5 ms·

Is there a single person here interested in photonic computing that wants to explain to the class if there's any "there" there?

by mapt 6mo ago

brcmthrowaway 6mo ago
There's a lot of people here with esoteric knowledge of lasers, because they're generally incredible devices (along with masers). Someone should be able to comment. I wish we had a large laser manufacturing ability in the West. I would say 95% of lasers of all kinds are manufactured in China.
db48x 6mo ago
It’s like any other fundamental research: you don’t know how much it’s worth until people start using it to solve real problems. This is something that is literally impossible to guess ahead of time. The most abstract mathematical techniques could turn into a trillion–dollar industry (number theory begat RSA encryption which now underpins _everything_ we do). But I will say that precise control of laser wavelength is critical to today’s communication technologies. I doubt their new techniques will be useless.
- QQ00 6mo ago
  Hopefully the billions money in AI will find some of its to turn this into real life applications. AI inference would love some more faster more efficient communication. I mean, Photonic computing already got the attention of these big tech companies.
topspin 6mo ago
There is there there... The substance is they've created a way to fabricate a device that can make the optical frequencies they wish. That is useful: it means a designer isn't limited to frequencies that are economic to generate with existing techniques, which is a constraint that lasers currently struggle with: low cost, compact, efficient laser sources (the kind that fit on a chip, and are fabricated by cost effective processes,) only exist for a limited number of frequencies. The story is typical tech journalism pabulum, but the underlying paper does discuss efficiency. It's about what you'd expect: 35 mW -> 6 mW @ 485 nm, for example. An obvious use case is multimode fiber communication: perhaps this makes it possible to use more frequencies for greater bandwidth and/or make the devices cheaper/smaller/more efficient. But there are other, more exotic things one might do when some optical frequency that was previously uneconomic becomes feasible to use at scale.
- criticalfault 6mo ago
  I wonder if this could also work for (e)uv
  - jonplackett 6mo ago
    I wondered this too - why are you being downvoted for asking? All the difficulty to create that laser it seems fair enough to ask!
  - jmusall 6mo ago
    Probably not because EUV gets absorbed incredibly quickly by anything other than vacuum. This is why it is created in low density gas, thin liquid or solid samples (high harmonic generation) or electron clouds (free electron laser).
2ndorderthought 6mo ago
Depends on the cost. We already have variable wavelength lasers. We have had them for years. They are currently expensive, large, and not the easiest things to control electronically. I have an application in mind for this technology outside of photonic computing. Again, it depends entirely on price, tunability, bandwidth of the profile, etc. My understanding of the photocomputing field is limited but I never thought the major issues were wavelength related? Maybe someone can educate me. If anyone wants to send me one of these I would be pumped.
dado3212 6mo ago
I think it's more relevant for quantum computing. The ions we choose for ion trap quantum computers are in part due to what wavelengths are excitable by modified telecom lasers, because they're the wavelengths that are easiest to produce and where the most research/stability/miniaturization has been focused. If the laser wavelength is configurable to this degree then it no longer becomes a constraint, and maybe you can choose single ions with different characteristics.
nine_k 6mo ago
Immediately: * You can pack many more different colors into fiber optic communication lines. Every color carries a few tens of GHz in modulation, but the carrier light is in hundreds of THz; there's a ton of bandwidth not used between readily available colors. * You can likely do interesting molecular chemistry by precisely adjusting laser light to the energy levels of particular bonds / electrons. * Maybe you can precisely target particular wavelengths / absorption bands for more efficient laser cutting and welding, if these adjustable lasers can be made high-power.
- suzzer99 6mo ago
  * Concert lasers just got a lot cooler.
  - inetknght 6mo ago
    Concert tickets will still remain very hot though.
    - firtoz 6mo ago
      Even with the latest updates? > Jury Finds Live Nation Acts as a Monopoly in a Victory for States In a verdict that could have far-reaching consequences in the music industry, the live colossus that includes Ticketmaster was found to have violated antitrust laws. https://www.nytimes.com/2026/04/15/arts/music/live-nation-antitrust-trial-verdict-monopoly.html?unlocked_article_code=1.cFA.CBSL.KeDGgmPu-G7F&smid=url-share https://www.nytimes.com/2026/04/15/arts/music/live-nation-an...
- summa_tech 6mo ago
  Fiber has fairly narrow windows in which it is as transparent as it needs to be to go long distance. We're already pretty good at filling these windows with conventional semiconductor lasers. What this is actually interesting for is being able to access arbitrary atomic transitions, many of which are outside the range of conventional semiconductors (too short, usually - there's a big hole between green and red for semiconductors). That's why they talk about quantum stuff.
SilentM68 6mo ago
[flagged]
Lerc 6mo ago
Not an expert in the field but it seems to me the key points are. Generating any wavelength. (this article) Accurately measuring wavelength. (otherwise there's no information benefit to arbitrary wavelength generation) Wavelength insensitive holographic gates. (If they work on that frequency, and in a way that does not change the frequency) I don't know what properties such devices currently have Assuming all of those, your ability to compute increases to your ability to distinguish wavelengths. You could theoretically calculate much more in a way you could never detect, but then you get into some really interesting tree falling in a forest issues.
morphle 6mo ago
The short answer if there is any "there" there for photonic computing is no, maybe. You need to understand quantum physics[3,2]. For example, photonic computing, photonic logic does not have a switch equivalent as semiconducting (CMOS transistor) or superconducting (Josephson Junction JJ) but we have a photonic Mach Zener interferometer (MZI) and a photon detector. Photonics and superconducting electronics is always going to be much larger in size (and therefore more expensive) than semiconductors build from few atoms. In quantum physics photonics we have advantages like quantum impedance, you can replace wires with photon transmitters and photodetectors and thus switch with only a few photons instead of large numbers of electrons. With photonics you can have billions of cheap low power data channels instead of high power wire bundles. But MZI as JJ will probably always be a few orders of magnitude larger than transistors so switching is not going to be better, but interferometry is. Shorter answer still: just low power communications and information processing yes, computing no. Bulk CMOS manufacturing is still cheaper than all the alternatives we have discovered or invented, until we learn to manufacture atom by atom or compute with single photons or electrons (also dependent on molecule by molecule self-assembly), we will stay with CMOS and Moore's law. Just listen to David B. Millers[1] lectures [2], his lectures are a shortcut to reading all his papers[2] that explain it all, especially [3]. Email me, I'll give you a private lecture. Your question's anwer is/was a summary of our whole lives research [4]: [1] https://appliedphysics.stanford.edu/profile/35 https://appliedphysics.stanford.edu/profile/35 [2] https://www.youtube.com/@davidmillerscience https://www.youtube.com/@davidmillerscience [3] Attojoule Optoelectronics for Low-Energy Information Processing and Communication https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7805240 https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7805240 [4] Wafer Scale Integration Free Space Optics Computing https://www.youtube.com/watch?v=vbqKClBwFwI https://www.youtube.com/watch?v=vbqKClBwFwI

