HNHacker News
TopNewBestAskShowJobs

betterunix2

3,500 karma · joined September 18, 2017

submissionscomments
betterunix2··on On Web3 Infrastructure
That is an impressive list of cryptographers working on blockchains, but at major cryptography conferences there is less and less blockchain work being presented, to the point where CRYPTO'21 didn't have any blockchain sessions at all, while EUROCRYPT'21 had a single session where blockchain work was combined with work on privacy and law enforcement. To be fair, three sessions at CCS'21 were dedicated to blockchain research, but CCS is structured to allow more topics, it is not a conference specific to cryptography, and they had two sessions dedicated to MPC and a third on federated learning which touched on MPC. It is a small sample but representative of a larger trend of cryptographers becoming less interested in blockchain research.

I have not seen ZKP innovation outstrip MPC innovations at all. In the past decade I have seen a rapid expansion of research in MPC following both a strong push by DARPA and growing interest among large tech companies and banks. There has been a revival of interest in set-intersection protocols and related functionalities, a lot of impressive work in garbled circuits and other generic protocols that have greatly reduced their resource requirements, machine learning applications, and various other ongoing lines of work. At worst I would say that ZKP and MPC research have been roughly equal in terms of the pace of innovation, which should surprise no one as the two topics have strong connections.

Moreover, while there is certainly a lot of ZK research being published year after year, most of it has nothing to do with blockchains and is not coming from anything related to blockchains. There are plenty of academic researchers publishing ZK work, and I still see lots of industry ZK research that has nothing to do with blockchain. The same is true of all the other topics you mentioned -- some blockchain-inspired work here and there, but a lot more research from elsewhere.

Sorry to hear that your MPC work has not made it into production, but maybe that is because it is not as practical as you claim. Personally I like to say that the only test of "practicality" that matters is whether or not it is useful in a real-world application. Obviously your SNARK work cleared that bar, which is great but does not really say much about the pace of innovation. I can say that most of my published research at this point has been put into production -- an equally meaningless statement since I have been working for a big tech company for a long time, and the research I have published in that time has all been the result of work I did to address various privacy and security problems that company faces. My judgement of where the innovation is happening is based on the research I am seeing people present at various conferences. Maybe I am looking in the wrong places, and there is actually a whole world of cryptography conferences where people are excited about blockchain work?

betterunix2··on On Web3 Infrastructure
My job would be much simpler if I could deploy generic MPC; all I would have to do is maintain a library and maybe a compiler, without having to design and implement an entirely new protocol every time someone came along with some new feature or use-case. The engineering effort of productionizing a generic protocol is a one-time cost and my coworkers and I could do that work relatively quickly. On the other hand, special-purpose protocols are typically difficult to modify, we are have to do our security proofs over and over whenever we roll out anything new, and we must go through the engineering effort over and over.

Engineering effort is not what holds back the deployment of generic MPC protocols. Those protocols are just too expensive to run in the majority of real-world MPC applications. Even special-purpose protocols are sometimes too resource-intensive to be deployed. I do not see that situation changing without a radically different approach to generic protocols. I also do not understand what is so uniquely exciting about deploying a generic ZKP or MPC protocol. If it works in a giving setting and no special-purpose construction could be used, great, but it is not some kind of badge of honor.

As for Zerocash, you had originally said that blockchains are where we can finding the largest deployments of non-trivial ZKPs, which is why I pointed out that Zerocash and its followup work do not really involve "blockchains" beyond a particular instantiation of a ledger. If the construction can be implemented without any blockchain at all -- which the authors of the original paper took the time to point out -- then I do not see how any of the research on ZKPs motivated by Zerocash and its followup work supports your claim at all. You are saying that Zerocash is actually ecash, which kind of makes my point for me: we are not actually talking about something in the "blockchain space."

