An Idea To Increase the Efficiency of Bitcoin Mining by an Order of Magnitude

[xcrunner2414]

Meta—It was suggested that this post be forwarded to this bitcointalk forum. It was originally posted on Reddit here: https://www.reddit.com/r/BitcoinMining/comments/sab5yk/an_idea_that_could_increase_the_efficiency_of/?utm_source=share&utm_medium=web2x&context=3. Obviously this is not a technique that could be developed by anybody except the most highly specialized scientists and engineers, but if the idea has merit then I think development should begin as soon as possible. So, if you know anybody in the R&D department of a large Bitcoin mining corporation, please forward it to him/her.

Tl;dr—Instead of using a local energy source (local work) to accelerate and manipulate charged particles (electrons, which are not massless) in intelligently designed devices to perform logical operations (compute proof), why not use intrinsically energetic and intrinsically massless particles (photons) that originate from the largest source of energy (work) in the solar system (the sun) as a source of coherence (for logical operations), and incoherence (if its needed for random guessing), to compute the proof of the next block? All that is needed is a novel, intelligent design.

There is already a large amount of research being performed in the development of Application Specific Photonic Integrated Circuits (ASPICs) and Programmable Integrated Photonics (PIPs). https://en.wikipedia.org/wiki/Optical_computing. To my knowledge, these devices can be used to perform basic logical operations, and I don’t know of any reason why they couldn’t be designed to perform the SHA256 hash function.

Due to the fact that these devices are intended to be deployed in locations where they are needed, ASPICs are currently being developed to use locally-powered lasers as a source of coherence. However, it is well-known that Bitcoin mining is location agnostic. As it turns out, lasers can be produced from solar radiation: https://en.wikipedia.org/wiki/Solar-pumped_laser.

Of course, Bitcoin’s Proof of Work algorithm includes randomly guessing a nonce. As it turns out, random number generation can be performed using either a coherent or incoherent source. https://arxiv.org/pdf/1611.00224.pdf, https://arxiv.org/abs/2108.08743.

Therefore, is it, or is it not, conceivable that the work that is required to prove the sequential ordering of Bitcoin transactions can be performed by the sun? And, is it, or is it not, also conceivable that the proof of that order can be computed by the radiance of the sun?

Therefore, isn't it conceivable that the sole source of all new bitcoins could, someday, be the sun? Doesn't this also solve the oft-cited problem of Bitcoin's security post-2140?

Edit: See this video, for example, of a professor from the Netherlands giving a talk about optical computing. https://youtu.be/UqeH7ozVOpQ. The video is more than an hour long, but the summary is that optical computers have been following Moore’s Law during the past couple of decades the same way electronics followed Moore’s Law in the latter half of the 1900s. They are now making optical computing chips as neural nets in AI, and these chips are much more efficient than their traditional electronic counterparts.

Edit 2: One of the companies that is mentioned in this professor’s talk is LightElligence. https://www.lightelligence.ai/technology. “By processing information with light, our chips offer ultra high speed, low latency, and low power consumption representing orders of magnitude improvement over traditional electronic architectures.” This particular company specializes in optical chips designed for AI applications. If Bitcoin continues to grow, I expect there will become a greater incentive for companies to develop similar ASPICs designed for SHA256 hash.

[1713865195]

1713865195

[1713865195]

1713865195

[1713865195]

1713865195

[larry_vw_1955]

Therefore, is it, or is it not, conceivable that the work that is required to prove the sequential ordering of Bitcoin transactions can be performed by the sun? And, is it, or is it not, also conceivable that the proof of that order can be computed by the radiance of the sun?

Solar power is already being used for bitcoin mining. So yeah...anything like that sounds like it could be possible. needs to make financial sense for the miner tho.

[n0nce]

Therefore, is it, or is it not, conceivable that the work that is required to prove the sequential ordering of Bitcoin transactions can be performed by the sun? And, is it, or is it not, also conceivable that the proof of that order can be computed by the radiance of the sun?

