Quantum supremacy, now with BosonSampling
scottaaronson.com
scottaaronson.com
Quite incredible! And for anyone who's worked in optics or photonics, a real experimental feat. (just look at that table!)
It's not about what you want, it's about who wants you.
like getting a job...
At first, I thought this was more of a type of single-use quantum circuit, where you can only use it to solve this one problem. And you'd have to build a new rig to solve another function. And on and on, until you sorta get a general purpose super quantum computer, with all these individual rigs solving different functions, that can finally do useful work.
Whereas the Google Sycamore is more of a general purpose quantum computer that can be reprogrammed to solve anything.
But, this guy believes that this might be a first step to a universal photonic quantum computer.
The last part was incredible. They verified the results on a supercomputer, and went up to n=40. But the supercomputer time costed them $400,000. So they decided to stop there.
Public Key Crypto Code breaking, here we come. Setec Astronomy [1]
Their optical detectors are superconducting and a lot of the apparatus is temperature controlled. (doesn't really contradict your point, but I thought it was interesting to point out).
I really appreciate this write by Scott Aaronson. It anticipates all possible inquiries. And is heavily grounded in CS Theory. The subtext of this open collaboration hints at universal cooperation and co-prosperity. But I fear the fallout may be closer to the "arms race" winner-take-all mentality!
I believe my main confusion is why is this experiment being called a computation: to me this seems to be simply the observation of some physical phenomena which happens to be very hard to calculate with a classical computer.
> An analog computer or analogue computer is a type of computer that uses the continuously changeable aspects of physical phenomena such as electrical, mechanical, or hydraulic quantities to model the problem being solved. In contrast, digital computers represent varying quantities symbolically and by discrete values of both time and amplitude.
[1]: the google achievement was designing a specific problem for which their quantum computer could perform better than classical computers. Now, this is a solution to a preexisting problem. Not to say that Google's result wasn't important, but this one is “one rank” higher.
Actually, some people weren't totally convinced that Google's result really deserved the “quantum supremacy” label. Sure, the guys at IBM[1] were a bit bitter, but their argument make sense (and they weren't alone on that).
This result unambiguously deserve such qualification: this the kind of results everybody expected to be called “Quantum supremacy”, so that's a bit sad that they won't get as much credit as google's team.
[1]: https://www.ibm.com/blogs/research/2019/10/on-quantum-suprem...
On the other hand, the device used in this experiment is a single-purpose non-universal sampler with very limited applications.
On the scale :
0. Computer solves a specifically crafted problem < 1. computer solves one preexisting problem < 2. computer solves several existing problems < 3. computer solves all theoretically solvable problems.
Google's result is at 0, while this is a 1.
(By solve/theoretically solvable I mean: solve faster than a classical computer, and should be solvable faster with quantum than classical)
No! Our world is harder to simulate than a classical one, which can only have classical computers.
It is impossible in principle to have any evidence about simulationism. All knowledge about everything from logic up is from inside the simulation. Just accept that it's an epistemological tarpit and give up.
But instead what is being done is to construct a system where each of the components is simulateable on a classical computer -- we completely understand how they work. But when you put a number of the components together, the complexity of the classical program required to simulate the system becomes exponentially larger, rendering the classical simulating infeasible, while the complexity of the realized physical system increases only linearly.
Since molecules are a quantum computer in themselves, this would imply (if it were possible) that to simulate quantum effects on a classical computer for even a small number of molecules would require a quantum computer exponential in the number of molecules.
So to simulate one mole of air in its full complexity would require not just ~6e23 bits of computing power but instead 10^(10^23) bits of classical computing power.
By contrast, simulating a very long series of classical coupled pendulums, though very complex, is a problem that increases linearly (or sublinearly) with the number of pendulums. Even though this system has nonlinearities that make it nearly impossible to compute in closed form, and exhibit chaotic behavior, simulation can be done feasibly.
If we hypothetically assume we are a program, there is nothing to derive about the stimulants their reality or/and capabilities.
They might be as well a 4D being running some tests. We might never know. maybe their physics is different then our who know.
Furthermore you do not need to simulate ever molecule in out universe. I am sure there would be optimizations of stuff that doesn't need to be simulated as its not observed.
