The teapot test for quantum computers

[@christophclear1438]

By not having smashed the tea pot, alas tea pot supremacy remains a conjecture.

[@perronjk]

and had he smashed the tea pot how could we verify the result was correct without running the classical simulation?

[@darryljohnson8516]

I just wish the resolution was good enough to see all the books in the background

[@NilodeRoock]

Exactly what I was thinking...

[@sfermigier]

My own obsession every time someone posts a video or photo with books in the background.

[@kilogods]

They’re all math books

[@jonrjeffrey]

I particularly enjoyed the line about having a time machine, but only using it to watch missed soap opera episodes.

[@verystablegenius4720]

He should do a pod-cast

[@rudilapa6569]

I spewed coffee. Such a biting, witty indictment!

[@aidanhschofield]

Richard, this was epic! Please rant more often.

[@richarde.borcherds7998]

It's not so easy to think of mathematical topics suitable for ranting. Any suggestions?

[@chetanvuppulury7537]

@@richarde.borcherds7998 Not strictly a mathematical topic, but the closure/reorganisation of math departments (like the recent news from Leicester)

[@aidanhschofield]

@@richarde.borcherds7998 It has to be something about which you have a strong opinion. So my suggestions are unlikely to appeal to you. However, it would be graceless not to give you one so do you have a strong opinion about the Newton-Leibniz priority argument? or perhaps a strong opinion that such arguments are a waste of time (and were a waste of time for the next 100 years).

[@richarde.borcherds7998]

As far as I know, almost everyone now agrees that they discovered calculus independently and N was not graceful about it. I can't think of much to add.

[@matheme7atica]

@@richarde.borcherds7998 A discussion on "finitist", "intuitionist", etc mathematical beliefs would be interesting

[@john-r-edge]

My late Uncle was a Maths lecturer at Birkbeck College, London University, and took an interest in a wide range of subjects. Visitors to his flat would often admire his large collection of books. He would reply, saying the collection would be much more impressive had he actually read them all. I hope it would not be too rude to ask Mr Borcherds to reveal what percentage of his (awesome) book collection he has read?

[@richarde.borcherds7998]

The word you are looking for is "tsundoku".

[@BlueSoulTiger]

en.wikipedia.org/wiki/Tsundoku

[@NothingButFish]

A bookshelf of unread books is much more valuable than one of read books.

[@BlueSoulTiger]

@@NothingButFish ... assuming you can recall what you've already read?

[@saulberardo5826]

Although Mr. Borcherds answer suggested he hasn't read the majority of the books in his bookshelf, I'm still curious to know what the ballpark estimate would be

[@888Xenon]

@Richard E. BORCHERDS Excellent rant, couldn't agree more! As an aside though, do you remember where you got the hollow polygon panels used to construct the convex uniform polyhedra scattered around your office?

[@richarde.borcherds7998]

It is the "Polydron Frameworks Archimedean Solids Large Set" made by polydron www.polydron.co.uk/mathematics-shop/polydron-frameworks-archimedean-solids-large-set.html

[@888Xenon]

@@richarde.borcherds7998 Fantastic, thank you :)

[@peterboneg]

Great video. I'd like to send to the people who keep messaging me about quantum computers, but I think you should have said at the end that they can't currently factor numbers better than digital computers.

[@EdwinSteiner]

"Teapot supremacy" needs its own T-shirt!

[@nonindividual]

This can also be used for separating other instances, outside quantum computers, of hype from substantive progress.

[@AndrewKnightMIT]

Absolutely outstanding point. Thank you for clarifying "quantum supremacy."

[@brettaspivey]

Oh an excellent rant, I agree 100%

[@kochasaito4628]

Sir, you are so serious and funny, I can not even imagine the test of eating ants!

[@DoggARithm]

I loved how he briefly turned into Douglas Adams for this

[@millwrightrick1]

The best use for a quantum computer is for it to be used to help achieve nuclear fusion. Both seem to just about 20 years in the future, no matter what year one starts from.

[@annaclarafenyo8185]

Nuclear fusion was achieved in 1951. It can be used to generate energy today, by blowing up bombs in underground chambers. Oil tycoons use nuclear fusion in tokamaks as a shiny unachievable boondoggle so as to stonewall transition away from carbon fuels which destroy the atmosphere.

[@samdirichlet7500]

Quantum computer will have arrived when a quantum computer can do (or assist in) a 100 bit Hadamard transform. Until then, it’s just people looking for grants.

[@WhosBean]

I agree with 99.9% of this, but one note: if "directing traffic" means city-wide direction and not at just a junction or something then the problem is actually massively complicated. It essentially boils down to a number of overlapping traveling salesman problems.

[@benjamindavid7371]

I wonder if quantum computers could actually find the optimal solution. Digital computers are already pretty good at finding approximate solutions to TSP (so I've heard), and I could certainly imagine these approximate solutions be regarded as directing traffic.

[@mishaerementchouk]

No they cannot (in polynomial time). It is proven that quantum computers do not speed up NP-complete problems. Moreover, as far as I know, quantum computers cannot speed up achieving the same degree of approximation as classical algorithms. This is definitely true for the max-cut problem but is most likely true for any APX-hard (that is problems where even approximations become hard at some point) problem.

