It's not decentralized if there's a trusted oracle.
So many schemes in the "blockchain" space fail to understand the core problem Bitcoin solves. Bitcoin solves trust. You can transact Bitcoin with anybody and you don't have to trust a single person, even the sending party. If someone sends you Bitcoins and the transaction is 6+ confirmations deep in the blockchain, you can be certain that you own those Bitcoins and nobody else can take them from you without your private keys. Not a bank, not a government, not a miner, not a "core dev", not a full node operator. The only necessary assumption is that 51% of the hash power is not conspiring against you.
Whenever you see a new "blockchain" scheme coming out, ask yourself who you need to trust for it to work. If the answer is not "nobody", then it is not worth any more of your time.
It pretty explicitly declares who Dai-holders are trusting- those who hold the rarer MKR coins. MKR-holders, presumably, appoint oracles by manipulating the Ethereum DAO that's running the whole business- but they also seem to have control over the rest of the system as well, so its not as though they would have to even be as subtle as manipulating oracles to mess things up. If they want to declare their own farts to be valid collateral to borrow Dai, I don't see a mechanism provided to stop them.
Its really easy to see the MKR-holders as just a formalization of the movers of high-finance in the fiat currency world- they would have a similar oversight role and powers as the high-level bankers and governmental financial regulators that control the workings of conventional markets. I suppose the idea is that MKR-holders have it in their best interest to defend a system they've invested in. But if the motivation of the whole digital currency movement is autonomy this seems to defeat the point.
51% of MKR holders versus 51% of miners is a different threat model, but "ability to make MKR" and "ability to own mining kit" both seem to depend on financial expenditure, so I wonder if the difference in practice will be that big?
> The only necessary assumption is that 51% of the hash power is not conspiring against you.
That's not true. Even if a miner has 51%+ hash rate, they can't "steal" your coins. They can, however, prevent you from moving your coins but that puts them at risk of getting forked. (unless they have way more than 51% of the hash rate)
Yes it is true. If a miner has 51% hash rate, they can in theory steal your coins by rewriting history to replace the transaction that granted you the coins with a different transaction that sends the coins somewhere else. Or, if your coins are too old for that to work, they can DoS you by refusing to mine on any block that includes your transaction, essentially rendering your bitcoins worthless.
In practice these attacks probably need significantly more than 51% hash rate, but there is some % hash rate where they would work. Of course, evidence that such an attack was being carried out would immediately crash the Bitcoin market, which would destroy the value of the miner's investment in mining hardware. (A crude form of proof of stake?) So it's hard to imagine that such an attack would actually happen in reality, unless it was carried out by someone with the intent to destroy Bitcoin and a few billion dollars to burn.
> If a miner has 51% hash rate, they can in theory steal your coins by rewriting history to replace the transaction that granted you the coins with a different transaction that sends the coins somewhere else.
Nope. Unless they have your private key, they can't sign your coins no matter how much hashing power they have.
What they can do is unspend their own coins. So if they paid you BTC$1,000 and you gave them a yacht they can go back and send that money to someone else (double spending). But even that is limited:
You can think of it like a race. With 51% of the network you're slightly faster than the other half. That means you can probably win disagreements going forward. But in order to change past transactions you have to go backwards in the race and then catch up again.
Your ability to do that is proportional to how far back you are going and how much above 50% you have. If I want to go back one block (ten minutes) and I am going at 1.1x the rest of the network, it will take me 100 minutes to catch up (I can do 11 blocks to their 10).
If I want to go back 2 blocks (20 minutes), it will take me 3 hours to catch up. You can see how this gets out of hand quite quickly.
They don't need to sign your coins. They need to sign the previous owner's coins. And they don't need to literally be the previous owner, they just need the previous owner's cooperation. And in fact they don't even need that: they can omit the transaction without knowing anyone's keys, and they've still deprived you of your Bitcoins. And they don't need to go back in time, they can omit all of your future transactions and thereby make your Bitcoins worthless, or hold them for ransom.
True, but the farther back they want to rewrite history, the more hashpower they need to outrace the rest of the network, since they're starting from further behind.
> replace the transaction that granted you the coins with a different transaction that sends the coins somewhere else
If the party that sent you the coins controls (directly or indirectly) >50% of the hash rate, they can reclaim the coins they sent you by omitting the transaction in a fork.
No amount of hash rate lets you send coins to arbitrary places - you need the sender's key to sign a transaction.
They don't need any keys to omit transactions and therefore deprive you of your coins, or hold them for ransom. It's true that they can't gain control of the coins without the keys, but that doesn't change the fact that you lose the coins. And they can potentially get keys from people by promising to pay them in stolen Bitcoin.
