I've wanted to do a video on these guys for literally 5 years. Let me know what your think! And check out shortform and get a FREE trial and $50 OFF the annual plan! at www.shortform.com/DrBen
When the two photons reach the end of the fiber optic cable and need to be separated again (transition from time to space separation) how is that done? How do they make sure the first photon goes through the delay circuit and the second, which has already been delayed, doesn't?
it is no secret that part of quantum computing will require some form of "time cloaking" computing Though I am loath to use the term "time cloaking" when all the tech really does is compute signal processing at binary oracle machine scale of "cloak" detection A true time cloak, to my view, should be able to obfuscate the detection of mass when encrypted with constants like the speed of light
btw...yes, the term "utility scale quantum" is not lost on me as not a proper use of quantum parallelization The term "quantum computing" seems a bit tortured in the context of signal processing that this "top secret" company is using
I’m blown away that a single photon can reliably be isolated and detected then gated for signaling. Props to the development team for their dedication and perseverance!! Truly amazing!
@JuttutinI thought I was so cool and on the cutting edge back in the late 90’s because I was installing OC-192’s with DWDM…we could combine Single-mode long range transceivers at 1550 & 1310 with multimode drops for private line/local data traffic. To see this quantum tech on fiber at this scale just makes my jaw drop….brings new meaning to single mode.
This is beyond insane. How can one species be both capable of taming quantum randomness at the level of individual photons, and simultaneously incapable of agreeing that the earth isn't flat.
There is a lot more stupid humans than non stupid humans. It’s that simple. Not every brain-body is created the same. Doesn’t mean that one person with less intelligence is anymore or less valuable than a person with high intelligence. Someone has to make chick-fil-a or Panera for the smart folks working 16hrs a day to contain and control quantum properties, because they won’t always make a lunch or coffee for themselves. Every roll/person/job has a place. What really matters, is purpose. A person waking up with and for a purpose. Not matter what that is.
My operating theory at this point is that the ads being sold for a video matching what ads you're a desired audience to is critical for content from subscriptions to show up in your feed timely or not. And when they show up days or weeks or months late its because they now have a desired overlap of ads that match what ads the algo wants you to see. I notice none of this when im on YT premium then bam its back when I dont pay for a year
In all reality, we are already there with electronics. The normal person has absolutely zero understanding of how these devices work. Most people can’t even explain the function of some of the most basic components.
@theniiitoo much. edit: Tried looking up how much it's become a part of the economy and didn't get much useful data than "omg ai Is the next next thing so many people are using it (Forced to)!!!1!!1!!" but the amounts I've seen range from 1.5%~ to 35%~ depending on if you count how much is in the stock market of ai companies to how much money ai stuff in general has brought to the US total GDP which is the lower end estimates around 1.5%~ to 3.5%~ which means theres a whole flippin lot being handed around rapidly (between the same few companies) which is... making it look like they're making billions but its the same three dudes handing a 20$ bill in a circle to pay for each other's "product" which makes others outside of the circle jerk to invest into the bubble and it is unsustainable cause one day people are going to DEMAND a final product and once any of the three collapse it all collapses taking everyone involved with it except for the people at the top who leave with the bulk of the money outsiders had put into it
I maybe only understood half of this Video but I'm still amazed at the small scale of single Photons they are operating at and trying to deciver anything from.
I don’t think, if I’m being honest, that I understood any of it. Too things amaze me. 1. That some people are smart enough to tackle such mind bogglingly complex issues. 2. That a bricklayer has a son-in-law (physics and engineering from Cambridge) who’s able to articulate such things to him.
Size of photons... Hehehe. Don't say that around my science teachers. Wouldn't it suck if our brains worked on both digital, analog, quantum at the same time while also using electrical, light, AND vibrations between it all.
The control of matter on the smallest scales that humans have achieved is more impressive than any magic from a fantasy book. Computer chips were already a miracle, this is just another level.
10:25 "We call this second photon the 'Sacrificial Photon'." (PR department hears this, has a panic attack.) "Or, after an emergency internal PR meeting, the 'Heralding Photon'." 🤣 This reminds me of when the Cape of Storms was renamed the Cape of Good Hope.
Assuming stable within tolerance temperature conditions, which makes the method very expensive compared to classical computational scales I mean, it is interesting tech to be sure, but not exactly a full comprehensive machine code quantum computer that has sorted error correction for classic compute algos at a quantum scale I would say this is more like a decent assembly path to pursue
@peanutpete assuming there is such a thing, they can only reclaim a bit of that in classical terms and it has more to do with processing speed than parrelized efficiency as a classical result
@memegazer - it does however also open up photonic computing at a rather large scale, not just quantum computing. One could see the technology approach used here to have a hybrid of photonic circuits (performing classical world computing, at very high speed, and lower energy requirement though) and quantum processing on the same system. The classical computing via photonics side could handle grunt work classic compute algos (at speed) whilst the quantum side can process problems (both pre-existing and novel) more adapt at solving using quantum computing. That way one gets a very fast classical computer (with a lower power budget), without having to waste quantum processing cycles on emulating processing of classic computer algos, with the addition of being able to process problems suited for the quantum domain in parallel, all within the same architecture (if not the same silicon die).