Is there a single person here interested in photonic computing that wants to explain to the class if there's any "there" there?

by mapt 6mo ago

Is there a single person here interested in photonic computing that wants to explain to the class if there's any "there" there?

brcmthrowaway 6mo ago

There's a lot of people here with esoteric knowledge of lasers, because they're generally incredible devices (along with masers). Someone should be able to comment. I wish we had a large laser manufacturing ability in the West. I would say 95% of lasers of all kinds are manufactured in China.

db48x 6mo ago

It’s like any other fundamental research: you don’t know how much it’s worth until people start using it to solve real problems. This is something that is literally impossible to guess ahead of time. The most abstract mathematical techniques could turn into a trillion–dollar industry (number theory begat RSA encryption which now underpins _everything_ we do). But I will say that precise control of laser wavelength is critical to today’s communication technologies. I doubt their new techniques will be useless.

QQ00 6mo ago
Hopefully the billions money in AI will find some of its to turn this into real life applications. AI inference would love some more faster more efficient communication. I mean, Photonic computing already got the attention of these big tech companies.

topspin 6mo ago

There is there there... The substance is they've created a way to fabricate a device that can make the optical frequencies they wish. That is useful: it means a designer isn't limited to frequencies that are economic to generate with existing techniques, which is a constraint that lasers currently struggle with: low cost, compact, efficient laser sources (the kind that fit on a chip, and are fabricated by cost effective processes,) only exist for a limited number of frequencies. The story is typical tech journalism pabulum, but the underlying paper does discuss efficiency. It's about what you'd expect: 35 mW -> 6 mW @ 485 nm, for example. An obvious use case is multimode fiber communication: perhaps this makes it possible to use more frequencies for greater bandwidth and/or make the devices cheaper/smaller/more efficient. But there are other, more exotic things one might do when some optical frequency that was previously uneconomic becomes feasible to use at scale.

criticalfault 6mo ago
I wonder if this could also work for (e)uv
- jonplackett 6mo ago
  I wondered this too - why are you being downvoted for asking? All the difficulty to create that laser it seems fair enough to ask!
- jmusall 6mo ago
  Probably not because EUV gets absorbed incredibly quickly by anything other than vacuum. This is why it is created in low density gas, thin liquid or solid samples (high harmonic generation) or electron clouds (free electron laser).