I don't see how any of that has any bearing on my objection, whatsoever. My point wasn't to say "WAT NO BITCOIN IS INVULNERABLE". My point was to correct the inaccuracy in the bit of your post that I quoted, lest it confuse others and for your benefit as well if it reflected actual misunderstanding.
replace the idea of "stealing your coins" with "depriving you of your coins" and it makes more sense. the attacker wouldn't be able to deposit the coins in any desired wallet, but would be able to create a fork of the chain that omits the transaction that put the coins in your wallet.
Replace "blockchain" with "merkel trees" and you get interesting things like certificate transparency, which is certainly "worth your time".
Hype aside, public, shared, consistent, irrevocable ledgers are useful for all sorts of things even if you're not attempting to solve the "no trust" problem.
You have a lot of people watching the trusted source to make sure that it's append-only and follows the rules for appends. It's a reputation thing; not impossible but it will be detected.
Actually I meant the source tree, and it's because Linus signed the git commit, and thus the tree and all its history. But yeah, the distro works too, which is the whole point: centralized trust does work without a solution to the "no trust" problem. And it works because you build an identity that is consistent over time. I can trust the 4.10 tree because it's signed by the same keys that have been signing kernel keys for years, which are themselves cross-signed by a bunch of trusted identities who have similarly built histories.
This is about tracking the source code using git. The specific thing being guaranteed is that they won't rewrite history; if they rebase you (and many others) will detect it.
Of course you're still trusting them in other ways.
Bitcoin only "solves" one part of trust. The blockchain transaction is perfect but to be of any use there is a corresponding exchange of some service or good or title. Even if the transaction is to exchange currencies from two different chains, trust or a trusted escrow or trusted multiparty consensus is needed.
Bitcoin has a trusted bunch of bootnodes hardcoded into each client. If they act poorly, its trivial to swap them out. Just like its trivial to swap an exchange rate oracle.
Not really. It also fetches new bootstrap nodes from DNS servers. But neither is very much trusted as it takes only 1 valid connection to relay other valid network addresses.
I'm not sure what you intend to mean. If you intend to mean that you must trust your ISP, that is false. ISPs cannot forge blocks due to the proof of work. The worst your ISP can do is a DoS attack delaying block delivery, and that is fairly easy to detect.
You can end up on a less work chain because your ISP filtered all connections to the real bitcoin network and you only connected to Sybil nodes. You never hear about the real blocks, just the handful of blocks mined by the attacker.
No matter what your ISP does you can still verify the proof of work. So your ISP's forged blocks still need the full difficulty PoW that the real chain had at the fork point. Which means that in order to generate a forged block more often than once a day, your ISP will need a significant fraction of the hash power of the whole network, costing probably tens of millions of dollars to buy and operate. Even so the forged blocks will come much, much slower than the real blocks, and that will become noticeable after just a few blocks.
Significant? No it only takes a few percent of the hashrate to make a few blocks to successfully Sybil an infrequently connected SPV node, or a full node that has been partitioned off for which you have sufficient time to prepare the alternate chain. You're not outracing the network, just making sure the victim doesn't have a view into the real network.
How can this work if there is no Oracle? Think about it, the price of USD/ETH has to be retrieved from somewhere. It's not in the blockchain, so it must provided in an Oracle.
There is not one trusted oracle. The oracles are determined by the MKR holders, and MKR holders can easily replace oracles if they act badly. MKR ownership is decentralized, hence I don't think it's accurate to say there's one trusted oracle.
> The tokens have been continuously sold off in small batches by the Dai Foundation. Currently about 55% of the total supply has been distributed, with the foundation still having another 45% for future fundraising.
And here where the "Scam" alarms go full retard, basically panicking the whole building I'm living in. The concept is yet to be finalized, let alone functioning and yet these guys started selling "tokens" a la "DAO".
I wonder when regulators are going to strongly hammer the ICO (or the crypto equivalent of IPO) by a 15 year of jail and catching bad jokers inland and overseas.
I don't see stability as a goal right now with leading cryptocurrencies (ETH, BTC). In fact, promoting cryptocurrency as a storage of value right now seems off
too: one day, not today.
More people will store more wealth in cryptocurrencies as the number of applications grow. Being able to pay for my Chipotle, or renew domain names, or buy stock in companies, or run a website all make me more willing to keep more of my wealth in ETH. Eventually stability will come as the relative utility of a currency pushes volume up.
The last thing I want is stability as a believer in the Ethereum platform. I want the price of Ether to go up pursuant with the growing utilities Dapps provide.
In the interim, solutions like Bitpay that pay merchants in USD over BTC are brilliant. They allow merchants to buy into cryptocurrency without accepting the volatility risk until it's low enough.
So this is trying to create a currency that is democratically controlled by the "MKR" holders - I don't really understand how this is fundamentally different than something like the US dollar. Sure, the central bank controls a lot of the system, but they are also 'democratically controlled' by our ACTUAL democracy.
Now, I am sure there will be an argument made that the central bank is beholden to special interests, who use their wealth and power to control our democracy... but won't that same thing happen with these 'MKRs'? The wealthy and powerful will have more control over them, and will certainly use that power and control in their own benefit.