Interesting idea, on high level (not sure if the science behind it is sound), but you forget that it won't be completely 'performed by the sun'. You will still have devices converting all this energy, radiation etc. into valid blocks. These would be the new ASICs that will always get faster (better use of sun energy / efficiency), smaller etc. and then I'm not sure what issue will actually be fixed. Because there will still be electric waste, race to produce more and better miners and so on.

If it's just about reducing power draw from the grid (that's not more efficient - just a different way to retrieve electric energy), that can be done today through solar cells.

Also, fwiw the nonce's randomness isn't really so important that it makes sense to use photons or similar, you could even just count up from 0.

[xcrunner2414]

Solar power is already being used for bitcoin mining. So yeah...anything like that sounds like it could be possible. needs to make financial sense for the miner tho.

Yes, indeed, solar power is already being used, but not in the way that is being suggested here. Currently, the energy of the photons from the sun is being absorbed by the PV panels and converted into electricity, which is then used to provide power to traditional electronic computers which perform computation by accelerating electrons within the device's transistors. What's suggested here is that all those photons from the sun be used directly within the optical logic gates of a novel device that is specifically designed for this purpose. It is apparent that solar radiation can be used to create coherent light (laser). From wikipedia:

A solar-pumped laser (or solar-powered laser) is a laser that shares the same optical properties as conventional lasers such as emitting a beam consisting of coherent electromagnetic radiation which can reach high power, but which uses solar radiation for pumping the lasing medium. This type of laser is unique from other types in that it does not require any artificial energy source.

Solar-pumped lasers are not used commercially because the low cost of electricity in most locations means that other more efficient types of lasers that run on electrical power can be more economically used. Solar pumped lasers might become useful in off-grid locations.

Additionally,

Optical computing or photonic computing uses photons produced by lasers or diodes for computation.

So, if computation can be performed by a source of coherent light, and coherent light can be produced by solar radiation without an artificial energy source... then doesn't that imply that it's theoretically possible to perform computation with zero operational cost? Obviously it hasn't been economical to mine Bitcoin this way because there is no such optical computing device in existence that is specifically designed to perform SHA256 hash function, let alone one that can perform the hash function from a single source of coherent light. But... it seems as if photonic integrated circuits are now starting to become economical for certain niche applications.

I'm not suggesting that this actionable at the moment. But, perhaps in 20 years, maybe something like this could be the pursuit of a joint venture between some government agency, polytechnic universities, and a large Bitcoin mining corporation.

[xcrunner2414]

Interesting idea, on high level (not sure if the science behind it is sound), but you forget that it won't be completely 'performed by the sun'. You will still have devices converting all this energy, radiation etc. into valid blocks.

Yes, the devices would need to be specially designed. But, the theory presented here is that the energy required to perform the computation would come directly from the sun without first converting it into electricity. The idea is that the photons that come directly from the sun would be the exact same photons that are used in the logic gates of this novel device, with no conversion to electricity.

These would be the new ASICs that will always get faster (better use of sun energy / efficiency), smaller etc. and then I'm not sure what issue will actually be fixed. Because there will still be electric waste, race to produce more and better miners and so on.

If it's just about reducing power draw from the grid (that's not more efficient - just a different way to retrieve electric energy), that can be done today through solar cells.

Technically, the idea suggested here is not a computer that is retrieving electric energy from the grid in a different way, but rather a computer that is powered primarily and directly by a naturally occurring fusion reaction in space.

[n0nce]

~

I see, I see. So your logic gates will operate on light rays instead of wires with electricity going through them. Sounds plausible. You will still need a small traditional computer that constructs blocks and inputs them into the ASIC (or ASPIC). So an interface between 'traditional electronic computer' and 'ASPIC that needs photonic inputs'. After the inputs were set, the circuit will need to double-hash, check difficulty target, choose a new nonce and repeat.

[xcrunner2414]

I see, I see. So your logic gates will operate on light rays instead of wires with electricity going through them. Sounds plausible. You will still need a small traditional computer that constructs blocks and inputs them into the ASIC (or ASPIC). So an interface between 'traditional electronic computer' and 'ASPIC that needs photonic inputs'. After the inputs were set, the circuit will need to double-hash, check difficulty target, choose a new nonce and repeat.

