Scott Aaronson is a breath of fresh air when it comes to explaining QC. It was painful to watch the Microsoft keynote a month or two ago where the "parallel solutions" explanation was given several times as the power of QC.
But if each possible solution is represented by a unique sum of probability amplitudes (with the goal that only one of those sums is non-zero) then why isn't it reasonable to describe that as "all the possible answers are tried at once"?
@@puffin88 "Eliminating all the wrong answers" (through destructive interference of the probability amplitudes) would be a much better simplistic description.
Excellent talk! It’s good to hear someone from the cutting edge explain things in layman's terms rather than relying on science writers or journalists to do it.
We are currently at the “vacuum tube stage” in development of quantum bits and quantum computing... and, there are several methodologies being studied leading to development of quantum-computing... some want to pursue photonic chips, some want to create nano antennas or rectennas, some want to study social organization of quantum bits, some want to grow organic computers from graphene, some want to move on from ones and zeros... some just want more research money.
Quantum computing is something I kind of want to grasp at least the outlines of but often end up chucking in the too-hard basket. Short explainers like this help me, so thanks.
My friend you are in a quantum super position as we speak. In another universe you achieve the unachievable in another superposition, but thats all parallel and doesnt matter
Excellent and interesting talk! I wonder - is there anyone working on the so-called non-conservative hypothetical of - what happens if quantum computing can't be done? I mean obviously the way in which things fail will be informative, but I wonder what are the serious theories *against* quantum computing working, and if there are any, would they be testable in some other way?
So algorithms like prime factorisation (shors algorithm) or searching, or a few others are proven to be possible. So it is proven that there are problems that quantum computers can work on. We also can experimentally achieve quantum computing with a very limited number of quantum bits - quantum computers do technically already exist, it’s just that they have like 5 working (entangled) bits at max. The point is that quantum computers can be proven to work and have been built on the small scale. Why would you assume they don’t work? EDIT: this is coming from someone who works in a lab where we do have quantum entangled qubits, and the initial conditions in our experiments states are reproducible, although currently we have to improve our fidelity (basically accuracy) and need to add one component to our ion trap called single ion addressing to turn it into a universal quantum computer. It is an extremely strange viewpoint to take that quantum computers may be impossible, because the science on how they work and the experimental tests of it are so well fleshed out...
In order for a QC to "compute" something, the qubits have to be completely isolated from the environment they are in so they only "react' with each other. This will be shown to be impossible because everything interacts with everything else at some fundamental level, so a QC will never be able to achieve the "quantum speedup" that the current theories claim they will.
High Physics postgrad here. What the guy was referring to is not so much whether Quantum Computers can or cannot do certain things, but that if it turns out to be essentially impossible to create quantum computers, that would mean that physics as we know it is wrong. However small quantum computers have been successfully created, so I do not think that is likely to happen.
I want to ask the question in a slightly different way to perhaps get a more efficacious/relevant response. For me, it is not so much pertinent to ask if they can work, but rather if they can be efficient enough to put to use. If it takes more work, space, effort, energy, and material to make them work as opposed to a classical computer, then we're wasting our time. So the proper question is: can they be more efficient than what we have?
I'm not a physicist or even a very good electronics engineer or computer scientist (although I do try hard with the latter two)... but I get the feeling that Quantum Computing is going to be CS's alternative to economical / sustainable fusion power.... It's going to be one of those fields that requires more money, more research, more money, more research, more money, more research and is perpetually "just a couple of years away now"
Ok n what about d-wave and Geordie Rose in his explanation and I quote " what the machine does is extract resources from these other parallel dimensions or universe" ??? My problem is he's the only one that talks about what these things do in that regard nobody else has touched on this!!