@MinusMedleyReally?! Did you watch the video? It seems legit and promising. Much better than a lot of fusion and nuclear companies which are clearly bulls@t
The animation and explanation at 21:27 gives the impression that the purpose of repeating the experiment is to eliminate noise and that the most frequent result is the correct answer. While eliminating noise is part of it, it’s important to understand that in quantum computation the result is actually the probability distribution itself, so you would have to do this repetition even in a flawless system.
Imagine the future when cameras use single photon detectors in order to record everything that came in with no noise. Kids in 100 years are going to have such cool stuff.
it is genuinely impressive how well you explained what is essentially "magic". Well done, that is talent. Kudos to you and your team for the quality of content!
This is absolutely mind blowing! It really feels like actual magic, but it's even more absurd to know that it's in fact what a group of genius dedicated people can accomplish in ten years of hard work. Amazing video!
Kudos! You've done such a great job explaining that I almost feel like I understand. Now I'm off to "explain it to my grandmother" to see if I do. Thanks.
I actually understood that and thats amazing that you could break down such an enormous complicated topic in an understandable way. Thanks so much! I wonder if some kind of cascading photon detector would be possible. Like a domino effect where the photon is triggering other photons until the cascade produce so much photons that its detectable by much easier measures.
You're forgetting how much trouble hey went to just to isolate a single photon...remember the goal is to compute at the 'edge' of the speed of light, to do that accurately very close to the speed of light requires discrete photons to ensure absolute coherence in your data stream.
They would seem to be entangled but it might be that they are also entangled with third parties so detecting one doesn't collapse the wavefunction for the other
The original photon pair is surely entangled, but detecting one doesn't matter because nothing is happening to the photons before then. It’s only after the red photon is detected that the blue photon is moved to the area where it gets "split" into a new entangled state which then has calculations done to it.
please tell me these guys know that in their hunt for Quantum computation, they have created/innovated the perfect component list for building light based inference chips? Like, are they cashing in yet? I want affordable RAM again.
I hear you but I don't think this helps much. 1) Light is fundamentally linear (modulo some weird effects) and you need non-linear effects at each stage in your inference and training. That means you have to constantly read your light out redigitize it, apply your non-linear transform and then do the next matrix multiplication. This doesn't make it useless, many companies are betting big on it, but I strongly suspect that optimizing for doing quantum computation is very different than optimizing for doing large linear operations on huge numbers of inputs and reading out the results quickly and efficiently -- for instance these guys want single photons while the photonic accelerator folks want as many high quality photons as they can get for their energy budget. 2) Well luckily there are quite a few companies doing the photonic accelerator thing. Unfortunately, light based inference doesn't really help with the RAM problem as it only does the computation efficenctly you still need to store all the data somewhere. Indeed it's probably even true that if you decrease the cost for compute you increase ram demands!
@petergerdes1094you load and hold the weights in RAM, in a photonic cascade you send the weights as pulse strenghts, since the trigger is solid, the only thing changing is the initial signal, I hashed it all out. It would definitely help with the RAM issue. Like, a lot.
@petergerdes1094the difference is that the architecture I have in mind doesn't read out the intermediary results during inference, only the final layer. The math is solid, and they have already done the heavy lifting to enable light-controlled gates. The weights are looped, so you can just spin up and partially read it from a loop of light which is the perfect lenght relative to the size of the relay. Sounds super complicated, and it is, but at the same time, it's just more and less pew pew, very much not rocketscience, and inference cost would plumet. Training is still super expensive, untill experimentally modulated inference variation for gradual self improvement puroposes post deployment is the norm.
Yes!! Finally a QC concept that feels like it could change the world. Others, while brilliant, don't seem very scalable to me - this one does!! Thanks for sharing Ben!
the entangled pair is used to produce a single photon then later they create the superposition of that single photon in that double pathway part, that's the qubit they are after.
At that stage of the process, it doesn’t matter that it decoheres: They're just getting a source of individual photons. It’s only after they have the individual photons that they split them and calculate with entangled states.
I have such respect for the creation of things that let us do things. From a bike to a car to a bridge, its fascinating to see challenges overcome. The difficulty of every step you've described is hard to fathom. Let alone bringing them together to function as a system. I find the physics of everything to do with EMR particularly difficult to understand. You've got a unique skill Ben, explaining these processes to us common folk. Thanks for the video. Eventually here, my head will stop spinning from what you've presented.