Yes. Like any worthwhile endeavor, designing such a device would likely be extremely difficult.

What do you think about this as a potential solution to the apparent lack of incentive to continue mining after the block subsidy is no longer issued? This is not a topic that I've spent a considerable amount of time thinking about, so please forgive me if I'm uninformed or simply being naive, but if there ever came a time when the Lightning Network was robust and stable, and when new channels were not created often, then the transaction rewards to the miners might not be sufficient to cover the cost of such a large power consumption (even if it is still only < 5% of the global power consumption). On the other hand, if all of Bitcoin's power consumption came directly from the sun's nuclear fusion reaction, without conversion to useable electricity, and if the optical computing devices were stable and durable, then nobody could possibly complain that Bitcoin was taking too much energy from the grid and maintaining the network may be as simple as infrequent maintenance of these durable devices.

[larry_vw_1955]

Yes. Like any worthwhile endeavor, designing such a device would likely be extremely difficult.

And just because it uses renewable energy doesn't mean it is something miners would do unless it made financial sense.

What do you think about this as a potential solution to the apparent lack of incentive to continue mining after the block subsidy is no longer issued?

More likely is that transaction fees will  make up for the block subsidy.

[xcrunner2414]

And just because it uses renewable energy doesn't mean it is something miners would do unless it made financial sense.

Well, technically, the renewable energy source suggested in this idea is like no other. It's more akin to how plants use sunlight to grow; not many people would consider plants as a user of renewable energy (at least, not energy that could sensibly be made available to the grid). But, since you bring up the financial aspect, I think large miners may someday want to collaborate with a government agency to fund a $10 million R&D project if it could potentially save them an order of magnitude, or more, in efficiency, especially if their operational cost is an order of magnitude greater than their contribution to the project's funding. If Bitcoin becomes as accepted and ubiquitous as the internet is today, then what government wouldn't jump at the opportunity to make their mining operation more environmentally friendly than that of every other nation?

More likely is that transaction fees will  make up for the block subsidy.

Personally, I think it's too early to predict whether or not the fees will cover the costs of the miners; 2140 is quite far into the future.

[larry_vw_1955]

Well, technically, the renewable energy source suggested in this idea is like no other. It's more akin to how plants use sunlight to grow.

Photosynthesis is the process by which plants use sunlight, water, and carbon dioxide to create oxygen and energy in the form of sugar.

I dont see a similarity. But if that technology existed then
fine we used to have asics now we have something better
Miners still need to make a profit Bitcoin goes on what else can you use the technology for?

More likely is that transaction fees will  make up for the block subsidy.

Personally, I think it's too early to predict whether or not the fees will cover the costs of the miners; 2140 is quite far into the future.

 Think of it like the trucking industry whenever gas prices go higher or new regulations get put in place to make things more expensive for the trucking industry guess who gets the price passed on to them?

[xcrunner2414]

Photosynthesis is the process by which plants use sunlight, water, and carbon dioxide to create oxygen and energy in the form of sugar.

I dont see a similarity. But if that technology existed then
fine we used to have asics now we have something better
Miners still need to make a profit Bitcoin goes on what else can you use the technology for?

The similarity to which I was referring was simply that this proposed mining technique, like plants, would use the power of the sun without relying on electricity; the power of the sun would go directly into the proof-of-work computation without first converting the photons' energy into electrical power. Similarly, plants use sunlight directly in the process of photosynthesis to make energy within their cells, no external electricity required.

[larry_vw_1955]

Photosynthesis is the process by which plants use sunlight, water, and carbon dioxide to create oxygen and energy in the form of sugar.

I dont see a similarity. But if that technology existed then
fine we used to have asics now we have something better
Miners still need to make a profit Bitcoin goes on what else can you use the technology for?