All that would happen is someone would just make it sound more sophisticated and complex, then make up a couple words to explain small nonsense no one cared about and doesn't make any change only waste more time and get high with the private funding they rack in
Perhaps we could agree to call it "Hilbert Computation", given that it's all about Hilbert spaces and the non-commutative algebras of operators that act upon them, not specially about woo-oo quantum theory? One has Hilbert spaces and non-commutative algebras of operators in classical signal analysis, for example, because of the omnipresence of fourier analysis, so that could equally well be a model of the Hilbert space mathematics. Hilbert computation perhaps should be taken to imply that we use mostly or only projective operators, so it's a particular restriction away from signal analysis in full generality, but hey. But then, QFT=signal analysis -modulation of the vacuum state and measurement of those modulated states- is my thing.
At 5:35 Dr. Aaronson suggests n^2 steps to factor an N-digit number. I can just about imagine a log(n) step solution for an N-digit number using traditional counting methods. I can also imagine a O(1) solution for a dynamic circuit. Need to build a system to test and prove it of course. Some friends and a whiteboard to work out a specific method I have in mind.
No you can't. This is prime Dunning-Kruger effect; if what you said is true then you could make millions and completely overhaul the field of cryptography. I'll be waiting for your paper.
You can check a solution to a Sudoku problem very quickly: just check that there are no repeated numbers in each row/column/square. This can be done in linear time, which is polynomial time, therefore Sudoku is in NP.
I divided the video into 30 second segments, chose six of those randomly for detailed analysis of the number of "uuuu"s and "um"s, and extrapolated to the full video length. My estimate is about 200 "uuu"s and "um"s, with standard error +/- 20. Not bad, but probably not close to a record for a TEDx talk.
But what you weren't able to count was the number of simultaneous uuu's and um's whose amplitudes cancelled each other out allowing you to see the solution...it was very meta of him....
And the dudes referring to people like Elon Musk and myself who think every bit of this is out of control and we need to take a step back and make sure that we are creating a future where we can coexist with all of this new tech... I refer you to Geordie Rose go listen to the lecture he gave in Canada quite frankly his statements were terrifying
La verdad siempre esperé gue Dios la revelaría un día tantos años en secreto siempre acompañado por aviones y helicópteros siempre que salía en el mayor secreto le doy gracias a Dios por la verdad gracias Pedro x desde Miami
Rather than disproving something people say, how about setting the goal to understanding how quantum particles work. It is science afterall. He himself ended the talk with advocating accurately learning & reporting, there's a lot of dishonesty going on.
This guy is hilarious! "But why is he pacing back and forth and saying ugh all the time" some would ask. Ugh because on planet Earth humans express emotions and make jokes; it's hard to explain. It's not a cause for concern though.
Good talk, but having my PhD in probability, I'd like it to be known that indeed probabilities CAN be negative. The interpretation goes like this: Let E be the event where a tree falls in the forest, makes a sound, and we do not observe the sound. Then by classical probability, the P(E)=0 since by the definition of E, we cannot observe it. Yet, we know this E happens quite often. We say then that E is owed an observational debt, that is P(E)
@@andresgoens If you're referring to the quantum wave function, then yes. In fact, the literature I am aware of discussing negative probabilities make specific mention to the quantum world where negative probabilities are not a contradiction nor evidence of a mistake, but the norm where it would be odd to not have them. Of course I'm reminded of a paper showing that for any generalized random variable X (one w/ negative probabilities), there exists classical random variables Y and Z such that X+Y=Z. What this implies is that negative probabilities suggests a problem in measurement (i.e. we've defined and are measuring the wrong variable X as opposed to X+Y which would be classical). It's an interesting idea to think that QM physicist are simply measuring the wrong thing.
Hilarious guy. Very interesting presentation. What about using these quantum properties to communicate large distances faster than the C? Pipe dream or is there something there?
In 1969 everyone though that before 2000 we will live in other planets, but we are very far from that. Quantum seems more like a dream than a fact, controlling energy at such deep details is kind of science fiction.
This guy like all the rest who get up in front of a bunch of people and try to make some sense of quantum computing. They really don't know what they have or its true intention they stumble over themselves to convince people that it's a good thing, what they don't believe is the chaos quantum computing is going to lead too.