It's only one photon at a time so only one qubit. However, the computers fundamentally don't seem to work like normal quantum computers as they just directly extract the probability distribution by iterating the system a huge amount of time
@luceafarul5740 That does not rule out the possibility of having many qubits on a single chip. All quantum systems have many Extra cores, so to speak for error correction
@jp34604 they can have mutliple qubit if they have multiple photons at the same time that become entangled but what the video is showing/talking about is single qubit (or even qudits or however they are called when they are in a superposition of more than 2 states)
Interesting, There is a similar company here in Colorado called Quantinuum that is using light as well. One of my former colleagues is now working for them. We worked together using fiber optical components to perform coherent optical mixing using these techniques.
10:30 How does the detector logic work downstream? presumably you wanna detect a photon in order to know that you have a clock signal on that line, but you only do the detection because you want to do something with that information (like open a gate or something), so how do you transmit a signal from that detector that is useful in any way when the blue photon is traveling at the speed of light, basically racing and being in front of the signal from the red photon ? Feels like even if you detect it, the blue photon already passed and you can't get "in front of it" no matter what you do (unless the wire of the blue photon has a longer path and loops back around to meet the end of the detector) 28:20 again, how does the red photon detector transmit information faster than the speed of the blue photon which is basically travelling at the speed of light? How does the information from the detector react the switching circuit before the blue photon?
Q1 it's obvious Q2 i think he said by using higher refractive material they are able to "slowdown" light enough to be able to activate the switch. To be honest, I'm sure none of what he's saying is actually how it works and it's super dumbdown by a lot to make it more simple and also not share too much public information.
@yungifezdifference is in the output variability. Normal bit computer can only output 0 and 1. But here, each curcuit/calculation can output and process probabilities between 0 and 1.
@DrBenMiles how are the paths switched at the other end of the tunnel to convert the time qubit to spatial qubit? there is no sacrificial photon here to actuate a detector that can trigger the BTO deflector. If we use signal from detector at source of tunnel, how does that electric signal travel so far in sync with the photons? So I don't understand how do they achieve the syncronization of the BTO deflectors at the two ends of the tunnel.
watch the video again. on the end side of the transition line, the "half photon" split particle that was faster (since it wasn't delayed in a maze), is then delayed in an equal size maze, synchronizing the photon halves before being combined together.
@panic-a-la-modethey work by persisting bits by toggling transistors. I’m not seeing how the qubits are persisted and then contribute to aggregate calculation. You need a “qubit resister”. That was not covered.
Just a follow up question: at 4:15 he states that light that we see from the cosmic microwave background still holds the same polarization as when the photons were created. How could you verify that? Because you can't possibly know their state before measuring them. Or is there just no way for the photons to change their state ?
As far as knowing something before measuring, that's why models are useful. There's no verifying our predictions of the early universe. There are only degrees of certainty as we extrapolate from what we do know. We have created a story of the early universe that is consistent with everything that we do know. We try to understand the limits of our knowledge, which takes us only so far as moments after the big bang. Before that is a mystery and could likely remain a mystery forever. From what we have gathered, there wasn't much to change the character of light as it escaped billions of years ago. There were some things at play, but the greatest influence was the expansion of the universe. This did lower the light's energy state from infrared to microwave, but the last thing to affect polarization happened in the moment it was released.
@REM-4444 I'm not a scientist so take this with a grain of salt, but afaik kelvin can't be measured in degrees like Celsius or Fahrenheit because a degree implies a step in an arbitrary scale (eg. 1°C being 1/100th of the temperature between the freezing and boiling points of water), and since kelvin is a unit of measurement based on the scientifically defined absolute zero and specific changes in thermal energy, the usage of "degrees" would be technically incorrect.
@Jediguil 5 kelvins, not 5 degrees kelvin; kelvin has no "degrees," it is an absolute scale and the word degree by its very nature implies something relative or proportional to something else.
Yes because the first people to get their hands on this technology will totally use it for stealing passwords rather than selling security to the richest people on earth.
I'am just asking myself, why theses people have even considered to use such an archaic method like heat for the creation of a phase shift within their wave guides. Whoever has carefully followed the topological insulator topic, will easily recognize, these topological objects are made from Cooper pairs, while the large band-gaps are surface effects. Cooper pairs have the unique ability to flip between a bosonic and fermionic state upon a control signal according to the Landau-plane (temprature, current-density and magnetic flux density). When excited by a material-specific RF-signal at resonance and polarized by a DC electrical or magnetic field, theoretically a periodic violation of the unviolable Pauli's principle would take place. Since this principle holds and the energy of each excited electron can neither be destroyed nor converted into heat, nature fights back and creates extra space by "diluting" spacetime. Extra, or diluted spacetime within a confined piece of spacetime is termed spacetime curvature. And according to the direction of the curvature, it creates gravitational red-, or blue-shift. Therefore frequency and phase of photons, being in superposition can be exactly controlled, what could simplify the whole system.
@global-village-elder Of course, there is a need to duck, because the described effect - every graduated engineer or physicist can replicate in a lab - and is contained within the accepted Lagrangian equation set of the accepted physical standard model, works on a microscopic quantum scale, up to large, multi-MW macroscopic proportions. Since there is a strong military and economic relevance, it took only one "wrong" term, reference, or person and the comment would go. This is the freedom of science.
