Bitcoin Forum

If you don't know what AsicBoost is, in their own words: AsicBoost is a patent-pending method to lower your total cost per Bitcoin mined by approximately 20%. (http://www.asicboost.com/)
The key word here is patent. Such a technology has the potential to dominate mining but patents go against bitcoin's FOSS design. The idea to make AsicBoost's technology impossible to be used on bitcoin has already been brought up (https://www.reddit.com/r/Bitcoin/comments/4iuk6k/what_is_the_motivation_for_forking_out_asicboost/) and embraced (https://lists.linuxfoundation.org/pipermail/bitcoin-dev/2016-May/012652.html) by certain Core developers. The reason I'm calling this a dilemma is because making AsicBoost irrelevant as Peter Todd suggests, would be something inherently illiberal and against free market principles.

What are your thoughts on this?

That Core Dev Central Planner slope sure is a slippery one.

First, centrally planned blocksize production scarcity to incentivize the use of layer 2 and altcoins. Now, doing favors for the chinese mining cartel to protect them from more efficient competition. Then, deleting Satoshi's coins to keep the public "safe"...

I dont see why this should be blocked anymore than ASICs were. If it works then eventually everyone will use this process and "standard" ASICs would be obsolete. Just like FPGAs, GPUs and CPUs.

Whatever Core central planners decides is OK by me. Not like any of us have a say in this.

Yes, but currently no one is required by law to pay another party to be able to mine.

If this company is serious, it is more likely that they will grant licensing rights to purchasers.
No one to my knowledge pays license rights to ASIC companies or any other miner manufacturer.
They only pay to buy the product and there is not an additional payment for right to use over time.

If the patent is approved, then the issue becomes, in the future no one can properly
compete unless they pay for that license which grants a right to use this new technology. Only those
who have the money and resources will be able to purchase the product and licensing right.

Of course, the patent owner can release it for free use, after their patent is granted, but this is not likely
and will ultimately lead to a new tech race which will result in even more centralized mining, if not creating
the final few mining companies.

Worst case it is like a tax, if it really works, then every coin mined, they collect a small amount, say .05 coin.
Overall miner save 20%.
This is a win win situation.
I don't think asicboost will closely control patent, running their mining business, not let anybody use it etc. that will be against free market principal.

Good point. I guess I was focused more on the tech than the idea of a patent.

Yeah, I think that is the real issue why people are questioning whether to restrict this or not.
(BTW this is the first I have heard of this new tech, so I don't know if OP story is real, but assume so.)

I don't know what the correct answer or course of action would be, but I believe that mining should
stay competitive between many different groups as long as possible. Something like this could wipe
out the ability to properly protect the network, placing it totally into the hands of the few.

This new tech can be compared to when man split the atom. It can lead to innovation, but also devastation.
For the sake of the network and the future, I think (uninformed opinion) it should be a restricted use.

http://lists.linuxfoundation.org/pipermail/bitcoin-dev/2016-May/012652.html

Patents themselves (using government to give one party a monopoly at the expense of all other parties) are against free market principles.

No, that's wrong. If ever miner licensed the optimization and so was able to mine 25% cheaper, all that would happen is that the difficulty rises 25%. Everyone ends up mining the same number of coins, and the patent troll collects his 5% tax from every miner.

That's not "win win" at all. It's "win lose lose lose lose", where all the miners lose and only the patent troll wins.


Why wouldn't they? Given the change between collecting 5% of every coin that is mined and not collecting 5% of every coin that is mined, which would you pick?

Bitcoin is meant to be permissionless. Allowing a single entity to patent part of the mining process such that we have to ask their permission to continue mining efficiently is not acceptable. It's not even as if there is any novelty in the patented scheme. The "new" idea is basically "avoid unnecessarily duplicating work".

Which basically means free market died around 1474 (https://en.wikipedia.org/wiki/History_of_patent_law) :(
BTW, you having a cow gives you "a monopoly [on that cow] at the expense of all other parties," which is "against free market principles."

And this is different from bying more efficient ASIC hardware ...how?
BTW, you don't understand what a patent troll is. For starters, it is not the actual author of a work/inventor.

We don't have a free market. I assumed that much was clear.


If you were able to patent cows in general, or "milking" as a process such that everyone else had to license "your" technology from you to be allowed to own and milk a cow then you would have a point.