(Also, I have a somewhat controversial view that NIZKs and signatures are not actually "zero knowledge," since the verifier obviously cannot compute a NIZK or signature without receiving a message from the prover/signer and thus gains knowledge when it does receive those strings. Not that it matters in any way for this conversation, since the value of innovation in NIZKs or ZKSNARKs is not in doubt, but I did want to mention that signatures are a poor example of real-world deployments of ZKPs.)

betterunix2··on Google doesn't want to pay Sonos for technology it copied
It is not that simple. If you are willing to shell out more money for "professional" products, you can get products that standalone or you can buy a standalone system (e.g. security cameras with an NVR etc.). What we are getting is a two-tiered society where people with more money have the option to "completely" own their appliances/electronics/homes, while the majority of people cannot afford to do so.
betterunix2··on Google doesn't want to pay Sonos for technology it copied
It is not true that anyone knowledgeable in a field finds everything in that field straightforward. For example, Schoof's algorithm for counting points on elliptic curves was so non-obvious to relevant experts that he struggled to get his paper through peer review. Diffie also struggled to get his ideas about public-key cryptography published because experts in relevant fields did not understand how such an idea could even make sense. Another example is Gentry's original FHE construction, which was not at all obvious even to experts in lattice theory or any other relevant field of cryptography or math. It pains me to say this, but Bitcoin is also an example; check out the response on the cryptography mailing list, where several prominent experts in the field were confused by the concept of electronic payments that do not require any bank to issue and redeem the money.

Those are just what I know off the top of my head from my own field. While there is plenty of incremental research in any field and plenty of situations where a motivated expert would have arrived at the same basic concept, it is not outlandishly uncommon for a truly novel, non-obvious idea to be presented. The problem for patent examiners is that they are not experts and the pace of software innovation leaves them baffled by the applications they are examining; there are also too few patent examiners to handle the volume of applications that are submitted.

One way to address the problem is to just abolish software patents entirely. Software was never meant to be patentable, at least not if you recognize software as a form of applied math (happy to argue this one all day long) and accept the idea that math is not (or should not be) patentable.

betterunix2··on On Web3 Infrastructure
Some interesting research is being done by some cryptocurrency companies like ZCash and Algorand. Their work on ZKPs and SNARKs has been interesting, but it is worth pointing out that they are not the only people working on this. Moreover, the serious cryptographers working on anything related to blockchains have more or less stopped talking about the permissionless setting (where Bitcoin, Ethereum, and basically all of the popular blockchains in use are) because security is too hard to define in a meaningful way. In the "permissioned" setting where parties have well-known identities there has been a bit of interesting research on maintaining a shared cryptographic data structure.

Meanwhile, academic and (non-blockchain) industry researchers have been pushing the state of the art in every subfield within cryptography, ZKPs included. Big companies have been deploying MPC as a means of addressing privacy concerns and regulation, and the cryptographers working on that (full disclosure: I am one of them) have been pretty active in publishing their work. Academic researchers have further advanced the results and addressed the problems raised by industry researchers, sometimes breathing new life into almost-forgotten lines of research (like set intersection protocols).

So sure, I can grant you that there has been some interesting work on cryptography within the blockchain space, but it is not nearly as exciting and significant as you suggest. I actually have a lot of respect for the ZCash team, whose work really is top-notch and who I see (or saw pre-COVID) at high-quality conferences like CRYPTO and RWC. On the other hand they are a small and very unique team within both the blockchain ecosystem and the cryptography research community, and their research work is only nominally related to blockchains (it is inspired by an application that did not even require a blockchain in the first place). Beyond the ZCash and a few other groups with serious cryptographers the blockchain space is a desert in terms of interesting cryptography.

betterunix2··on On Web3 Infrastructure
MPC has been deployed at large scale by numerous companies, at least for the ads industry; I know because I actually work on exactly this full time and I have seen the numbers (but unfortunately I cannot share specifics). There is nothing special about general-purpose protocols that makes them more "legitimate" or whatever; we use specialized protocols in practice because it is almost always more efficient and thus less expensive to run (and MPC is usually right on the threshold of being too expensive).