INFORMATION IS (not) STORED IN THE BALLS (of individuated light particles but rather in the exact phase alignment between the superposed, spatially-encoded qubits and the discrete time bins the qubits are later converted into as time-encoded qubits.)
Wow, first you explained Majorana in a way I could understand (after two or three views) and now this. I feel like quantum computing expert 💪 Thanks, Ben! Btw. you explain things so much better then Veritasium and your jokes are superb 👌
Videos like this, much like the one on ASML's engineering, really highlight the true ingenuity of humanity. It’s awe-inspiring to watch, and I hope companies like these continue to succeed.
outstanding topic and explanation. i recall reading a journal article about optical circuitry years ago and the many challenges faced. very nice to see how they problem solve.
![The "Happiest" Scary Place - Subliminal [Full Game]](http://i.ytimg.com/vi/kb0_-LYvHbI/mqdefault.jpg)
![[Honkai: Star Rail на русском] Proi Proi (Cover by Sati Akura)](http://i.ytimg.com/vi/IJzQ4Ffahnc/mqdefault.jpg)
I've wanted to do a video on these guys for literally 5 years. Let me know what your think!
And check out shortform and get a FREE trial and $50 OFF the annual plan! at www.shortform.com/DrBen
When the two photons reach the end of the fiber optic cable and need to be separated again (transition from time to space separation) how is that done? How do they make sure the first photon goes through the delay circuit and the second, which has already been delayed, doesn't?
it is no secret that part of quantum computing will require some form of "time cloaking" computing
Though I am loath to use the term "time cloaking" when all the tech really does is compute signal processing at binary oracle machine scale of "cloak" detection
A true time cloak, to my view, should be able to obfuscate the detection of mass when encrypted with constants like the speed of light
btw...yes, the term "utility scale quantum" is not lost on me as not a proper use of quantum parallelization
The term "quantum computing" seems a bit tortured in the context of signal processing that this "top secret" company is using
"This tech is kinda cool."
yeah for sure, fermion condensate baby, pun intended, bc you can still dust for fingerprint conformity
Been wanting to ask for a while... what is the status for new rockstar scientist designs?
I’m blown away that a single photon can reliably be isolated and detected then gated for signaling. Props to the development team for their dedication and perseverance!! Truly amazing!
I mean the same can be done with other thingsz.,,
@Gam31n you may have missed the point of the video...
@JuttutinI thought I was so cool and on the cutting edge back in the late 90’s because I was installing OC-192’s with DWDM…we could combine Single-mode long range transceivers at 1550 & 1310 with multimode drops for private line/local data traffic. To see this quantum tech on fiber at this scale just makes my jaw drop….brings new meaning to single mode.
This is beyond insane. How can one species be both capable of taming quantum randomness at the level of individual photons, and simultaneously incapable of agreeing that the earth isn't flat.
Bc theres 8 billion of us
There is a lot more stupid humans than non stupid humans. It’s that simple. Not every brain-body is created the same. Doesn’t mean that one person with less intelligence is anymore or less valuable than a person with high intelligence. Someone has to make chick-fil-a or Panera for the smart folks working 16hrs a day to contain and control quantum properties, because they won’t always make a lunch or coffee for themselves. Every roll/person/job has a place. What really matters, is purpose. A person waking up with and for a purpose. Not matter what that is.
There's a lot of space under both ends of the bell curve.
@DLWELDhighest σ of any human distribution.
Because we haven’t cured schizophrenia.
generally youtube would recommend this after 7months of release. Surprised I got this so early!!!
Ye same here. I don't understand why, but I'm happy!
Same here.
Just subscribe and you get the recommendation right away
I have no idea why I'm not getting this incredible video in 2027. What's wrong with RUclips???
My operating theory at this point is that the ads being sold for a video matching what ads you're a desired audience to is critical for content from subscriptions to show up in your feed timely or not. And when they show up days or weeks or months late its because they now have a desired overlap of ads that match what ads the algo wants you to see. I notice none of this when im on YT premium then bam its back when I dont pay for a year
"Instead of encoding in space, they encode in time"
That's a fancy way to describe serialization
But here we have a serialization of probabilities ;)
To the RUclips algorithm, I want more videos like this
Same I also want videos like this
"Any sufficiently advanced technology is indistinguishable from magic" - Arthur C. Clarke, 1962
In all reality, we are already there with electronics. The normal person has absolutely zero understanding of how these devices work. Most people can’t even explain the function of some of the most basic components.
Isaacmantx no need to call us dumb like that 😭
my favorite quotation bro
Thank you so much for making complex subjects so easy to comprehend.
You're welcome! 🙏
I wouldn't call that easy but yeah I have fish brain. You're awesome man.
meanwhile the whole economy is around making increasingly realistic videos of Will Smith eating pasta
Next we add cheese and a lot of pepper.