This is different than buying more efficient hardware because I am free to buy the hardware from any vendor I like, or create my own. There's no law telling me that if I want to use the more efficient mining algorithm I need to gain permission from a particular person.


"a patent troll is a person or company that attempts to enforce patent rights against accused infringers far beyond the patent's actual value or contribution to the prior art"

In this case the idea being patented is the that of not repeating the same calculation multiple times. It's hardly a novel idea. Requiring a licence before allowing people to omit repeating the same calculation multiple times isn't right.

Do you also feel entitled to reverse-engineer, slavishly copy, and sell the jet engine which cost me millions to design?
Should I let you copy my GDSII files without paying me? Because I wouldn't want to impinge on your natural rights to just snag whatever you want.

No one is stopping you from mining, making ASICS, optimizing your shit to make it faster/more efficient. You just can't use AsicBoost's work. Not without paying for it. Because it's not yours.


If the idea is unoriginal, begging, and self-evident, how is it that none of the ASIC manufacturers hit upon it in all these years? Is it because "repeating the same calculation multiple times" is appropriately Goldbergian for Bitcoin? So they decided to lose some money & just keep doing it all these years for the heck of it, as a whimsical baroque flourish?
Anyhow, i'm certain there are hundreds of other [equally] self-evident things you could omit, without paying a cent to anyone. Do that :)

The speed increase of 20% sounds innovative.
But patenting of mining, sounds like a strange way. Patents is part of the state, or government law. Yes, the state eyeing Bitcoin, but not all states is so loyal to Bitcoins. In my opinion, a patent on it is superfluous.

what will happen if this is only the first of many patents? will we change code everytime that happens in the future?

No more hashs = more secure.

You know, that's a good question. And in AsicBoost's case, we don't even know how they're planning to market their technology but I'm sure that harming bitcoin would be against their interest. The HK agreement that Todd cites was a roundtable (https://medium.com/@bitcoinroundtable/bitcoin-roundtable-consensus-266d475a61ff#.5k2bmosri) that several people from bitcoin ASIC manufacturing companies (Bitmain, BitFury, Spondoolies-Tech) and pools (GHash.IO, F2Pool, BW, BTCC) attended. I'm sure that those people wouldn't want patented technologies competing with them but this isn't really in the best interest of the bitcoin user. There's a lot of money at stake but forking bitcoin specifically to make a technology irrelevant mostly helps big miners.

I don't think we want the dev's to start trying to actively manage what is and is not allowed, unless there is an overwhelming consensus that something needs to be proscribed to maintain network security. (Which is not the case here.) If someone has a bright idea to improve mining, why not let them make a profit off it? The chance to make a profit provides a powerful incentive to innovate.

if everyone will save 20% then it's like none are saving anything, the same condition as before, the same profit

The AsicBoost debate is really interesting, I appreciate those folks that are taking the time to develop arguments here (whatever the position). I haven't really formulated an opinion on this issue, but this debate has jogged my memory about a patent I saw months ago while searching the Patent Office's records.  I'll have to go back through some old notes to find what I think I read originally, but its something like the patent found here: https://www.google.com/patents/WO2015138506A1.  The title is: "A system and method for detecting intrusions through real-time processing of traffic with extensive historical perspective." A word search for "Bitcoin" will take you to this paragraph:


This means an internet provider, who probably sub-licenses routers to its customers fwiw, may be able and/or required to detect bitcoin mining and possibly node activity. This is a bigger potential threat than licensing a particular bit of ASIC tech IMHO; that's not to say the same strategy being debated RE: ASICBoost isn't applicable to mining detection.  Consider what would happen if miners have to go through KYC compliance or meet some other regulation. Internet providers may be forced to act as deputies and report any detected mining or other related activity.  If that is the case, home-miners could be persuaded to give up mining; miners able to continue might just decide to relocate and/or congregate in more friendly environments. I'm not intending to create a staw-person or a slippery slope argument here, I just think there may be bigger fish to fry (or to consider frying) vis-a-vis the patent scene.  

Interesting theorization, but I doubt that this is what parties involved with mining that attended the roundtable are worried about. Would ISPs go that far to protect a patent, especially in China? They way I see this, existing ASIC manufacturers see AsicBoost as competition. Given that they wouldn't be able to use the tech freely due to the patent, they'd rather have devs make AsicBoost irrelevant other than having to submit to the competition.

I don't want to use their work. I want to use an optimization that I came up with independently, but which they are attempting to claim some kind of ownership of.