As for zerocash, the last time I looked into it what I saw were a set of security definitions that assume a reliable ledger of some kind; whether or not that ledger is implemented using a blockchain at all is not addressed in the theoretical work. The practical deployment relied on Bitcoin, but since Bitcoin security is not well-defined (or at least not convincingly defined) that makes the rest of the security argument dubious. As far as I know Zexe has the same problem: yes, the security definition is much stronger, but they do not address the realization of the ledger functionality itself and thus any real-world deployment that relies on e.g. Bitcoin, or really any permissionless blockchain, has the same theoretical shortcomings. Ultimately the permissionless setting itself is the problem; zerocoin could be implemented using a ledger managed by a trusted party, and it would achieve its security goals without those theoretical problems.

I should also be clear that when I say ecash does not share this problem, it is because ecash has a well-defined security model and all functionalities needed to realize an ecash system also have well-defined security. We can instantiate ecash using any of the security assumptions we commonly use for digital signatures, and in theory ecash can be instantiated from MPC (by using a generic MPC protocol to implement a blind signature, then using the blind signature to implement ecash), which itself can be instantiated with standard cryptographic assumptions. So ecash has a security definition that is as well-defined as a cryptographic security definition can be.

betterunix2··on Forced to sell chipless ink cartridges, Canon tells customers how to bypass DRM
Paperless tech is not there. If it was there it would not be so inconvenient to do simple things like underlining some text or crossing out a word. I have tried all the above and it is still too cumbersome to really be comfortable. It's not just me; despite the widespread availability of these devices and the technology generally, people have not really made the switch.

When cars first came out their technical advantages over draft animals were so obvious and undeniable that the world rapidly switched from animals to machines and has never looked back. We have had tablets with styluses and other paper-replacing technology for decades now, but thus far the advantages compared to just using paper have not clearly outweighed the disadvantages and the world has not rapidly switched. To put that in perspective, I have not had a desk phone for more than a decade now, because cell phones are clearly a better thing to give employees; at the same time my desk is within 100ft of a heavily-used printer/scanner/copier.

betterunix2··on On Web3 Infrastructure
Except that ZKPs had already seen real-world use before Satoshi's whitepaper was circulated; in fact, there was an already-defunct startup that was selling ZKP-driven authentication tech. Secure multiparty computation is even more advanced than ZKPs, was already deployed in several real-world applications prior to Bitcoin, and has probably driven more research on ZKPs (as a building block in MPC protocols) than anything in the blockchain space thus far. As for how widely productionized the technology is, while I am not sure how you define "non-trivial" ZKPs, U2F was almost certainly a more widely used ZKP application than any blockchain tech, and there are plenty more real-world ZKP applications having nothing to do with blockchains that we could list.

David Chaum dreamed about a world where electronic payments could be anonymous and secure, but the demand was not there and his startup never took off. "Blockchain" sucked most of the oxygen out of the room when it comes to further work on ecash, which is unfortunate given that even the most technically complex ecash proposals were overwhelmingly more efficient than any blockchain-based payment tech ever could be. For what it's worth, the most recent ecash proposals also advanced the research on NIZKs and ZKPs more generally (it is actually hard to avoid some kind of NIZK in a system that supports offline payments) and had ecash been deployed more widely we probably would have seen at least as much research and productionization activity as we see in the blockchain space.