The similarity to which I was referring was simply that this proposed mining technique, like plants, would use the power of the sun without relying on electricity; the power of the sun would go directly into the proof-of-work computation without first converting the photons' energy into electrical power. Similarly, plants use sunlight directly in the process of photosynthesis to make energy within their cells, no external electricity required.

so what else would the technology be used for? is bitcoin it's most compelling application?

[xcrunner2414]

so what else would the technology be used for? is bitcoin it's most compelling application?

Well, I think that depends on the technology to which you're referring. There's already commercially viable photonic integrated circuits (PICs) being employed in niche areas of the market. As noted in the OP, a company called LightElligence is making PICs for AI applications.

If you're referring to the entire technique, though--the technique of using sunlight directly for computation--then I think that would have very limited applications, mostly because the availability of sunlight is intermittent. But, I think the efficiency that it could bring to the Bitcoin mining market is reason enough to motivate its development, especially if Bitcoin continues to become a more important and prevalent financial technology.

[NotATether]

So, if computation can be performed by a source of coherent light, and coherent light can be produced by solar radiation without an artificial energy source... then doesn't that imply that it's theoretically possible to perform computation with zero operational cost? Obviously it hasn't been economical to mine Bitcoin this way because there is no such optical computing device in existence that is specifically designed to perform SHA256 hash function, let alone one that can perform the hash function from a single source of coherent light. But... it seems as if photonic integrated circuits are now starting to become economical for certain niche applications.

Not exactly zero operational costs - you still have the data center maintenance costs - but this manages to eliminate the need for large power sources, and with it, the main talking point that people use against bitcoin mining (that "crypto mining causes pollution" - but that is a discussion for yesterday). It can be deployed in places like Africa with almost zero disruption to the power grid. You'd still need a backup generator for the systems running the Bitcoin nodes submitting the blocks, and a decent internet line, but yeah. It looks doable.

[j2002ba2]

Optical computing is extremely slow and clunky. Optical stuff is great for communication though.

It's easy to see why it wouldn't work: the photon just takes too much space - with wavelength hundreds of nanometers it just cannot compete with the density of present day chips. Additionally photons don't interact themselves, so at least one electron is needed for anything to happen. Or to happen probably, maybe. It's way worse: the event probably doesn't happen. The computation is lost then. Which means huge switching devices, and relatively slow speed.

What about the power needed? One must use lasers for all this. Very inefficient.

Edit: typo.

[BlackHatCoiner]

I don't understand much from optical computing, but here's a simple question: What's bad with converting the energy of the photons to electricity? Does it make the procedure less efficient?

[xcrunner2414]

I don't understand much from optical computing, but here's a simple question: What's bad with converting the energy of the photons to electricity? Does it make the procedure less efficient?

Whether or not it's bad depends on your moral philosophy, so can't determine that for you. Regarding the question of efficiency, the short answer is "yes, it's less efficient to convert photons into electric power and then use that power to accelerate subatomic charged particles, with mass, inside electrically conductive material to perform computation (which inevitably produces a lot of heat because the particles are charged) than it is to simply use the intrinsically energetic and intrinsically massless photons to perform computation directly." Of course, this is all moot because humanity has not developed this technique of using sunlight to perform computation. This answer also depends on the design and engineering of the technique; a bad design can turn a more efficient physical process into a less efficient human endeavor.

[larry_vw_1955]

so what else would the technology be used for? is bitcoin it's most compelling application?

If you're referring to the entire technique, though--the technique of using sunlight directly for computation--then I think that would have very limited applications, mostly because the availability of sunlight is intermittent.

i'd say i couldnt agree with you on "very limited applications". such a technology seems like it would be useful in general purpose computing. if it can compute hashes i'm sure it could run a c program.

[BlackHatCoiner]

Whether or not it's bad depends on your moral philosophy, so can't determine that for you.

My moral philosophy? What makes direct usage of the photons' energy more or less moral than the other way?

[BlackHatCoiner]

You and OP had different idea about "bad". He misunderstood that you refer "bad" in morality area, while you actually meant "bad" in technical area.

Yes, but I don't understand how's morality related with the discussion at all; otherwise it wouldn't be tiring from me to just add an adverb and avoid confusion.