How embarassing. hahaha When someone has double-talk propensity they can babble on about protons and other fantastical excerpts from bs books; what is the practical application of your Nerd Wet Dream.
Scott Aaronson is a breath of fresh air when it comes to explaining QC. It was painful to watch the Microsoft keynote a month or two ago where the "parallel solutions" explanation was given several times as the power of QC.
Go to Y Combinators channel, they have life saving content
But if each possible solution is represented by a unique sum of probability amplitudes (with the goal that only one of those sums is non-zero) then why isn't it reasonable to describe that as "all the possible answers are tried at once"?
@@puffin88 "Eliminating all the wrong answers" (through destructive interference of the probability amplitudes) would be a much better simplistic description.
Excellent talk! It’s good to hear someone from the cutting edge explain things in layman's terms rather than relying on science writers or journalists to do it.
We are currently at the “vacuum tube stage” in development of quantum bits and quantum computing... and, there are several methodologies being studied leading to development of quantum-computing... some want to pursue photonic chips, some want to create nano antennas or rectennas, some want to study social organization of quantum bits, some want to grow organic computers from graphene, some want to move on from ones and zeros... some just want more research money.
One of the best TEDx talks I've seen.
Quantum computing is something I kind of want to grasp at least the outlines of but often end up chucking in the too-hard basket. Short explainers like this help me, so thanks.
My friend you are in a quantum super position as we speak. In another universe you achieve the unachievable in another superposition, but thats all parallel and doesnt matter
Next to Sabine Hossenfelder talking about quantum physics, this is the next best thing I have listened to.
this guy is our hero
Am I the only one who thinks this guy is freaking hilarious ? hahahahahahaha
Smarter than you
You should check out "Quantum Computing Since Democritus", you'll get plenty of this humour
Excellent and interesting talk!
I wonder - is there anyone working on the so-called non-conservative hypothetical of - what happens if quantum computing can't be done? I mean obviously the way in which things fail will be informative, but I wonder what are the serious theories *against* quantum computing working, and if there are any, would they be testable in some other way?
So algorithms like prime factorisation (shors algorithm) or searching, or a few others are proven to be possible. So it is proven that there are problems that quantum computers can work on. We also can experimentally achieve quantum computing with a very limited number of quantum bits - quantum computers do technically already exist, it’s just that they have like 5 working (entangled) bits at max.
The point is that quantum computers can be proven to work and have been built on the small scale. Why would you assume they don’t work?
EDIT: this is coming from someone who works in a lab where we do have quantum entangled qubits, and the initial conditions in our experiments states are reproducible, although currently we have to improve our fidelity (basically accuracy) and need to add one component to our ion trap called single ion addressing to turn it into a universal quantum computer. It is an extremely strange viewpoint to take that quantum computers may be impossible, because the science on how they work and the experimental tests of it are so well fleshed out...
Great talk .... very refreshing very informative
In order for a QC to "compute" something, the qubits have to be completely isolated from the environment they are in so they only "react' with each other. This will be shown to be impossible because everything interacts with everything else at some fundamental level, so a QC will never be able to achieve the "quantum speedup" that the current theories claim they will.
High Physics postgrad here. What the guy was referring to is not so much whether Quantum Computers can or cannot do certain things, but that if it turns out to be essentially impossible to create quantum computers, that would mean that physics as we know it is wrong. However small quantum computers have been successfully created, so I do not think that is likely to happen.
I want to ask the question in a slightly different way to perhaps get a more efficacious/relevant response.
For me, it is not so much pertinent to ask if they can work, but rather if they can be efficient enough to put to use. If it takes more work, space, effort, energy, and material to make them work as opposed to a classical computer, then we're wasting our time. So the proper question is: can they be more efficient than what we have?
This is the first talk about quantum computers that shed at least some light on this topic. Great talk, thank you very much !
Hi Claire
Nice name
This dude is awesome
That seasoning analogy is great. I’m gonna use it for AI for sure.
Scott, great talking !!! Very good in explaining something that everyone is filling their mouths saying it's magic!!!!!!!