On the other hand, blockchain research has struggled with a foundational question that does not present a problem for any of the technology I mentioned above: how to properly define security. Especially in the permissionless setting the effort on defining security has been unconvincing so far, requiring a very stretched approach to formalizing computational resources that is hard to actually map onto a real-world application. Satoshi did not start with a well-defined problem he was trying to solve with Bitcoin, and such an approach -- clearly identifying the problem you are actually trying to solve and verifying that the definition is logically consistent and realistic -- is exceedingly rare in the blockchain space, while in mainstream cryptography research it is a de facto requirement. So while blockchain tech has not experienced a spectacular failure due to some theoretical shortcomings, the theory itself is not well developed compared to the theory of cryptography in general (including ecash, which can be rigorously defined and proposed systems can be proved to satisfy the definition).

betterunix2··on On Web3 Infrastructure
Well, let's see, the technical program for this year's RealWorldCrypto conference includes...side channel attacks, symmetric cryptography, privacy, attacks on privacy, cryptography for the ads industry, messaging, post-quantum crypto, threshold crypto, and zero knowledge proofs. Funny how Blockchain did not make the cut for a popular conference on practical applications of advanced cryptography.

Looking through the CRYPTO'21 and EUROCRYPT'21 conferences, there is only one session between them involving "blockchain" and it is not even dedicated to the topic (it also includes papers related to law enforcement and privacy).

This may be hard for blockchain enthusiasts to hear, but it is not really a hot topic among cryptographers. There was a bit of interest a few years ago as people tried to figure out if any good security definitions can be developed, and the results were not very convincing. Beyond that almost all the academic interest in block chains has focused on the "permissioned" setting where security can be defined in a meaningful and useful way.

So, yes, "crypto means cryptography" and "blockchain" should refer to block cipher chaining modes.

betterunix2··on On Web3 Infrastructure
"IACR, was founded by people who were working on cryptocurrency"

This is extremely misleading. IACR was proposed by David Chaum in 1982, a year before his first paper on ecash, and its first board members in 1983 included Whitfield Diffie (who had not done any work on payments) among others. Chaum's ecash ideas look nothing like "cryptocurrency" as it exists today, nor do any of the ideas presented in subsequent research on the topic.

The person you replied to correctly pointed out that "crypto" was coopted by the blockchain space and is now being used to mean any number of distributed systems technologies. I have seen people wearing t-shirts saying things like "crypto means cryptography" and making jokes about reclaiming "blockchain" to refer to block chaining modes at various cryptography conferences (many organized by IACR) over the past decade. Moxie was right when he quipped that the "crypto" spaec involves very limited use of actual cryptography.

betterunix2··on On Web3 Infrastructure
...why do you have to have a peer-to-peer network managing a blockchain just to verify membership data? It would be simpler to just have whoever controls memberships sign a user's public key to indicate membership (and then you add some metadata for expirations/membership levels/etc.). A user seeking to assert membership presents that certificate and proves knowledge of the private key.

Really though, I am not sure what problem is being solved here. Why do we need this in the first place? What is the attack scenario we are trying to address?

betterunix2··on The Pirate Bay Token
Actually it is backed by US law and the strength of the US government in terms of enforcing those laws. In particular, the dollar is backed by US tax and debt laws; US taxes must be paid in dollars, debts in the US can be cleared with dollars, and US courts deal only in dollars when working out damages or in bankruptcy proceedings. As long as people doing business in the US follow those laws, which ultimately requires them to respect the government's ability to enforce the law, demand for dollars will exist and thus the dollar has value.

So it is not really about reputation or violence being used against those who would "interfere" with the dollar (whatever that means). I would argue that the value of the Yen is more influenced by reputation than that of the dollar, since people tend to treat Yen as a "safe haven" during economic crises (but ultimately the Yen's value primarily follows from the Japanese government's ability to enforce its laws).

betterunix2··on The Pirate Bay Token
As an American, I have to pay taxes and abide by US laws and the rulings of US courts, and all the above work exclusively with USD. So I cannot switch away from the dollar completely; whatever business I conduct in some other currency, I would still have to convert back to dollars at least some of the time. So you cannot crowd out the dollar with alternatives, because hundreds of millions of Americans are required to use the dollar regardless of the availability or popularity of the alternatives.
betterunix2··on Forced to sell chipless ink cartridges, Canon tells customers how to bypass DRM
There will always be some demand for printing, even if it is reduced over time. For example I and several my teammates routinely print research papers so that we can take pens/highlighters and mark them up or put a post-it note on some page of interest. Sure, you can theoretically do all the above on a computer, but even with the highest-end tablets I still find it cumbersome to do something as simple as drawing a big circle around a paragraph of text.