It seems like a common flaw in all of these 'alternative' systems people create to replace large scale institutions; they might seem different than what we have now, but when you scale it up to match the size of the current system they are aiming to replace, you end up seeing that the systems are nearly identical in practice.
I think one of the difficult things is that volatility (and speculators) do play a role in a coin succeeding, as speculators make the 'network effect' problem of altcoins easier to overcome.
There's next to zero reason for a random merchant to accept an altcoin unless there are people using it, and unless that coin is known. Speculators and volatility are very big generators of noise that gets the coin noticed and helps bring in actual for-the-sake-of-the-coin holders, which then makes the coin itself more attractive to a merchant (as there are people who wish to use the coin).
As much as coin purists may like to speak derisively of speculators, I believe that altcoins would be nowhere without those who have come in to the game to try to earn money off of the volatility.
I agree with you on volatility playing an important role in bootstrapping the network in the early days of digital currencies like Bitcoin and Ethereum.
I think stablecoins will play out differently though and are an important component of attracting new users. I believe decentralized applications will be built that bring mainstream users in, and these users will want a stable store of value to use these applications.
Stable coins still have volatility, however the volatility is pushed to a different set of players (in this case the mkr holders). These mkr holders have a big incentive to promote adoption of dai (the stable coin).
I think it was Ven that was the concept I liked where you build a currency on a bunch of commodities targeted toward stability. Clive Robinson on Schneier's blog had previously described that as the best solution then I stumbled on Ven.
So, my scheme was to put something like Ven in a sane country such as Switzerland under a nonprofit/foundation with strict rules aimed at preserving stability, security, innovation of ecosystem, and cap on admin overhead. If that works out pretty well, we can try a decentralized version of the same thing. Meanwhile, centralized can already work very well. Just eliminate as many incentives to mess it up as you possibly can to make it better.
The big brains haven't explored this option enough since they want a crypto/algorithmic solution to hard problems instead of using the well-understood systems of law, accounting, and safe investments. There's also a strong preference for decentralize stuff highly-likely to fail vs centralized stuff way better than what we have. I think that's ideological or even youth-related.
,,While the volatility of Bitcoin has decreased significantly since it launched in 2009, it is still not a good unit of account or a stable store of value. ''
I will gladly accept his BTC from 2009 and give him USD from 2009 for the 2009 price.
Sorry, but stability is not holy grail in itself. I would prefer constant supply to it for long term store of value
Would you sign a 30 year mortgage denominated in BTC? I don't think so, it's neither a good unit of account or store of value. Looking back at the past when you know the previous value is not a good test; looking at the unknown future is.
However, I think if I had a reliable income that was measured and paid out in BTC, then the answer would be yes.
That's really the benchmark for when BTC can be used as a proper currency, when people are prepared to set their value in BTC.
At the moment it makes a decent transfer mechanism for funds where the quantity is calculated relative to another currency at the time of transfer.
If Bitcoin were to become a standard currency it will necessarily have to go though more volatility. Bitcoin's value, should it succeed, is significantly higher than where it is now (conversely, should it fail, lower). When its value does not match its utility it must eventually move. Stability increases utility. These oppossing forces will cause occilations like a plucked string. The occilations should deminish over time as the forces reach equilibrium. The point of balance is anyone's guess right now.
> However, I think if I had a reliable income that was measured and paid out in BTC, then the answer would be yes.
I think you haven't thought it through if you still think that. 2BTC is currently worth just over 2g in dollars. In 10 years time, it might be worth 10g in dollars or it might be worth 4g in dollars. You'd be crazy to agree to pay 2BTC a month for 30 years with a currency that doesn't have a predictable future value. The dollar has a predictable future value; stability is not about not losing value, it's about predictable value.
> That's really the benchmark for when BTC can be used as a proper currency, when people are prepared to set their value in BTC.
No, people can do that now, but they do it in relation to exchange rates of more stable currencies. BTC will be usable as a proper currency only when its future value becomes somewhat predictable, predictable enough to denominate long term contracts like mortgages in it without relation to another currencies exchange rate. It'll take decades to reach that point.
A currency without a reasonable long term predictable value, isn't a currency, it's a commodity.
So many schemes in the "blockchain" space fail to understand the core problem Bitcoin solves. Bitcoin solves trust. You can transact Bitcoin with anybody and you don't have to trust a single person, even the sending party. If someone sends you Bitcoins and the transaction is 6+ confirmations deep in the blockchain, you can be certain that you own those Bitcoins and nobody else can take them from you without your private keys. Not a bank, not a government, not a miner, not a "core dev", not a full node operator. The only necessary assumption is that 51% of the hash power is not conspiring against you.
Whenever you see a new "blockchain" scheme coming out, ask yourself who you need to trust for it to work. If the answer is not "nobody", then it is not worth any more of your time.