I'm not a physicist or even a very good electronics engineer or computer scientist (although I do try hard with the latter two)... but I get the feeling that Quantum Computing is going to be CS's alternative to economical / sustainable fusion power.... It's going to be one of those fields that requires more money, more research, more money, more research, more money, more research and is perpetually "just a couple of years away now"
Ok n what about d-wave and Geordie Rose in his explanation and I quote " what the machine does is extract resources from these other parallel dimensions or universe" ??? My problem is he's the only one that talks about what these things do in that regard nobody else has touched on this!!
this guy makes me wanna go into algorithm research
This talk is still modern and always funny.
Gracias todo el universo es vidad inteligente total millones de años y que evolución gracias Pedro x desde Miami
Great Video, by the way Google has launched it's 72 qubit machine in March 2018.
Damn, I might enter QC research in a couple of years. I just hope that it is not too developed by then hahah
I will develop further
Thanks for an interesting talk!
Fantastic!
Good talk.
He is not nervous 😬.. it's his shtoyl...he is varsity professor and has great public speaking skills. See his other talks
I love it
More exciting if it doesn't work
Re write all the physics books
All that would happen is someone would just make it sound more sophisticated and complex, then make up a couple words to explain small nonsense no one cared about and doesn't make any change only waste more time and get high with the private funding they rack in
Amazing lecture considering its duration to be
just 15 minutes long
Le doy la razón, hay muchos locos de plaza, hablando de física cuántica, dimensiones y espiritualidad...
I love it
Perhaps we could agree to call it "Hilbert Computation", given that it's all about Hilbert spaces and the non-commutative algebras of operators that act upon them, not specially about woo-oo quantum theory? One has Hilbert spaces and non-commutative algebras of operators in classical signal analysis, for example, because of the omnipresence of fourier analysis, so that could equally well be a model of the Hilbert space mathematics. Hilbert computation perhaps should be taken to imply that we use mostly or only projective operators, so it's a particular restriction away from signal analysis in full generality, but hey. But then, QFT=signal analysis -modulation of the vacuum state and measurement of those modulated states- is my thing.
you missed the physics there. You need physical objects to operate in Hilbert space i.e., quantum mechanics.
Entertaining!
At 5:35 Dr. Aaronson suggests n^2 steps to factor an N-digit number. I can just about imagine a log(n) step solution for an N-digit number using traditional counting methods. I can also imagine a O(1) solution for a dynamic circuit. Need to build a system to test and prove it of course. Some friends and a whiteboard to work out a specific method I have in mind.
No you can't. This is prime Dunning-Kruger effect; if what you said is true then you could make millions and completely overhaul the field of cryptography. I'll be waiting for your paper.
Will quantum computers run some kind of centralized super A.I.?
Would it be wrong to call him a lil' adorkable?
When carl grows up after stealing Jimmy's computer project.
AT 7:24 : I am not sure that SUDOKU is NP ?
You can check a solution to a Sudoku problem very quickly: just check that there are no repeated numbers in each row/column/square.
This can be done in linear time, which is polynomial time, therefore Sudoku is in NP.
The study of what I can’t do guy
QUMPUTER - From now on, let's call them QUMPUTERS. Much better than the tongue twister 'Quantum Computers'!
I divided the video into 30 second segments, chose six of those randomly for detailed analysis of the number of "uuuu"s and "um"s, and extrapolated to the full video length. My estimate is about 200 "uuu"s and "um"s, with standard error +/- 20. Not bad, but probably not close to a record for a TEDx talk.
Did you count how many times he moves his weight from left foot to right foot
But what you weren't able to count was the number of simultaneous uuu's and um's whose amplitudes cancelled each other out allowing you to see the solution...it was very meta of him....
But did he break the dancing intensity record?
Peter Scott very meta is an understatement. Id estimate extremely, maybe even at an ultimate level of meta.
You could have Quantum Computer to do that!
7:23
i like his shirt
Who did he refer to as not embracing the future?
Trump, and presumably his voters too.