That said, printers are certainly a mature technology at this point and I doubt there is much room for disruption.

betterunix2··on The Pirate Bay Token
What exactly is the difference? Censorship for the purpose of protecting business interests is still censorship by any reasonable and honest definition. Even moreso considering the history of copyright law and its relationship with censorship regimes...
betterunix2··on The Pirate Bay Token
It's funny, many years ago when Bitcoin was relatively new I and a number of others pointed out that there is no particular reason why it should have any value, since there is zero barrier to creating your own cryptocurrency and therefore supply was effectively unlimited (and thus prices should decline long-term). Various unsatisfactory answers were given at the time, and since then the USD price of Bitcoin actually increased very significantly. Yet every time I see something like this, I immediately think back to that argument -- these alternative "tokens" or "coins" or whatever people call them are simple to create and seem to require minimal promotion to start attracting buyers. I wonder if we will eventually see this space become so crowded that those early objections wind up playing out...
betterunix2··on My First Impressions of Web3
Distributed storage does not make any difference for lock-in with a centralized API. For example, imagine a system for storing photos on some distributed system and a popular, centralized web front-end for users. Now what I will do with the centralized front-end is to give users a "value-add" by encrypting their photos, thus protecting their privacy, and better still I will use my proprietary key management technology to relieve end users of the various problems with losing private keys. Lock-in achieved, and all you accomplished with distributed storage was to outsource the maintenance of the storage infrastructure.

We already see this with blockchain payments. The vast majority of merchants who accept cryptocurrency payment do so through a service that manages their wallet and typically offers some kind of value-added features to lock them in. There is no reason to believe the same will not happen with Web3, if it is not happening already.

betterunix2··on My First Impressions of Web3
I was referring to technical solutions, not fines or regulatory measures. For example, before Bitcoin cryptographers published a mountain of research on designing secure and anonymous electronic payments, but relied on a central bank that issued and redeemed the money. The bank was constrained mathematically so that it could not link user transactions, unless some subset of users had cheated in some way (double spending). So there was a central party but certain forms of abuse were impossible, and those systems were overwhelmingly more efficient than Bitcoin or even a proof-of-stake approach ever could be (this is because transactions are "truly" peer-to-peer, meaning that only two parties do any work at all when a payment is made or when money is withdrawn from or deposited with the bank; moreover the work required to perform transactions amounts to verifying a few signatures/NIZKs). Another example is the use of oblivious RAM for secure cloud storage, which both protects user data and ensures that "most" of the access pattern (everything but the number blocks of data a user has accessed) remains private. There are also many examples of real-world deployments of secure multiparty computation that limit abuse by large/centralized parties in various ways while still allowing those parties to operate and even expand their business (without having to collect more user data than they already collect).
betterunix2··on My First Impressions of Web3
Do you actually think blockchain tech is remotely competitive with the big platforms? Blockchain payment systems have had more than a decade to become popular and still are not even remotely competitive with the big payment processors. Most of the world will only read about "Web3" on some news site or blog, then ignore it because it does not even come close to meeting their needs.

Consider how many people post something on Facebook in a single day, and now consider what it would take if each post had to be replicated across tens of thousands of independently operated systems. Big tech companies scale in large part because of their centralization, which allows them to coordinate large numbers of physical machines to efficiently provide service to their users. You may not like the ads-centric business model but on a purely technical level it is pretty clear that the big tech companies have a big advantage in terms of operating their infrastructure, and overcoming that advantage is not going to be easy for any distributed system.