Are you telling me that my idea is not mine because you had the same idea? That makes no sense.

No, I'm telling you that while I think it's pure coincidence that you had the same idea after reading theirs, you're just shit out of luck. Just like you would be shit out of luck if you published Crime and Punishment (which you totally wrote on your own) a few years after Dostoyevsky published his.
Even though I totally believe you.

But chillax, the patent hasn't been granted yet, merely submitted (pending). If AsicBoost is as self-evident as you claim, no patent will be granted :)

Exactly.  These socialist morons think patents are evil.  Some companies spend 10s of millions to get an idea working - then the others just press one button and copy it and start selling it.  It is no wonder the US has shitloads of VC money and Europe has none.  If you are wondering why it is nearly impossible to make a startup in Europe - it is because the patent system is so weak compared to the US.

The reality is 99% of this hardware is built in China and they all copy each other's designs without regard to intellectual property.

If the speedup was 2000% and all the hardware was made in the USA where patents are strongly enforced might be relevant but a minor 20% optimization is nothing we should be hard forking to avoid. Irrelevant - many design advances in ASIC, widely copied, have yielded far more than 20% boosts and Bitcoin has managed to survive.

So even though you believe that we came up with the idea independently of each other, one person should be allowed to use the idea and the other shouldn't?

That really just makes no sense.


You appear to be confusing patents and copyright. They're quite different things. You can't independently write the same novel. If I tried to publish a copy of an existing work then clearly I have copied it.


You have more faith in the patent system than I do.

I could tell you that I was being sarcastic, that coming up with their idea after reading it is called "retention" and not "invention." I won't, I'll just say "they found it first, and got dibs on it." Just like they would have dibs on the dollar bill they found on the sidewalk, which you would have found if they didn't. It's theirs. To do with as they wish.

Oh stop it. Of course I'm not confusing patents and copyright, I'm using whatcha m'call an analogy. Crime and Punishment is such an obvious novel that anyone could have written it. Especially after reading it. Much like you read the AsicBoost pdf.

It's imperfect, it's exploitable, but it's far better than nothing. Possibly due to the fact that inventors aren't a particularly powerful lobby, so the laws protecting their work are unlikely to be stacked in their favor.
As others have pointed out, this is all pointless rhetoric, since the Chinese (who'll most likely be making your next batch of silicon) don't particularly care about patent law & could release their next gen chips through some Belize corporation, perhaps with some catchy/original name, like AsicMiner.
So even if AsicBoost kids have the wherewithal to decap and conclusively prove, at gate level, that their patented methods are being used, good luck to them finding an entity to sue.

May seem like a double standard but the fact he is creating a aspect that could change bitcoin and at the same time limit who has access. So attempting to create a monoply, that is against bitcoin in my opinion and it strips away the mask of double standard.

Unless I'm missing something (possible, I only skimmed through their pdf), we're talking about ~20% boost? Better designed silicon from the same node process could offer more (and did in the past, and how!).
'Far as limiting access, a huge chunk of Bitcoin hashpower was manufactured by a guy known only as Friedcat. How would one mount a case against that?
Plus straight from the horse (http://www.uspto.gov/patents-getting-started/international-protection/protecting-intellectual-property-rights-ipr)):
Besides, patent hasn't even been granted. Might never be, if dooglus is on the money.

Well, it's possible that the patent never materializes in reality. However things tend to happen very fast in the Bitcoin universe. Therefore it seems appropriate to be prepared for the worst. When it comes to financial software, it's always good to have slightly paranoid people like Peter Todd on board. Better be safe than sorry.

Clearly this is not about being against innovation. The problem is, that this kind of innovation might not be equally available to all miners, which poses the risk of a malevolent mining cartel.

ya.ya.yo!

Of course it's not gonna be equally available to all miners, nothing ever is. Not government subsidized 2-4 cent power, just for starters.
As far as "malevolent mining cartels" go, scratch the malevolent part (Bitcoin has no anti-monopoly rules, afaik), and sure, we got them already :(

http://s32.postimg.org/ru8bnc6j9/bitchrs.jpg

This AsicBoost debate is very important for the future, I'm not entirely sold on the idea to hard-fork Bitcoin just to eradicate someone's innovation. If we allow this fork, then how many will be needed if someone else decides to file a new patent because they designed a chip that gives 1000% more hashing power?
Are we really going to prevent intellectual property?

True. But can they export patent-infringing hardware to the West?


