note, I grew up in communism, and it would be funny if it wouldn't be sad how much of a feeling of deja-vu this all is. Many people mistake this for a form of "enlightenment" but it's really just a form of herd behavior most of the time :(

Naively, i kind of feel like advantage is a better term. Supremacy feels like it has implications that quantum computers have overcome all obstacles, which seems a bit much for a contrived problem with no real world applications

As someone who works in the quantum info / computation / foundations field I prefer "advantage" to "supremacy" not just because of all the racists using the word "supremacy" but also because advantage is more accurate. For me "supremacy" would indicate that quantum computers are convincingly beating classical ones for all or most problems at sufficiently large scale. I.e. if you're google or whatever you're buying quantum computers and using them for a big chunk of you computing needs.

This scenario is (in my opinion) overwhelmingly unlikely to occur during my lifetime (or ever). Classical computers aren't going anywhere. Quantum advantage conveys the reality that there are some problems that using quantum computers gives you advantages for, but others for which you should be using classical computation.

"Advantage" sounds quite a bit better than "double-plus good", but I get your point.

I personally like the sound of "quantum supremacy" and think it sounds cool, but it really does not seem worthwhile to die on this hill. If some people have ruined the word and given it extremely negative connotations in the minds of many of my colleagues, why would I insist on using it. For a comically extreme extension of this principle: the swastika used to be a pretty positive symbol, but it was ruined as such. It is not the first time that we as a community have decided that a common symbol/word is "ruined" and I do not particularly care where we place the line on what is considered "ruined".

>> Still, the progress since conceiving this field in the early 90s is astronomical.

I know next to nothing about quantum computers but I note that the first programmable computers were created a few years after Alan Turing's "On computable numbers" in 1936. In fact, after Turing described a Universal Turing Machine, programmable computers started appearing left and right with many groups in the USA and Europe at least building their own [1].

What I mean by this is that, as someone who doesn't know much about quantum computing, it seems that any practical applications of the approaches that are frequently hailed as breakthroughs in the lay press seem to be taking a much longer time than I'd expect, given the prior of classical computers and how fast they were possible to implement in practice and start using for real.

Probably this means that creating a quantum computer is a much harder problem - but then, that's perhaps something to keep in mind when trying to make sense of the reported progress in the field, for those outside it?

_____________

[1] Oh and let's not forget Charles Babbage and his analytical engine, never fully built but nevertheless perfectly serviceable (at least on paper). And the work of just one man (who incidentally also invented feature creep at the same time).

I would say it is legitimately much more difficult. Pretty much since the conception of quantum mechanics up until around year 2000 most luminaries of the field would say "it is impossible to harness single-particle quantum effects, all human tech would be based on collective behavior". And to realize a quantum advantage you need those single-particle effects. Kinda like asking Babbage to build the analytical engine before the wheel (let alone the cog) was invented.

Another thing to keep in mind: Creation of a working quantum computer would be a monumental achievement, letting humans harness the full computational power of nature's laws, a defining moment in a civilizations technical capabilities, etc. However, there are only a handful of computational problems at which quantum computers would be exponentially better than classical (most notably simulating novel materials and drugs and certain optimization problems). While these problems are simply impossible on classical computers, for the majority of other problems classical and quantum computers would be on equal footing.

Thanks for the clarification!

The photonic quantum computer generates up to 76 output photon clicks, which yields an output state-space dimension of 1030 and a sampling rate that is ~1014 faster than using the state-of-the-art simulation strategy and supercomputers.

Art of War: Appear weak when you are strong.

IME, it's plausible that nation/state strat pertinent to QC advances is to hide the true extent of cutting edge breakthroughs. This makes sense considering what's at stake. China is suspected to engage in bioengineering for their soldiers and it's not a stretch to speculate they also seek such synthetic advantages with their vast armies of engineers and scientists. This would lay fertile ground for secret breakthroughs in, e.g., QC.

I assume that both China and USA/fyev seek to obtain secret radical advances in key areas such as QC. The urgency for adopting quantum resistant crypto suggests such concerns might be shared by powers that be. QC breakthroughs are unpredictable, so the smart move is to assume the enemy could make secret advances at any time.

So they are going to hide their breakthroughs by announcing their breakthroughs?

If that's the case here, then it isn't indicative of any secret progress.

Or maybe it's a misdirection to get the enemy to go down a path they know is a deadend.