I personally prefer to focus on mitigating/preventing abuses by a central authority/component of a system, which almost always results in a far more efficient and reliable solution that trying to eliminating all centralization.

betterunix2··on My First Impressions of Web3
Github relies on a distributed storage architecture (local git repositories on developers' machines) and in theory anyone can take a project from github and duplicate it on gitlab etc. In reality nobody bothers and a project hosted on github will remain exclusively hosted on github and nowhere else, and likewise with other git hosting services. For the most part nobody cares if the data is hosted on a distributed system or a centralized one, because the overwhelming majority of users will rely on the front end. Changing RPC urls is not as easy as you might think, especially for systems that are widely deployed and have heterogenous clients (which in theory would be the case if Web3 ever took off, which I personally doubt).
betterunix2··on My First Impressions of Web3
Why would we ever move toward decentralization? It is almost always easier to have at least some central point of control in any distributed system, even the Internet (IANA, RIRs, etc.). It is also very difficult to remove a centralized control point after a system is already deployed, especially if the system supports heterogenous clients (as it is likely that some clients will be slow to switch to the new design, and many will make bad assumptions about the system architecture).
betterunix2··on My 70 year old mother has been using Linux on the desktop for the past 21 years
My mother has been using Linux exclusively since I was in high school. Occasionally I have to head over to her place to deal with some weird issue but it has become less and less common over time, to the point where I am almost surprised when problems arise.

Of course she has very straightforward needs. If Firefox and VLC work she has 95% of her needs met, and the remaining 5% are related to her printer which has not been a problem for a long time now. For some people who need specific Windows software this is probably much harder, and likewise for anyone who purchase some obscure hardware that only has Windows drivers.

betterunix2··on Farmers facing fertiliser sticker shock may cut use
Pretty sure he meant carbon emissions, but technically you are correct, "steel" is a mixture of iron and a relatively small amount of carbon. As far as I know (but I am not an expert) there are not many applications for elemental (pure) iron.
betterunix2··on Omicron at 100% Prevalence, Colorado
It is still more dangerous than the other endemic coronaviruses that cause the common cold. In general the selection pressure on viruses (and other parasites) is to be less disruptive to the lives of their hosts -- not just being less deadly, but also not incapacitating hosts for extended periods of time. How long it takes for a virus to become endemic-without-ill-effects can vary widely, but the hope here is that COVID19 will do so quickly, becoming "just another common cold" within another year or two.
betterunix2··on Omicron at 100% Prevalence, Colorado
Ebola is not endemic to humans. It is not lethal to the bat species in which it is endemic.
betterunix2··on Omicron at 100% Prevalence, Colorado
At this point that is the best case. COVID zero is a pipe dream, there are just too many vaccine refusers and too many people refusing to do anything to avoid spreading the virus. If COVID follows the typical evolutionary pattern of viruses and becomes less dangerous, we can get back to normal life -- the question is how long that will take.
betterunix2··on Omicron at 100% Prevalence, Colorado
Yet it is predominantly the unvaccinated who are overwhelming hospitals with more severe cases. In NYC it is something like 15x difference -- 30 per 100000 unvaccinated versus 2 per 100000 vaccinated people wind up requiring hospital care. Those may be small numbers, but with omicron clearly being both more contagious and evading prior immunity it rapidly becomes problematic.
betterunix2··on Omicron at 100% Prevalence, Colorado
Yet the unvaccinated are still overwhelmingly more likely to wind up requiring hospital care than the vaccinated.
betterunix2··on Omicron at 100% Prevalence, Colorado
Not very hard to find the current, best understanding of the phenomenon:

https://www.ninds.nih.gov/Current-Research/Coronavirus-and-N...

betterunix2··on Show HN: WordLines – A zero-knowledge proof-based blockchain puzzle game
"there are no clean non ZK mechanisms to prove your solution outside of sharing the solution itself"

This is not entirely true. Witness hiding proofs are not zero knowledge (no simulator) but also do not reveal the solution. It is a subtle but important difference that can matter in some situations -- for example, a witness hiding proof could potentially allow the verifier to prove a related statement (e.g. proving that the proof protocol was executed).

← PreviousPage 5 of 34Next →