Vitalik had trouble solving the "fork" issue, that is, that when two forks arise in a PoS coin, there is no incentive for PoS miners to select one over the other. That is, they are incentivized to mine both simultaneously at the same time. He thought he may have solved it today, but didn't seem terribly interested in pure PoS, especially because you need to typically do a 100% premine.
There are other people I know who previously wanted pure PoS chains arguing about them for one reason or another now. I would guess that Satoshi himself/themselves also considered this (as it seems like the more logical means to construct a cryptocurrency network), but abandoned it for whatever reason.
There are other people I know who previously wanted pure PoS chains arguing about them for one reason or another now. I would guess that Satoshi himself/themselves also considered this (as it seems like the more logical means to construct a cryptocurrency network), but abandoned it for whatever reason.
Thanks a lot tacotime for making the thread and representing us. I actually did just formalize the PoS algorithm I had briefly sketched out this morning: http://blog.ethereum.org/?p=39/slasher-a-punitive-proof-of-stake-algorithm .If we get real overwhelming demand to use proof of stake, we will use something like this or something better, but that's not really likely IMO. An interesting idea there would be to also add some kind of proof-of-excellence based distribution model (eg. we'll put 10000 trillion ether into a contract which gives ether to people who solve very hard math and CS problems like integer factorization, AI board game algorithms, etc) so we still have some more issuance over time and aren't 100% premined.
Now, to get to the main point, the issuance model.
First, as mentioned by LeoC, proof of burn is a non-starter for our use case, because PoB cannot be independent. That is to say, with PoB an Ethereum client would need to also have a Bitcoin blockchain and would be dependent on the Bitcoin blockchain for validity checking. I absolutely do not want to have external dependencies in Ethereum - with the obvious exception of people creating SPV clients inside of contracts, which I think is a really cool idea with amazing potential for cross-chain atomic swaps, two-way pegs and the like.
Second, and more importantly, while I do understand the sentiment that cryptocurrency algorithms should be neutral and privilege no specific parties I would like to contest the claim that the concept of neutrality in the context of issuance is both (1) possible and (2) desirable. When you're putting BTC into the Ethereum fundraiser, you're financially empowering and incentivizing an organization that has produced substantial real results already, and is pledged to developing the Ethereum protocol. When you're putting BTC into a proof of burn, through the deflation effect you're effectively donating your money to large BTC early adopters.
When we talk about neutrality and fairness, what we really mean is that we want the currency supply to be distributed among as many large and diverse constituencies as possible. And this is actually the reason why I came up with Ethereum's compromise distribution model. With Bitcoin, the only way to get BTC was mining, and so BTC privileges the class of people with computing power. With Mastercoin, it was buying into the fundraiser, privileging the Bitcoin-wealthy. With Ripple, it's working for Ripple Labs or getting a bit at a giveaway, privileging the founders. With Ethereum, there are many ways to get ether: you can participate in the fundraiser, you can participate in the Ethereum project and receive salaries or bounties, you can mine, and we may potentially even end up doing a giveaway or two. This actually increases fairness compared to the pure mining model, since different groups of people are privileged in each of these categories (people with money, people with hardware, people with connections and development skills, etc) and all of these have a chance.
With regard to the "founders get some for free" factor, what people need to understand is that every currency, including BTC, QRK, MSC and Ripple, so far has privileged its founders in some fashion. The only difference is that in some cases the privilege is more subtle than in others. With Bitcoin, Satoshi got his 0.25X by mining it for a year when nobody else had heard about Bitcoin yet. With MSC, JR got his by running a fundraiser relatively quietly on Bitcointalk and then targeting media attention only after it was over. We are not seeking any avantage through obscurity; we will be targeting the limelight from before the first day that the fundraiser launches.
Ultimately, our ideology with regard to the distribution long-term is this. We believe that centralization and decentralization both have their value, and must be used at the appropriate times. Startups, in practice, generally have to be dictatorial. However, institutions that are at the base level of society should ideally not be controlled by anyone. To that end, the way that we are structuring both the organization and the issuance model is that we will have a large amount of influence at the beginning, but that influence will quickly decay over time as the years progress. Once we release the coin, we will inevitably immediately lose control, and over the next one or two years we even hope to turn the Ethereum organization itself into a DAO, sustaining itself through VC-like investments with ownership spread among thousands of people in the community. It's the same with the profit structure; yes, if Ethereum succeeds, we will profit handsomely, and fundraiser participants will profit handsomely. However, because we have permanent linear inflation, this profit is a one-time fact whose importance in the system will diminish to zero over a sufficiently long period of time. As T approaches infinity, ether becomes zero percent premined. Satoshi, on the other hand, has up to 5% of all BTC forever.
All in all, we believe that we have come up with a very interesting issuance model that combines together the benefits of all of the various models in small doses; we have Ripple-style premining, Mastercoin-style fundraising, but ultimately Bitcoin-style mining dominates over the long term, eventually coming to eclipse every other component of the issuance. It is ultimately up to the community to decide whether or not this issuance model is valuable.
Now, to clarify some facts:
* The founders will not be getting any BTC from the pool. We may get some not particularly high wages to live on, but our financial interest in the project is heavily concentrated in the value of our premine share.
* The founders are not getting 25%. The founders are getting 12.5% after one year when those funds actually become accessible, and the percentage will rapidly dwindle due to supply expansion, becoming 6.25% after 5 years. I'm sure that there exist five people today who together own 6.25% of all bitcoins, and Satoshi alone likely has that much. In Mastercoin, three people have 25%.
* The BTC from the pool will be going entirely to the organization, and we will start using it to pay expenses even as the fundraiser is running in order to get development moving quickly. We have plans on how we can spend money all the way up to $25 million; the higher-end expenses consist of things like Bitcoin Decentral-style incubators, a large amount of funds into cryptocurrency scalability research, funding very high-quality interfaces, decentralized exchanges, and other such infrastructure, and even funnelling money into projects that might work alongside Ethereum (eg. Bitmessage, KryptoKit, OpenTransactions)
* We are not trying to be a highly powerful HFT exchange. We are trying to be a generalized workhorse that can do anything. Identity systems, decentralized Dropbox, sub-currencies, you name it. HFT can be done through either a dedicated chain (eg. BitShares can release an Ethereum contract containing an SPV version of the BitShares blockchain and use that to allow units to move back and forth between the two blockchains, and then do HFT on their own chain) or using a semi-centralized system like OpenTransactions
* We are going to release clients in 3-4 langugaes, likely Python, Go, C++ and NodeJS. The main reason for this is to avoid monoculture from day one and to ensure consistency since it's very unlikely that the same bug will appear in all four.
* We will not put Zerocoin into Ethereum. But what you can do is set up a Zerocoin-based mixer as an Ethereum contract.
* We have not figured out how to do 100% trust-free fiat. The closest we've gotten is doing contracts for difference off of voting pools of price feeds. But even that is a massive improvement over having to trust an issuer.
* We will likely use 3-of-5 multisig to hold our funds. We will be the first cryptocurrency fundraiser that we know of whose exodus address actually has a 3 at the beginning.