Quantum computing already has it's own smaller speculative market tbf.
@chandaman95 and how much of the us economy is based on it?
It's integral
@theniiitoo much.
edit: Tried looking up how much it's become a part of the economy and didn't get much useful data than "omg ai Is the next next thing so many people are using it (Forced to)!!!1!!1!!" but the amounts I've seen range from 1.5%~ to 35%~ depending on if you count how much is in the stock market of ai companies to how much money ai stuff in general has brought to the US total GDP which is the lower end estimates around 1.5%~ to 3.5%~ which means theres a whole flippin lot being handed around rapidly (between the same few companies) which is... making it look like they're making billions but its the same three dudes handing a 20$ bill in a circle to pay for each other's "product" which makes others outside of the circle jerk to invest into the bubble and it is unsustainable cause one day people are going to DEMAND a final product and once any of the three collapse it all collapses taking everyone involved with it except for the people at the top who leave with the bulk of the money outsiders had put into it
Wow!!! How does this ONLY have 95k views ?!?!?!?!!
Shadow band by nonsense....
Not everyone has the intellect to comprehend, nor the luck to come across this with all the noise in the world
I maybe only understood half of this Video but I'm still amazed at the small scale of single Photons they are operating at and trying to deciver anything from.
I don’t think, if I’m being honest, that I understood any of it. Too things amaze me. 1. That some people are smart enough to tackle such mind bogglingly complex issues. 2. That a bricklayer has a son-in-law (physics and engineering from Cambridge) who’s able to articulate such things to him.
Size of photons... Hehehe. Don't say that around my science teachers.
Wouldn't it suck if our brains worked on both digital, analog, quantum at the same time while also using electrical, light, AND vibrations between it all.
@RedactedDeathWhy dont they have a size?
@drlauch2256Uh...that guy didn't said photon have no size
Fantastic explainer, thank you Ben!
The control of matter on the smallest scales that humans have achieved is more impressive than any magic from a fantasy book. Computer chips were already a miracle, this is just another level.
10:25 "We call this second photon the 'Sacrificial Photon'."
(PR department hears this, has a panic attack.)
"Or, after an emergency internal PR meeting, the 'Heralding Photon'." 🤣
This reminds me of when the Cape of Storms was renamed the Cape of Good Hope.
Very cool to use the red photon detection to indirectly detect the useful blue photon.
Assuming stable within tolerance temperature conditions, which makes the method very expensive compared to classical computational scales
I mean, it is interesting tech to be sure, but not exactly a full comprehensive machine code quantum computer that has sorted error correction for classic compute algos at a quantum scale
I would say this is more like a decent assembly path to pursue
I figured they'd use the waste light.
@peanutpete
assuming there is such a thing, they can only reclaim a bit of that in classical terms and it has more to do with processing speed than parrelized efficiency as a classical result
@peanutpetethey would not, because using the waste light would not necessarily signal that any photon was created
@memegazer - it does however also open up photonic computing at a rather large scale, not just quantum computing.
One could see the technology approach used here to have a hybrid of photonic circuits (performing classical world computing, at very high speed, and lower energy requirement though) and quantum processing on the same system. The classical computing via photonics side could handle grunt work classic compute algos (at speed) whilst the quantum side can process problems (both pre-existing and novel) more adapt at solving using quantum computing.
That way one gets a very fast classical computer (with a lower power budget), without having to waste quantum processing cycles on emulating processing of classic computer algos, with the addition of being able to process problems suited for the quantum domain in parallel, all within the same architecture (if not the same silicon die).
This gives me hope for the future of science.
Vaporware, can bet it's only another 30 years away.
@MinusMedleyReally?! Did you watch the video? It seems legit and promising. Much better than a lot of fusion and nuclear companies which are clearly bulls@t
tech stagnation dnt be 2 optimistic.,,
@Gam31nYou think technology and science is stagnating?
@MinusMedley
Thoroughly stupid comment
That coffee setup tells me just how much funding they are getting.
I absolutely love that there's a computer that has a "world line" as part of it's design and construction.
As someone working with snspds for almost 10 years it’s cool to see that a big channel like this one discuss them!
Using light makes this so much easier to understand and your graphics.
They are secretly making the most innovative technology ever
The animation and explanation at 21:27 gives the impression that the purpose of repeating the experiment is to eliminate noise and that the most frequent result is the correct answer. While eliminating noise is part of it, it’s important to understand that in quantum computation the result is actually the probability distribution itself, so you would have to do this repetition even in a flawless system.
Thank you, that was helpful.
Cool to get such an insight into this and their actual state they are in.
15:00 I see MASTER YODAA!!!!
Imagine the future when cameras use single photon detectors in order to record everything that came in with no noise. Kids in 100 years are going to have such cool stuff.
Blue and Red photons?
Close enought, welcome hollow purple irl 😂
Superb and captivating video! :)
thanks for sharing a amazing knoledge
it is genuinely impressive how well you explained what is essentially "magic".