If u think Trump is the problem your way off
And the dudes referring to people like Elon Musk and myself who think every bit of this is out of control and we need to take a step back and make sure that we are creating a future where we can coexist with all of this new tech... I refer you to Geordie Rose go listen to the lecture he gave in Canada quite frankly his statements were terrifying
Charles Bingham that is an interesting remark. Maybe Trump is the symptom, not the disease. Thanks for making me realize that.
It doesn’t lie
Cool dude
PODERIA SER USADO PARA ENCONTRAR A CURA DAS DOENÇAS E PROBLEMAS SOCIAIS PARA JUSTIFICAR OS GASTOS
La verdad siempre esperé gue Dios la revelaría un día tantos años en secreto siempre acompañado por aviones y helicópteros siempre que salía en el mayor secreto le doy gracias a Dios por la verdad gracias Pedro x desde Miami
Ah damn why did I go for Biology :(
Rather than disproving something people say, how about setting the goal to understanding how quantum particles work. It is science afterall. He himself ended the talk with advocating accurately learning & reporting, there's a lot of dishonesty going on.
Attractive guy!
Good to see his passion on the subject but unfortuntelly the results are the same: too much talk, and nothing was said.
This guy is hilarious! "But why is he pacing back and forth and saying ugh all the time" some would ask. Ugh because on planet Earth humans express emotions and make jokes; it's hard to explain. It's not a cause for concern though.
B''H, Christ, Jesus, Amen.
Very insteresting topic, but this guys seemed like he was nervous or something. It was hard to watch at points.
He's not nervous, this is just how he is
Peer Svensson Sorry, didn't mean to offend. Very interesting topic!
Peer Svensson I'm glad he spoke about it. He's probably an extremely intelligent person. Much more intelligent than I am!
Oh no, none taken :)
:)
uuuuuuuuuuuuuuu
That's why it's called Quantuuuuuuuuum Computing!
Quantum computing isn’t computing
it is in a sense
Good talk, but having my PhD in probability, I'd like it to be known that indeed probabilities CAN be negative. The interpretation goes like this: Let E be the event where a tree falls in the forest, makes a sound, and we do not observe the sound. Then by classical probability, the P(E)=0 since by the definition of E, we cannot observe it. Yet, we know this E happens quite often. We say then that E is owed an observational debt, that is P(E)
that's interesting! does this generalization to negative probabilities correspond to real negative amplitudes in QM?
@@andresgoens If you're referring to the quantum wave function, then yes. In fact, the literature I am aware of discussing negative probabilities make specific mention to the quantum world where negative probabilities are not a contradiction nor evidence of a mistake, but the norm where it would be odd to not have them. Of course I'm reminded of a paper showing that for any generalized random variable X (one w/ negative probabilities), there exists classical random variables Y and Z such that X+Y=Z. What this implies is that negative probabilities suggests a problem in measurement (i.e. we've defined and are measuring the wrong variable X as opposed to X+Y which would be classical). It's an interesting idea to think that QM physicist are simply measuring the wrong thing.
Ahahahahaha.....
Hilarious guy. Very interesting presentation. What about using these quantum properties to communicate large distances faster than the C? Pipe dream or is there something there?
Uhuhuhuhuhuhuhuuuuuhuuhuhuhuhuhuhuhuhuh
In 1969 everyone though that before 2000 we will live in other planets, but we are very far from that. Quantum seems more like a dream than a fact, controlling energy at such deep details is kind of science fiction.
not too good
UUUUMM UUUUMM M AAAAA
This guy like all the rest who get up in front of a bunch of people and try to make some sense of quantum computing. They really don't know what they have or its true intention they stumble over themselves to convince people that it's a good thing, what they don't believe is the chaos quantum computing is going to lead too.
How embarassing. hahaha When someone has double-talk propensity they can babble on about protons and other fantastical excerpts from bs books; what is the practical application of your Nerd Wet Dream.
Yes, the topic he's talking about it so interesting but this guy has zero speech skills, is very annoying.
The study of what I can’t do guy