Hacker News Reader

Comment by naskwo | Hacker News Reader

Given the order-of-magnitude increase in speed (https://www.theregister.co.uk/2016/04/18/d_wave_demystifies_...) combined with the assumption that the price of these chips will go down as commercialization becomes more viable, isn't it naive to even consider the viability of cryptocurrency? I mean, if the core underlying value of this "asset class" is the fact that there is a scarcity of numbers (like collector baseball cards) combined with ease and anonymity of transactions (like diamonds), then unlimited computing power will demolish the value of cryptocurrencies as much as alchemy would diminish the value of Au.

View on HN

There are several misconceptions about what universal quantum computers could do. One of them is talking about exponential increase of computing power with each new qubit. It just doesn't work like that. This brief talk from Scott Aaronson tackles this issue quite well: https://www.youtube.com/watch?v=JvIbrDR1G_c

While quantum computers threaten some cryptographic schemes, it will just be the end of those schemes, not the end of cryptography in general (see: post-quantum cryptography). As long as we still have problems that are hard to compute and easy to verify, we can have cryptocurrencies. Quantum computers do not give "unlimited computing power" and do not threaten the existence of cryptocurrencies in general.

As far as I know, there isn't a conclusive decent public-key algorithm that is quantum resistent.

We know that Diffie Helmann, RSA and their elliptic counterparts are broken in quantum. That was essentially all we had. I know that 'lattice based cryptography' is a potential option, and believe there are more potential options. However, as far as I know it is not even clear whether these are safe against classical computers.

I should note that I don't know much more about 'lattice based crypto' than the name and the hope it is quantum resistent.

Merkle-Damgard hash functions (such as SHA256) are merely weakened by QC. SHA256 is considered good enough and as such, BTC mining should be able to weather QC. There is a variant of Shor's Algorithm that can solve ECDSA, so you might have problems with your BTC wallet; a fork using a new resistant asymmetric algorithm would be needed.

Memory-constrained algorithms, such as scrypt, should be quantum resistant based on what I know (which isn't much). If I'm not horribly misinformed, LTC mining should therefore be resistant. Again, LTC wallets would face problems.

Does the same hold for hashes based non on Merkle-Damgard but the sponge construction?

It looks like we don't know, excluding some very specific sponge functions, based on some quick research.[1][2]

[1]: https://eprint.iacr.org/2017/771.pdf [2]: https://crypto.stackexchange.com/questions/419/what-security... (NB: comments)

Order of magnitude increase =/= unlimited, right? Also, it's my understanding that there is already plenty of work done on post-quantum cryptography. I don't know if it's entirely a solved problem, but my understanding is that it's at least solvable. Perhaps someone with more domain experience can weigh in.

There are numerous options for quantum-resistant algorithms. Existing cryptographic currencies can fork to whatever seems most suitable. This wouldn't be a trivial change, but it can certainly be done. https://en.wikipedia.org/wiki/Post-quantum_cryptography

There is also no such thing as 'unlimited' computing power with regard to proof of work. Difficulty simply adjusts, and the cost of a network attack will always be related to the rewards of mining.

crypto-currency is not about scarcity of numbers its about proof of work. Bitcoin started with relatively little computation power and has now scaled to where more power is spent than a small country uses in the mining operation... Its unfeasible to mine bitcoin on a cpu or gpu because ASICs are just so much more efficient. None of this scaling in hashing rate really changing the underlying concept which is that you prove you did some work in order to add blocks. If quantum computers do in fact result in a couple of thousand orders of magnitude raise in the hashing rate... well then miners will use different hardware the difficulty of the network will adjust each 10 minutes as new hardware gets added, and after the change you'll just have a large army of quantum miners instead of ASIC miners.

Remember the exitsanse of quantum computers does not prove P=NP, which is to say, even with quantum computers you can still have the class of problems which are difficult to solve but easy to verify which is the essential component of proof of work.

> If quantum computers do in fact result in a couple of thousand orders of magnitude raise in the hashing rate... well then miners will use different hardware the difficulty of the network will adjust each 10 minutes as new hardware gets added, and after the change you'll just have a large army of quantum miners instead of ASIC miners.

If it's a couple of thousand orders of magnitude, the first person with a working quantum chip might entirely dominate the hashpower, long before the chip is available commercially. So you'd end up concentrating hashpower into a couple of BigCorps's hands. Like, if I went back in time to the beginning of BTC with an ASIC I'd probably completely dominate the baby blockchain and outcompete the entire network.

No, there are quantum resistant protocols that could be switched to instead when the gap starts to close.

There are several quantum resistant crypto currencies. I.e. IOTA

Can you please explain what makes it quantum resistant?

Not all crypto becomes easy to crack with a quantum computer: https://en.wikipedia.org/wiki/Post-quantum_cryptography

With its use, the scarcity of numbers remain, that is, finding a clash that would allow to forge a transaction stays infeasible, and mining stays hard.

BTW while mining is so important with Bitcoin and Etherium, it's not a necessary part of a cryptocurrency. Mining is an incentive to keep doing blockchain validation. Some currencies exist without it (e.g. NXT).

All crypto becomes easier (well, except for the one time pad). But some algorithms (AES, SHA) only require doubling the number of bits to achieve the same security level, which is rather trivial to do.

To date the only practical replacements for RSA and EC achieve key derivation but not signatures. Without signatures we're not really 'complete' since certificates are hard to do with only key derivation.

Post-quantum cryptograph research is a very active area right now. If these researchers do not find a practical signature algorithm, PKI as we know it will change very significantly.

Ironically, blockchains might elevate in importance as a result, since the evolving blockchain is a useful construct for quantum resistance. It has cryptographic agility built in.

"IOTA uses hash-based signatures (https://www.imperialviolet.org/2013/07/18/hashsig.html) instead of elliptic curve cryptography (ECC). Not only is hash-based signatures a lot faster than ECC, but it also greatly simplifies the overall protocol (signing and verification). What actually makes IOTA quantum-secure is the fact that we use Winternitz signatures. IOTA's ternary hash function is called Curl."

IOTA isn't even resistant to differential cryptanalysis.

Care to add sources?

A flaw was found in the hash function they rolled themselves which doesn't build confidence there aren't more vulnerabilities to be found:

https://medium.com/@neha/cryptographic-vulnerabilities-in-io...

https://www.forbes.com/sites/amycastor/2017/09/07/mit-and-bu...

To make things even worse than rolling their own hash function, they're claiming the flaw was intentional! This makes them hostile and untrustworthy to the open source community if this is true and further erodes confidence they know what they're doing if this is their attempt at PR spin:

https://hackernoon.com/why-i-find-iota-deeply-alarming-934f1...

> Next, and in my mind most damningly, Sergey Ivancheglo, Iota’s cofounder, claims that the flaws in the Curl hash function were in fact deliberate; that they were inserted as ‘copy protection’, to prevent copycat projects, and to allow the Iota team to compromise those projects if they sprang up.

These points have all been talked over extensively and are not a concern: https://medium.com/@comefrombeyond/cfbs-comments-on-https-ww...

That blog response is completely unconvincing and makes me trust them even less. They have no proof of the extraordinary claim that the flawed hash function was intended (which is bad whether this is true or false) and their network going down for three days is a huge red flag that the protocol cannot be used in a decentralised way. He talks a good game but if you have any technical background you can see he isn't addressing what's being asked and is avoiding valid criticism via hand waving and character attacks.

I disagree with you entirely.

The creator requested an open, live debate to put these claims to rest. They haven't accepted and probably won't.

Reply on news.ycombinator.com