Say up until recently a Pluto might needed less resources to simulate as it was never really seen, now it surface needs to be simulated etc. I am sure you get the point.
I myself am not really concerned about the simulation hypothesis and I think its less likely. A simulation needs to be useful and if we were to run our simulation and start a possible chain then the cost of running the whole thing is felt the most at the actual real universe. If the cost is too high (ie time resolution is too slow for the liking of the owners) it will be terminated or rebooted.
So if we are in simulation then not starting a simulation would be the wisest choice to prolong our existence.
Either way its just an idle friday chitchat with no way to check or prove anything :)
In the spirit of sensationalizing "China" and "supremacy", does anyone know if this technology is relevant to some Star Wars weaponry for example?
[1] https://science.sciencemag.org/content/early/2020/12/02/scie...
They should have requested AWS free credits :)
Ha
That wasn't a coincidence.
> Just after the lockdown of Wuhan in January, the same week the U.S. confirmed its first case of the novel coronavirus, Chinese civic organizations in dozens of countries on five continents began buying masks and other personal protective equipment.
> In Nagoya, Japan, volunteers drove to pharmacies and bought 520,000 masks in three days, according to an account carried by Xinhua, China’s official news agency. By Jan. 26, the head of a Chinese chamber of commerce in Toronto, just back from a trip to Beijing, started making calls to members telling them they needed to join the effort, the report said. Almost 100 people drove to Toronto—some overnight, on icy roads—and were dispatched to buy supplies. Planes out of Kenya and Milan were packed with boxes and suitcases filled with PPE bound for China. An overseas Chinese association in Argentina sent some 25,000 masks within a week of receiving the request.
https://www.bloomberg.com/news/articles/2020-09-17/behind-ch...
Seems like it was a global phenomenon. There were articles in the local newspapers where I live (US) interviewing Chinese expats who were buying all the PPE they could find and shipping them home.
While it put other countries in a bad position, I think it was probably a perfectly normal thing for an expat to do since people didn't know how bad it would get outside of China at the time. Most times there is a disaster somewhere in the world, the people from that country who live elsewhere try to help.
Keep in mind that this was also right around the time the WHO reported there was “no human to human transmission” based on data from the CCP.
Outside of China, nobody was talking about needing masks until March. I'm pretty sure the author is talking about the fact that by March, China had already produced tons of new masks but average people in the US couldn't access them because they were being reserved for health care workers.
https://science.sciencemag.org/content/early/2020/12/02/scie...
Because China(chinese people) funded the science? It may also hint that the governance system is productive.
Any time, any one who brings up their nation as being the best - I've got no room for you in my life. I despise nationalism and national pride in all forms. It only breeds hatred and unproductive relationships, and creates echo chambers. This goes for all countries.
We're tiny blobs of muttering meat in the universe. Temporarily here.
You are suggesting you know better than a system that millions of years of evolution have created. You are also assuming that you know all the risks of NOT organising in tribes, that somehow if all humans were some untribal lonely individual entities, the world would work better. Many many billions of people and animals seem to disagree.
All of these thing you mention are simply attempts to attenuate evolution, not conquer it. Moreover, all of the things you mention are a new feature set developed from evidence given to us by evolution.
About the only area where biological evolution is relevant now is pathogens, because these evolve on timescales comparable to technological development of humans.
Humans managed to create a means to communicate with each other. End of evolution right there. We figured out germ theory and then vaccines. The DEATH of human evolution. We have the Internet now so that must mean human evolution just stops, because integrated circuits or something.
We look back on proto-humans, who had many tools, with the same arrogance. There’s no reason our species won’t evolve further to look back on the “internet” and laugh.
Timescales, huh.
Evolution for humans just doesn't matter, because at this point, we're doing meaningful changes to ourselves and our environment at a rate faster than it takes for a human to reach reproductive age. We didn't wait for evolution to adapt us to colder climate, fix our eyesight or change melanin levels; we just developed clothes, glasses and sunscreen - which also made these features not influenceable by natural selection pressures. And whatever mutation happens that doesn't kill a person, we work around it socially or technologically - which means its impact on reproductive fitness is usually close to zero. Biological evolution is no longer the force driving the shape of our species.
There is an evolution of humans happening, but it's social and technological, not biological.