Well done, that is talent.
Kudos to you and your team for the quality of content!
4 minutes of video and my mind just blow up. So fucking genious to use light as a source of computation at quantum level.
5 years of asking your classmates for one yes is pretty in line 😂
Enlightening 😊 what a cool technology!
This is absolutely mind blowing! It really feels like actual magic, but it's even more absurd to know that it's in fact what a group of genius dedicated people can accomplish in ten years of hard work. Amazing video!
Very interesting, well explained as always. Thank you.
The "time encoding", made to make the photons resilient to noise across chip-to-chip transportation, is truly genius! Congrats!
This is the milestone. Quantum is here now.
Quantum computing before GTA6
Kudos! You've done such a great job explaining that I almost feel like I understand. Now I'm off to "explain it to my grandmother" to see if I do. Thanks.
I actually understood that and thats amazing that you could break down such an enormous complicated topic in an understandable way. Thanks so much! I wonder if some kind of cascading photon detector would be possible. Like a domino effect where the photon is triggering other photons until the cascade produce so much photons that its detectable by much easier measures.
You're forgetting how much trouble hey went to just to isolate a single photon...remember the goal is to compute at the 'edge' of the speed of light, to do that accurately very close to the speed of light requires discrete photons to ensure absolute coherence in your data stream.
this is like watching asml in their early days of development. amazing work!!
Nicely explained. Wouldn't the generated photon pair be entangled themselves? Detecting one of them would affect the survivor, no, yes, no?
I would also love to hear the answer
i don't think destroying one destroys the other.. i think it's just the state connection is lost.. i could be wrong through
They would seem to be entangled but it might be that they are also entangled with third parties so detecting one doesn't collapse the wavefunction for the other
The original photon pair is surely entangled, but detecting one doesn't matter because nothing is happening to the photons before then. It’s only after the red photon is detected that the blue photon is moved to the area where it gets "split" into a new entangled state which then has calculations done to it.
Yes, the photons are entangled. Detecting one confirms the presence of the other.
Fascinating that these guys have figured out the answers to so many complicated questions that they created. Bravo!!!
please tell me these guys know that in their hunt for Quantum computation, they have created/innovated the perfect component list for building light based inference chips? Like, are they cashing in yet? I want affordable RAM again.
I hear you but I don't think this helps much.
1) Light is fundamentally linear (modulo some weird effects) and you need non-linear effects at each stage in your inference and training. That means you have to constantly read your light out redigitize it, apply your non-linear transform and then do the next matrix multiplication.
This doesn't make it useless, many companies are betting big on it, but I strongly suspect that optimizing for doing quantum computation is very different than optimizing for doing large linear operations on huge numbers of inputs and reading out the results quickly and efficiently -- for instance these guys want single photons while the photonic accelerator folks want as many high quality photons as they can get for their energy budget.
2) Well luckily there are quite a few companies doing the photonic accelerator thing. Unfortunately, light based inference doesn't really help with the RAM problem as it only does the computation efficenctly you still need to store all the data somewhere.
Indeed it's probably even true that if you decrease the cost for compute you increase ram demands!
@petergerdes1094you load and hold the weights in RAM, in a photonic cascade you send the weights as pulse strenghts, since the trigger is solid, the only thing changing is the initial signal, I hashed it all out. It would definitely help with the RAM issue. Like, a lot.
@petergerdes1094the difference is that the architecture I have in mind doesn't read out the intermediary results during inference, only the final layer. The math is solid, and they have already done the heavy lifting to enable light-controlled gates. The weights are looped, so you can just spin up and partially read it from a loop of light which is the perfect lenght relative to the size of the relay. Sounds super complicated, and it is, but at the same time, it's just more and less pew pew, very much not rocketscience, and inference cost would plumet. Training is still super expensive, untill experimentally modulated inference variation for gradual self improvement puroposes post deployment is the norm.
so when are they gonna go public? Seems like prime time for getting public investors on board
15:07 Bachmann-Turner Overdrive
Very apt. With this computer one can literally describe it as "You Ain't Seen Nothing Yet".
Yes!! Finally a QC concept that feels like it could change the world. Others, while brilliant, don't seem very scalable to me - this one does!! Thanks for sharing Ben!
8:10 hmm... red+blue = ???
🟣
Crazy new stuff, amazing!
18:47 you didn't put a shadow in your animation under a photon, did you?
I think it's to clearly distinguish the paths they take
It's a misleading simplification, to actually show what is going on
You had me at space to time converter circuit. I also work with horizontal flip flop bubble valves.
10:30 Isn't the heralding photon entangled with the blue clock photon and so detecting the heralding photon causes decoherence in the clock?
the entangled pair is used to produce a single photon then later they create the superposition of that single photon in that double pathway part, that's the qubit they are after.
At that stage of the process, it doesn’t matter that it decoheres: They're just getting a source of individual photons. It’s only after they have the individual photons that they split them and calculate with entangled states.