20% isn't minor. It will be the difference between being able to mine profitably or not.

There is a physical limit to how fast hardware can mine. Hardware which is allowed to skip repeating part of the hashing process will be able to mine faster and cheaper than hardware that cannot. Eventually the faster hardware will be profitable and the slower hardware will be uneconomical to run.

Bitcoin was designed to allow anyone to be able to mine it. A mining environment where one entity is able to mine 20% faster than all the others because they are using government power to cripple the competition doesn't work out well for Bitcoin as a whole.

I think the problem is more with the patent. It would effectively create a monopoly on money, if only one company can dominate the manufacturing and supply, because they are protected by the patent. A global patent would stop other competitors from entering the mining scene.

It's supposed to be a free market. ^hmmm^

By "government power," you mean subsidized Chinese 2 to 4 cent per kWh power, right? That's more than a 20% edge on my rate tho. Life is so unfair :(

We already have a centralized mining horizon (80% of mining is in China). I don't see how we are in a free market right now.
This dilemma is a true one and creates increasing troubles to the mining industry approaching the halving.
For the first time I am a bit worried.

Don't know how accurate are your stats ,even if its true ,I don't see see any relevance in it being called "centralized".Mining isn't profitable anymore,with electricity costs and everything ,difficult for an individual to make profit out of it especially in the western countries where prices are sky rocket high.China has mining farms on a large scale,early adopters.The mining farms are too expensive to be shut down considering the amounts invested in them.

Again,the lesser the miners it will be more lucrative for them with the low block rewards.

You are, you are free to create your own mining equipment.
Or having to ever pay for anything ever. MS should make Windows freely available to everyone, so as not to put Linux people at a disadvantage. Everyone is entitled to an iPhone, not just any smartphone. And rich parents, not just any parents, because otherwise life would be totally unfair.
You are. AsicBoost isn't taking away your CPU, your GPU, your FPGA, or the chickencoopful of ASICs you're mining with now.

What can I say? Do the math and check for yourself here
https://blockchain.info/pools?timespan=24hrs (https://blockchain.info/pools?timespan=24hrs)

You are a bit superficial regarding the matter: I think it's very important knowing who controls the mining operation.

It won't necessarily translate to a few people/companies are the only ones able to mine. Once the patent got approved, should they decided to sell the rights to use the tech to ASIC companies, these companies will simply pass the costs to their end-users, us miners and the huge mining organizations. In the end everyone gets to upgrade, and the situation would stay the same as it is now.

In the end if everyone pays money to acquire that 'improvement', the it is pointless as nobody gains the advantage anyways (everyone's hashrate increases by 20%). In other words, this is pointless as it gets.

The solution is simple and was proposed by some developer, i.e. block such improvements with a change that is required to make them inefficient.

Exactly. If they don't 'sell' their patent to others, they're going to be abusing their position and pretty much undermine all connection. On the other hand, if they sell it nobody has an advantage and they generate profit. I disagree with either case.

You dont count the fact the 20% faster units should have higher initial cost price to pay the patent rights. How much more the initial cost should be? To get most revenue for the patenter it should be very close to the +20% expected lifetime profit of the unit. In other words, in ideal world the expected lifetime profit for normal and +20% faster patented unit should be the same eventually.

Err...You mean like everyone upgrading from CPUs to GPUs to FPGAs to ASICs to better_ASICs? That sort of pointlessness?

Reading up on this, I'm struck by a couple things:

1. Bitcoin core is threatening to make changes to defeat this, and any other, patents that would effectively force miners to pay the patentee in order to mine profitably.

2. Patents drive innovation, providing a means for inventors to profit from their work.

3. The ASICBoost inventors point out that if Bitcoin Core suppresses their patent, the impact will be this: Future innovators with similar technology will be motivated to go underground, developing secret partnerships with (select) miners. That way they still get paid something for their work, but only a fraction of the mining community gains the benefit.

Point #3 bears careful consideration. Let's say the ASICBoost team had realized what Core's response would be, and just secretly sold their technology to the highest bidder after quiet negotiations. That bidder would hold a major advantage, strong enough to grind down competition over time.

In the long run, this could lead to some serious centralization of the mining industry, and none of us want that.

The point being, Core acting to suppress patents does not make the inventors quit or their inventions magically disappear. It winds up endangering the decentralization of the bitcoin mining network.