You simply can't just say these don't matter any more, because it's ignoring reality. Every time you're turned on by someone attractive, its evolution thats doing that. And good luck thinking rationally when you are truly in love. (or to take a different example, when your body has a temperature in reaction to a disease. Good luck overriding your absolutely terrible mental state with rational human thought).
I think what you're missing is that... nobody in this thread is saying that.
In any case, I think we have managed the counter the forces of natural selection that affect our physical health, but not those that affect our mental health.
I'm suggesting that it might be wise to try to keep our baser instincts at bay -- as well as we can, at least. Sure, we'll fail from time to time (or often even), but that doesn't mean we should give up that struggle.
The internet has created other strongly bonded echo-chambers. Elon Musk lovers/haters for example.
can not agree more on this
See: every significant accomplishment by SpaceX or NASA for a repeating case.
If you really want to see objectors get triggered, check out the past threads that tangent off about SpaceX employees chanting USA USA USA on video, that's always quite amusing.
Science might be a global cooperative effort, but this paper comes from China.
Maybe it is a sign of a society still focused on building stuff rather than redistributing and administrating what is already there.
Easy, the original title of TFA, "Quantum supremacy, now with BosonSampling".
Editorializing breaks site guidelines.
Haha... that's not at all the case; academics are petty people with fragile egos who are often under the thumb of lab directors with even worse egos and often come into that role with an ax to grind. Racism, be it veiled or otherwise, is also no stranger to the realms of Science. Much less the dregs called the peer-reviewed system. Consider the plight of Dr. He Jiankui who many researchers in the West lauded in private, but scorned in public when the 'evil Chinese Scientist' revealed his results to the Public because grant money and political standing were at risk.
Collaborations are limited in so far as their personal utility for more grants or prestige, lets not elevate researchers and their horrible practices to places they do not belong.
I think you would find the process entirely sad and depressing, as most of us did in our undergrad, when you see it for what it really is.
I've only heard universal condemnation from scientists, where can I learn more about this private support for him? Or am I going to take your word for it?
Not necessary, but I'm just going to assume you aren't a biologist nor do you follow the Bio-hacker culture, right?
Start with Josiah Zayner's talks from the time of that event. I think it was in the conference ahead of the trailer release of 'Code of the Wild' that stands out. But he's on record talking about it as well, I think Dr. Church also touched on it as well. The Bio Hack the Planet conference talks might also have more details, as I recall him talking about the events leading up to Dr. Doudna's misplaced/misunderstood role in CRISPR that eventually got her a Nobel Prize and it segued into that discussion, too.
Back on PC Edit: This piece [0] is also worth a read, here is an excerpt from Science:
> Yet opposition was not unanimous. A few months before He met the couples, a committee convened by the U.S. National Academies of Sciences, Engineering, and Medicine (NASEM) concluded in a well-publicized report that human trials of germline editing "might be permitted" if strict criteria were met. The group of scientists, lawyers, bioethicists, and patient advocates spelled out a regulatory framework but cautioned that "these criteria are necessarily vague" because various societies, caregivers, and patients would view them differently. The committee notably did not call for an international ban, arguing instead for governmental regulation as each country deemed appropriate and "voluntary self-regulation pursuant to professional guidelines."
>> That circle included leading scientists—among them a Nobel laureate—in China and the United States, business executives, an entrepreneur connected to venture capitalists, authors of the NASEM report, a controversial U.S. IVF specialist who discussed opening a gene-editing clinic with He, and at least one Chinese politician. "He had an awful lot of company to be called a ‘rogue,’" says geneticist George Church, a CRISPR pioneer at Harvard University who was not in the circle of trust and is one of the few scientists to defend at least some aspects of He's experiment.
The questionable stance, and reluctance comes when academic researchers know that their grants and good standing with a chosen University (and the prestige and clout they garner) comes under threat. Maybe its because I've seen this kind of thing play out in real time when I was in School and then entered the Industry that I'm not all surprised by it... but it's also a sense of remorse mixed with dread and despair as a person who was on track to be a lab Clinical Lab Scientist but what disgusted at what 'really happens' internally that shaped my view(s) and made me leave the Industry entirely.
0: https://www.sciencemag.org/news/2019/08/untold-story-circle-...
I also wonder how the Google result was reported in China. I wouldn't be surprised if it was reported as "Americans have done X".