Detecting the heralded photon won't cause decoherence. Only when you lose the heralded photon you lose information and the system becomes mixed
I have such respect for the creation of things that let us do things. From a bike to a car to a bridge, its fascinating to see challenges overcome. The difficulty of every step you've described is hard to fathom. Let alone bringing them together to function as a system. I find the physics of everything to do with EMR particularly difficult to understand. You've got a unique skill Ben, explaining these processes to us common folk. Thanks for the video. Eventually here, my head will stop spinning from what you've presented.
You never said how many qubits are on a chip? And how many chips are in a cabinet?
It's only one photon at a time so only one qubit. However, the computers fundamentally don't seem to work like normal quantum computers as they just directly extract the probability distribution by iterating the system a huge amount of time
@luceafarul5740
That does not rule out the possibility of having many qubits on a single chip.
All quantum systems have many
Extra cores, so to speak for error correction
@luceafarul5740
That's not true. It is still potentially possible to have multiple qubits on a single-chip.
@jp34604 they can have mutliple qubit if they have multiple photons at the same time that become entangled but what the video is showing/talking about is single qubit (or even qudits or however they are called when they are in a superposition of more than 2 states)
Interesting, There is a similar company here in Colorado called Quantinuum that is using light as well. One of my former colleagues is now working for them. We worked together using fiber optical components to perform coherent optical mixing using these techniques.
10:30 How does the detector logic work downstream? presumably you wanna detect a photon in order to know that you have a clock signal on that line, but you only do the detection because you want to do something with that information (like open a gate or something), so how do you transmit a signal from that detector that is useful in any way when the blue photon is traveling at the speed of light, basically racing and being in front of the signal from the red photon ? Feels like even if you detect it, the blue photon already passed and you can't get "in front of it" no matter what you do (unless the wire of the blue photon has a longer path and loops back around to meet the end of the detector)
28:20 again, how does the red photon detector transmit information faster than the speed of the blue photon which is basically travelling at the speed of light? How does the information from the detector react the switching circuit before the blue photon?
Q1 it's obvious
Q2 i think he said by using higher refractive material they are able to "slowdown" light enough to be able to activate the switch.
To be honest, I'm sure none of what he's saying is actually how it works and it's super dumbdown by a lot to make it more simple and also not share too much public information.
thats a lot of energy to keep it THAT cool
For me this just seems like a normal bit computer running closer to speed of light then 70% c in our computer(copper wire)
Yeah, I don't understand how superposotion or entanglement would be achieved here
@yungifezthe photon enters a superposition regarding which "path it takes" when they are split
@yungifezdifference is in the output variability. Normal bit computer can only output 0 and 1. But here, each curcuit/calculation can output and process probabilities between 0 and 1.
How long until encryption keys are broken?
the nsa already had a hardware backdoor for years now buddy 😅😢
Yesterday
This is one of the coolest videos I’ve seen of yours. Well done.
@DrBenMiles how are the paths switched at the other end of the tunnel to convert the time qubit to spatial qubit? there is no sacrificial photon here to actuate a detector that can trigger the BTO deflector. If we use signal from detector at source of tunnel, how does that electric signal travel so far in sync with the photons?
So I don't understand how do they achieve the syncronization of the BTO deflectors at the two ends of the tunnel.
watch the video again.
on the end side of the transition line, the "half photon" split particle that was faster (since it wasn't delayed in a maze), is then delayed in an equal size maze, synchronizing the photon halves before being combined together.
Great video and interesting logo for a company
I get that the qubits are being mass generated and transported, but how is one addressed and used to calculate something?
he mentioned logical gates like in a classical computer so performing operations using those gates is how it becomes useful
@panic-a-la-modethey work by persisting bits by toggling transistors. I’m not seeing how the qubits are persisted and then contribute to aggregate calculation. You need a “qubit resister”. That was not covered.
Just a follow up question: at 4:15 he states that light that we see from the cosmic microwave background still holds the same polarization as when the photons were created. How could you verify that? Because you can't possibly know their state before measuring them. Or is there just no way for the photons to change their state ?
I dont know
@urisinger3412 Ha ha, neither do I, LOL.
As far as knowing something before measuring, that's why models are useful. There's no verifying our predictions of the early universe. There are only degrees of certainty as we extrapolate from what we do know. We have created a story of the early universe that is consistent with everything that we do know. We try to understand the limits of our knowledge, which takes us only so far as moments after the big bang. Before that is a mystery and could likely remain a mystery forever.
From what we have gathered, there wasn't much to change the character of light as it escaped billions of years ago. There were some things at play, but the greatest influence was the expansion of the universe. This did lower the light's energy state from infrared to microwave, but the last thing to affect polarization happened in the moment it was released.
Very great video thank you for going into such depths while still explaining it fairly simple very interesting.
16:51 People of science we meet once again at yet another kelvin faux pas!