[@WhosBean]

@@mishaerementchouk quantum computers are not general purpose computers. In that sense they cannot surpass classical computers. But a quantum computer can solve NP hard problerms faster than classical methods when they are specifically built for that problem. This isn't even a property unique to quantum computes. For example, a maze can be solved using a diffusion model, which solves an NP hard problem in polynomial time but can only do that. Not solve mazes, just solve that one maze.

[@mishaerementchouk]

​@@WhosBean The quantum computers I am talking about are ideal. They contain classical computers as a particular case. Therefore, they can do whatever classical computers can do. They can also do something, which classical computers can't, as evidenced by the Deusch-Jozsa and Bernstein-Vazirani algorithms. Yet, there are no examples of quantum algorithms solving NP-complete problems in polynomial time. In a model without the collapse of quantum states, such algorithms are somewhat straightforward. The NP-hardness for particular instances doesn't make much sense. Simply because the notion of polynomial time is not defined for particular instances. Finally, algorithms based on classical physical models do not change anything in solving NP-complete problems. They, indeed, may lead to speeding things up comparing to classical simulations,say, by eliminating limitations due to communications, memory access, precision, and so forth, but all that speedup will be polynomial and will not change superpolynomial asymptotics. It seems like the same conclusion holds for quantum physical models as well.

[@theoschijf8155]

All we need now is one quantum computer to factorize the product of two approx 1000 digit size primes. That, a teapot nor a dig computer can do in a day or so. The record today for a qc is 35 -> 5 x 7, I believe. Best application for QC is to run a nuclear fusion reactor, another bad idea. Practical QCs will not happen, I say.

[@poo2uhaha]

these videos are the best

[@santiagomartinez3417]

Loved the trolling! :)

[@googleyoutubechannel8554]

"this will probably be out of date" .. this is still too generous to 'quantum computers', they haven't existed in any practical format in the past 20 years, and there's no reason to suspect, given the state of QM, that this will _ever_ change.

[@copernicus633]

Borcherds is a Fields medal winner, which is roughly the equivalent of the Nobel Prize. There is no Nobel Prize for math, but the fields Medal is math’s greatest honor. I too wish I could read the book titles in the shelves.

[@owenkeith1188]

Teapot Supremacy

[@RalphDratman]

So fun

[@isabellebarruhet225]

Et un sous titrage ? Comme sur un téléviseur ! Çà m'éviterai de quitter cette vidéo à laquelle j'entrave rien du tout - bye bye !

[@theoschijf8155]

Two questions: how thick is your carpet? Is you teapot a dumb plastic one?

[@qbqbqdbq]

Classical efficient factorization algorithm may exist. Shor's is not a good test given what we know; problem must be BQP-complete.

[@qbqbqdbq]

Further consideration: a teapot isn't a quantum computer because the quantum effects culminate in macroscopic classical behavior. By the same reasoning a digital computer is also a quantum computer because solid state physics depends on quantum mechanics.

[@annaclarafenyo8185]

The current state of the art is about 50 qubits, nowhere near 'several thousand', several thousand would be enough to factor nontrivial numbers. "Quantum supremacy" has not clearly been achieved, not even teapot supremacy.

[@mathstrek]

A several-thousand qubit quantum computer based on the "gates model" would certainly be enough to prove quantum supremacy. But I think Richard is referring to D-Wave's machines, which have 2-5k qubits but are based on the adiabetic model. Too bad no one quite knows how to run Shor's algorithm on those.

[@annaclarafenyo8185]

@@mathstrek It's impossible to use D-wave machines as quantum computers, their business model is basically pure fraud.

[@zavvie809]

@@annaclarafenyo8185 which is why he said "depends on how gullible you are about the hype from quantum computer manufacturers" - Dwave advertises themselves as quantum computers with 2000qbits! Now it's just up to how naive you are to trust them...

[@annaclarafenyo8185]

@@zavvie809 It's not acceptable to even suggest it is possible. You have to explicitly say that any such claim is fraudulent. Anything less is irresponsible.

[@varunachar87]

@@annaclarafenyo8185 he is not going to be heard as an expert on the matter. His reach is to an audience that completely understands this to be more of a friendly but skeptical rant than a serious assessment of the state of quantum computing.

[@SpencerBliven]

An entertaining video, but in the end it’s just another crack pot theory!

[@migarsormrapophis2755]

yeeeeeeeeeeeeeeeeeeeee

[@pythonwire]

I do not drop comments often BUT you are seriously wrong! Big dislike indeed

[@EdwinSteiner]

You forgot your argument.

[@shaimach]

Wrong! Before Google's experiment there where ZERO experiments in which quantum computers are better than classical one. After Google there was one. Everybody in the field knows this is the first step. But it is a very significant one. And how the quantum computer achieves this is important - quantum gates, readout, etc. The same sort of operations you need for general purpose quantum computation - just a small number of such operations. This is a straw-man argument.

[@caspermadlener4191]

Even if this person was right, never start a comment with "Wrong!", because it will get disliked to hell.

[@asherasher9249]

@@caspermadlener4191Wrong!

[@mishaerementchouk]

This is exactly what his “teapot supremacy “ analogy is about. Now, two years later, still no one came up with the “clever application” (as was put in the Google’s paper) of the sampler’s quantum superiority.