I feel like sharing a curious e-mail encounter I had with IOTA founder David Sønstebø. After writing a blog post about the limitations of the Bitcoin blockchai

I feel like sharing a curious e-mail encounter I had with IOTA founder David Sønstebø. After writing a blog post about the limitations of the Bitcoin blockchain, which got some attention on HN[1], I received an e-mail by someone named David (I didn't know who he was at the time), asking me: > Hey Rune, > I read your interesting article analysis of Bitcoin limitations. I was wondering if you have had any time to check out IOTA (www.iota.org) and the p2p Flash Network within it (https://blog.iota.org/instant-feeless-flash-channels-88572d9a4385 https://blog.iota.org/instant-feeless-flash-channels-88572d9...)? > [..] Assuming this was just some random person asking for my opinion on IOTA, I replied with a critique, pointing out two weaknesses I was familiar with: > The only thing I know about IOTA is that it’s centralized, through the so-called Coordinator. This makes it uninteresting to me. As I understand, the IOTA team argues that this is just a preliminary precaution, which will change as the size of the network grows, but I’m skeptical of this claim. In any case, until it actually becomes decentralized, I don’t think I will have enough interest in it to learn how it works. So I think I will wait for this to happen before taking the time to learn about the system. Also, I know that the IOTA team designed their own hash function, which turned out to be vulnerable to collision attacks, which sounds rather amateurish to me. He promptly replied to this critique -- which I thought I was offering someone seeking advice on the soundness of IOTA -- informing me that > As the founder of IOTA I can answer these questions: [...] Proceeding with a rebuttal[2] of the weaknesses I had pointed out, even though he was the one who had asked me, despite the fact that he didn't need any information about IOTA at all in the first place (given that he created it). [1] https://news.ycombinator.com/item?id=15427662 https://news.ycombinator.com/item?id=15427662 [2] > 1) The Coordinator is quasi-centralized, you as a programmer can easily opt out of it if you want, it is not enforced upon you as a user, it's simply a "best practice" at the moment. Tangle is made to scale, so the argument is indeed that the Coordinator is only now in place to prevent against the 34% attack that all DLTs suffer from until the network has scaled. I don't see anything controversial about this, it is the only way to reach a truly decentralized scalable ledger. It is no different from Satoshi firing up his first miners to get the Bitcoin network to work in the first place. The Coordinator will very shortly also be distributed to consist of numerous nodes, at which point it will be a lot more decentralized. These are all well-known steps towards the long-term goal. IOTA never claimed to be production ready, nor do we do any handwavy nonsense like Ethereum and Bitcoin core mantras "we'll solve it with some computer science breakthroughs in the future", IOTA's roadmap is very simple and straightforward. > 2) The hash function story has been so misrepresented and blown out of proportion it is comical. I'll spare you the pointless drama and conflict of interest from the guys carrying out the hit piece and only focus on what is important, in short: We spoke with the Keccak team back in 2014 about creating a trinary sponge based hash function for the inevitable arrival of trinary processors (we have designed our own as well) as ANNs, photonics, spintronics etc. all favor trinary over binary, hence we need a trinary hash function that is lightweight for IoT which utilize such chips. 'Curl' was born. The best way to thoroughly vet a hash function (which is vital) is to put it out there with a big incentive to crack it. This is what we did. Even if someone had broken the hash function entirely (no one ever did) there was no threat to the network due to the precautionary steps in place, and the fact that we had Keccak as back up as safety precaution #10. Right now we are working with several world-leading cryptographers, including your fellow countrymen of http://cybercrypt.dk/company/ http://cybercrypt.dk/company/ on further developing and optimizing Curl. This is far from amateurish, it is simply leading the way through genuine invention. Lightweight hash function development is a very active field of research. > 3) We also invented full Proof of Stake, the first decentralized exchange, the first decentralized voting protocol, decentralized marketplace in 2013, pioneered using blockchain for ID, Supply Chain and IoT in 2014-2015. So while we're known to push boundaries, we have so far always been vindicated later on. IOTA being the first ledger to disrupt blockchain itself is a testament to this, but is naturally very hard to swallow for a lot of blockchain maximalists. However, when it comes to actual researchers the reception tends to be very positive, and as for companies we work with everyone from Cisco to Maersk to Bosch to Microsoft to IBM to Statoil, so outside of the niche cryptosphere the interest is mounting daily, due to the fact that they have concluded that they can't use blockchain due to its inherent scaling and fee limitations. I mention this so you realize that we didn't just hack some random shit together like most people do in this space. > 4) [...]