I'm uninitiated. Teach me.
@REM-4444 I'm not a scientist so take this with a grain of salt, but afaik kelvin can't be measured in degrees like Celsius or Fahrenheit because a degree implies a step in an arbitrary scale (eg. 1°C being 1/100th of the temperature between the freezing and boiling points of water), and since kelvin is a unit of measurement based on the scientifically defined absolute zero and specific changes in thermal energy, the usage of "degrees" would be technically incorrect.
@Jediguil ah, so it's just an issue of rounding or being imprecise. thanks!
@Jediguil 5 kelvins, not 5 degrees kelvin; kelvin has no "degrees," it is an absolute scale and the word degree by its very nature implies something relative or proportional to something else.
Really much enjoyed this video. Very cool stuff!
Posted 2 minutes ago, how am I so late
You're out of phase
Because your photons can’t see time.
@sudo_softlock -And you’re underrated 🫵
Nicely done.
My jaw dropped at 26:00. Space to time quantum converter sir whatt?! The line between physics and magic is blurred beyond measurement...or is it ?!
Amazing content! Thank you for sharing 👍🏾👍🏾👍🏾
Good to see you are shedding light on such an underdiscussed topic
I see what you did there
This is freaking awesome! I'm so ready for this.
Photon-based computing? it's an interesting concept
What like fiber optic?
WOW, I feel so blessed to get this insight into the next step of computing!
There goes your encryption. All your passwords and secrets now belong to us!
Yes because the first people to get their hands on this technology will totally use it for stealing passwords rather than selling security to the richest people on earth.
Ah yes you can’t encrypt it at all /s. This is the dumbest take I’ve ever heard.
This is amazing!
It took the Brits! Well done lads. Btw, that was the best explanation of Quantum computing ever. Two wins for the Brits! 😀
Great content as always Dr.Miles:) thank you.🙏
new setup just dropped, we don't need GPUs or CPUs or whatever anymore, we only need light!
Thanks! Thanks to the editing, and explanations, this was really easy to understand :)
I'am just asking myself, why theses people have even considered to use such an archaic method like heat for the creation of a phase shift within their wave guides. Whoever has carefully followed the topological insulator topic, will easily recognize, these topological objects are made from Cooper pairs, while the large band-gaps are surface effects. Cooper pairs have the unique ability to flip between a bosonic and fermionic state upon a control signal according to the Landau-plane (temprature, current-density and magnetic flux density). When excited by a material-specific RF-signal at resonance and polarized by a DC electrical or magnetic field, theoretically a periodic violation of the unviolable Pauli's principle would take place. Since this principle holds and the energy of each excited electron can neither be destroyed nor converted into heat, nature fights back and creates extra space by "diluting" spacetime. Extra, or diluted spacetime within a confined piece of spacetime is termed spacetime curvature. And according to the direction of the curvature, it creates gravitational red-, or blue-shift. Therefore frequency and phase of photons, being in superposition can be exactly controlled, what could simplify the whole system.
No need to duck,
@global-village-elder Of course, there is a need to duck, because the described effect - every graduated engineer or physicist can replicate in a lab - and is contained within the accepted Lagrangian equation set of the accepted physical standard model, works on a microscopic quantum scale, up to large, multi-MW macroscopic proportions. Since there is a strong military and economic relevance, it took only one "wrong" term, reference, or person and the comment would go. This is the freedom of science.
This has the potential to turn everything upside down and a real endgame for computer development. Watch where you invest for sure.
INFORMATION IS (not) STORED IN THE BALLS (of individuated light particles but rather in the exact phase alignment between the superposed, spatially-encoded qubits and the discrete time bins the qubits are later converted into as time-encoded qubits.)
Bunch of cool stuff! Will it really at the end of the say do anything?
great video!
You are sending a photon but also now how another form of bit in that of the BTO. These guys have an amazing job. 🎉
Exceptional video, thank you so much!! Love it!!!
Wow, first you explained Majorana in a way I could understand (after two or three views) and now this. I feel like quantum computing expert 💪 Thanks, Ben!
Btw. you explain things so much better then Veritasium and your jokes are superb 👌
amazing content my man
Wow! This is soooo cool.
Videos like this, much like the one on ASML's engineering, really highlight the true ingenuity of humanity. It’s awe-inspiring to watch, and I hope companies like these continue to succeed.
I'm in aw at these technologies and glad to be in the right time to maybe be working myself on them when working hard enough
I've always thought photon qbiys are the real future. But these guys have 3x level-upped the game. Thanks for the video and keep up the awesomeness.
outstanding topic and explanation. i recall reading a journal article about optical circuitry years ago and the many challenges faced. very nice to see how they problem solve.
Very bullish on photonics, thank you for the video. It is fantastic.
Amazing content. Thank you very much for exploring and explaining this. Very cool company!
Absolutely fantastic.
great reporting from somebody on the ground
Great video fantatstic thank you
Cant wait to have one of these in my pocket 😂