Or a lure to get the enemy to work harder on a problem so they can steal the solution.

Or fake so that the enemy spends more resources faster.

Or any of many other possible scenarios. It's a nonstop game of spy versus spy.

Yes to all.

Mainly, American culture could use a strong dose of reality check. This appears to be the case in terms of publicly stating that we are retooling global intelligence missions from Cold War (t-40yrs) and War on Terror (t-20yrs), to the China threat.

Awareness of strategies, capabilities of the enemy, and what's at stake will help reduce Western complacency.

Is this a troll?

Spend some aws credits and verify the results of the paper for higher and higher numbers, if you don't believe it.

I'm not saying china can't be evil, but faking a breakthrough that can be verified (to a certain level) is a next level stupid way of being evil.

Pretty sure it was just a general comment on the topic of art of war, and not specific to the linked paper.

Pray tell, in your view, what is the antecedent of the pronoun "it" in geomark's comment, if not the paper in question?

Snarky tone of your response put aside:

Whether the comment was incorrect in this case related to the paper in question doesn't invalidate the strategy, generally. Might be easier to pass scientific misinfo/disinfo if it were more difficult to reproduce, of course.

The comment was likely done in haste, meant to convey general strat that warrants consideration when examining the paper -- but could be easily invalidated, as you pointed out.

Correct. It doesn't have to be fake. It could be a very real breakthrough. Then the possibilities branch out. Taking the next step might require huge resources. Will the enemy take that on? If so, can you just watch and spy, saving your own resources to apply to other or more promising efforts?

Some people might think this is all just being too imaginative. But that is what the spooks are constantly doing, creating layer upon layer of misdirection and obfuscation. To wit, a sibling comment[1] about the Star Wars program.

[1] https://news.ycombinator.com/item?id=25329030

I can't help but feel this amounts to: the chinese state might be doing something to futher some sort of geopolitical end, but we don't know what that end is or what the something that they are doing is.

And you're right, the chinese state probably has various plans to take advantage of whatever advantages they think will be good for their interests (just like every other country). But if we have no information or even plausible theories, what good is this baseless speculation? Its not exactly news that China is a rising super power and probably has schemes within schemes. Maybe this one is the obvious one, that a working quantum computer would be very prestigious and economically valuable to the first country to get there. Or maybe not. But if it isn't the obvious one, guessing at utter random without even reading up on the underlying paper almost certainly will not come up with whatever the scheme is.

Does anyone remember the "star wars" program of early 1980s? It was a massive fraud smoke-and-mirrors program to lure the USSR into an area of weapon development where it could not sustain itself economically.

It included plans to build a colossal network of space-borne X-ray lasers powered by nuclear blasts, which would be able to destroy ICBMs before they finished acceleration and deployed warheads and decoys. This would allow the US to break out of the mutually-assured destruction balance.

None of that was true. The US did not (and still do not) have a technology to produce nuke-powered X-ray lasers. The US did not have the launch capacity to deploy the advertised number of space platforms. Not nearly.

Still, this was (or is believed to be) instrumental in the peaceful demise of USSR due to extreme economic difficulties resulting from a symmetric effort.

It might not have been fatally flawed, but it wasn’t completely smoke and mirrors. It was a real an research program employing scientists and engineers that built and tested laser platforms for deployment. It wasn’t viable using contemporary technology, but neither is commercial fusion power, and we have been working on it for several decades. Would you consider ITER smoke and mirrors?

Yes, but I think that the achieved advances have been greatly exaggerated, IIRC.

> China is suspected to engage in bioengineering for their soldiers and it's not a stretch to speculate they also seek such synthetic advantages with their vast armies of engineers and scientists. This would lay fertile ground for secret breakthroughs in, e.g., QC.

In which a HN commenter proclaims that China can build a working quantum computer by bioengineering their scientists to be smarter.

I speculated that they may seek to use eugenics, nootropics or other methods to enhance the performance of their (vast) pool of scientists and engineers. This would offer a strategic advantage in accelerating such development, but would assure nothing.

The phrasing of your response re-interpretating/rephrasing my statement is misleading/inaccurate.

Does this apply to Google's supremacy announcement and IBM's classical challenge to that?

If those two are following Sun Tzu's advice, they'd probably be attempting to appear strong where they are weak.

Reply on news.ycombinator.com