I'm not a huge fan of the idea that the ASICBoost team, and similar groups, could reap huge profits off the miners. As someone noted above, the rational ASICBoost price would be just a little less than the added profit it provides compared to an equal miner that does not use ASICBoost. In other words, just under 20% of mining revenue. So ASICBoost could potentially suck up almost 20% of mining revenue. The Miners would have to pay electricity, hardware investment, personnel, rent and so on from the remaining 80%. That's a tough hit for them.

But if the alternative is them going out of business... I guess this brings forth Core's threat to defeat not just ASICBoost, but to find a way to defeat any similar enhancement. Keep in mind that Core would not only need to defeat public patents, but also any private invention that someone secretly cooks up.

If they could do that with some confidence, I think that might be the way to go. But if they can't be confident that they are forcing a level playing field on all miners, then I think they'd better lay off and let the patent holders take their profit. Because otherwise I think the threat of centralization due to some miners having a fixed advantage is just too great a risk.

Just my 2 satoshis...

Great points.  This whole issue is pretty complicated IMHO.  There are all kinds of possible consequences associated with changing the protocol.  I think it would be good to see some kind of research paper/memo from proponents - ideally some of the core devs - researching the prospective change using a cost-benefit analysis of some kind. Specifically, What are the most probable risks associated with doing nothing (is there any empirical evidence of these types of risks materializing in other fields, what was the outcome?), how have analogous issues in other industries been handled, what are the most probable risks of forking the protocol on the basis of this or similar threats, what are the known-unknowns here (and what are some analogous unintended consequences associated with similar changes and non-changes in other fields, how well do these analogies fit here), etc.?  Same type of paper from opponents would also be useful.     

TL;DR - A change in protocol may well be warranted.  However, making the change without a concerted and transparent attempt at analyzing the problem alongside the users risks a lot of backlash and an irreparable loss of faith in the Bitcoin system. Proponents implementing the change cannot and should not let themselves appear to be above public dialogue (i.e., by ignoring genuine concerns of opponents or concerned users) or like they're making things up as they go along.  Paradoxically, making the right change may be wrong if collapses the system.

No, of course not. We are talking about patents, not the price of power.

If you're choosing to mine in a place where you're paying so much for your power that you can't compete, that's your choice. It is also a separate issue from whether Bitcoin mining should be patent-encumbered.

I would support this if they hard forked to ban all asics. Because why asics should have advantage over a GPU? But then, GPUs have advantage over CPUs, thus Core should also make changes to forbid GPU mining. \s


 

Well, does it work?

If it does, for how long?

I think that ship sailed a couple years ago. In hindsight, maybe they should have acted to prevent ASIC mining and kept things at the CPU/GPU level. That would have maintained greater decentralization, I would think. But even with GPU mining we face these kinds of situations. When I was GPU mining X11 I recall how someone came up with a pretty good refinement that boosted hashrate. They sold it to one miner, but it ended up leaking to everyone pretty quickly, robbing them of any further profit.

A software-based innovation would probably be worthless to patent (piracy would be rampant), but one could easily envision inventors encoding their innovations in GPU or motherboard hardware and creating the same scenario: Only those who buy the patentees' hardware could survive.

But killing ASIC mining now would be virtually impossible. For one, it would render hundreds of millions of dollars worth of mining hardware worthless - so 100% of the existing bitcoin mining industry would be dead set against it.

No, that is not comparable.

That doesn't make sense. You don't really gain anything with that. People with capital are still going to flush out small miners within a very short amount of time.

I don't think anyone knows now, it's a patent.

Well, until changes are made that render it useless.

You missed the point.

The point isn't that the new technique is faster, it is that the new technique is patented.

If it was only faster then everyone else could also go faster to keep up.

But it is both faster and patented, so even if everyone else comes up with the same idea, they still aren't allowed to compete because of the patent.

If you're choosing to mine in a jurisdiction in which said patent is registered/effective that you can't compete, that's your choice. I hear Somalia don't give shit about no patents.

Subsidised power price and patents are not that much different. Former incentives centralisation of hash-power in one state (right now Chinese gov can seize control over miners/pools quite effortlessly, if they only choose to do so), the latter could give control to one commercial entity (depending how are they going to play it).

I'm not saying AsicBoost should be allowed, but there's some serious inconsistency here, and it's not even the first patent relating to ASIC mining, BitFury had one (specialised, more efficient chip) but somehow no one have seen any major problem with it back then.

PoS is starting to look less ugly...