2ndorderthought 6mo ago

Depends on the cost. We already have variable wavelength lasers. We have had them for years. They are currently expensive, large, and not the easiest things to control electronically. I have an application in mind for this technology outside of photonic computing. Again, it depends entirely on price, tunability, bandwidth of the profile, etc. My understanding of the photocomputing field is limited but I never thought the major issues were wavelength related? Maybe someone can educate me. If anyone wants to send me one of these I would be pumped.

dado3212 6mo ago

I think it's more relevant for quantum computing. The ions we choose for ion trap quantum computers are in part due to what wavelengths are excitable by modified telecom lasers, because they're the wavelengths that are easiest to produce and where the most research/stability/miniaturization has been focused. If the laser wavelength is configurable to this degree then it no longer becomes a constraint, and maybe you can choose single ions with different characteristics.

nine_k 6mo ago

Immediately: * You can pack many more different colors into fiber optic communication lines. Every color carries a few tens of GHz in modulation, but the carrier light is in hundreds of THz; there's a ton of bandwidth not used between readily available colors. * You can likely do interesting molecular chemistry by precisely adjusting laser light to the energy levels of particular bonds / electrons. * Maybe you can precisely target particular wavelengths / absorption bands for more efficient laser cutting and welding, if these adjustable lasers can be made high-power.

suzzer99 6mo ago
* Concert lasers just got a lot cooler.
- inetknght 6mo ago
  Concert tickets will still remain very hot though.
  - firtoz 6mo ago
    Even with the latest updates? > Jury Finds Live Nation Acts as a Monopoly in a Victory for States In a verdict that could have far-reaching consequences in the music industry, the live colossus that includes Ticketmaster was found to have violated antitrust laws. https://www.nytimes.com/2026/04/15/arts/music/live-nation-antitrust-trial-verdict-monopoly.html?unlocked_article_code=1.cFA.CBSL.KeDGgmPu-G7F&smid=url-share https://www.nytimes.com/2026/04/15/arts/music/live-nation-an...
summa_tech 6mo ago
Fiber has fairly narrow windows in which it is as transparent as it needs to be to go long distance. We're already pretty good at filling these windows with conventional semiconductor lasers. What this is actually interesting for is being able to access arbitrary atomic transitions, many of which are outside the range of conventional semiconductors (too short, usually - there's a big hole between green and red for semiconductors). That's why they talk about quantum stuff.

SilentM68 6mo ago

[flagged]

Lerc 6mo ago

Not an expert in the field but it seems to me the key points are. Generating any wavelength. (this article) Accurately measuring wavelength. (otherwise there's no information benefit to arbitrary wavelength generation) Wavelength insensitive holographic gates. (If they work on that frequency, and in a way that does not change the frequency) I don't know what properties such devices currently have Assuming all of those, your ability to compute increases to your ability to distinguish wavelengths. You could theoretically calculate much more in a way you could never detect, but then you get into some really interesting tree falling in a forest issues.

morphle 6mo ago

The short answer if there is any "there" there for photonic computing is no, maybe. You need to understand quantum physics[3,2]. For example, photonic computing, photonic logic does not have a switch equivalent as semiconducting (CMOS transistor) or superconducting (Josephson Junction JJ) but we have a photonic Mach Zener interferometer (MZI) and a photon detector. Photonics and superconducting electronics is always going to be much larger in size (and therefore more expensive) than semiconductors build from few atoms. In quantum physics photonics we have advantages like quantum impedance, you can replace wires with photon transmitters and photodetectors and thus switch with only a few photons instead of large numbers of electrons. With photonics you can have billions of cheap low power data channels instead of high power wire bundles. But MZI as JJ will probably always be a few orders of magnitude larger than transistors so switching is not going to be better, but interferometry is. Shorter answer still: just low power communications and information processing yes, computing no. Bulk CMOS manufacturing is still cheaper than all the alternatives we have discovered or invented, until we learn to manufacture atom by atom or compute with single photons or electrons (also dependent on molecule by molecule self-assembly), we will stay with CMOS and Moore's law. Just listen to David B. Millers[1] lectures [2], his lectures are a shortcut to reading all his papers[2] that explain it all, especially [3]. Email me, I'll give you a private lecture. Your question's anwer is/was a summary of our whole lives research [4]: [1] https://appliedphysics.stanford.edu/profile/35 https://appliedphysics.stanford.edu/profile/35 [2] https://www.youtube.com/@davidmillerscience https://www.youtube.com/@davidmillerscience [3] Attojoule Optoelectronics for Low-Energy Information Processing and Communication https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7805240 https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7805240 [4] Wafer Scale Integration Free Space Optics Computing https://www.youtube.com/watch?v=vbqKClBwFwI https://www.youtube.com/watch?v=vbqKClBwFwI