Monday, December 24, 2018

Ethereum "developers" don't understand incentives and trust minimization in block chain design

It's a simple fact Ethereum is mostly centrally premined as seen on their blockchain: https://i.imgur.com/PYylAma.png

It's a simple fact Proof of Work miner participation is driven to:

  1. Profit
  2. Cover the recent sunk costs of aging equipment (by selling coins)
  3. Cover the electric costs of running the equipment (by selling coins)
  4. Most efficient equipment is most profitable which incentivises buying new equipment
  5. Have incentive to mine and secure the most profitable fork or chain (hashrate tends to follow price and profitability)
  6. Determine security and liveness of individual forks or chains (low value of incentives to mine can kill off a chain by being unable to solve current difficulty or make the blockchain very cheap to attack)

It's a simple fact that distribution via an ICO of a premine is indistinguishable from premines:

  1. Just like premine, and unlike PoW, completely free to do for a single self-appointed permissioned party
  2. Only the single self-appointed permissioned trusted party gets to profit from the ICO
  3. Only the single self-appointed permissioned trusted party is allowed to buy into the ICO while getting all proceeds from themselves and others
  4. ICO design has a built in incentive for a single trusted party to cheat for any percentage of premined coins at no cost (link)
  5. None of this is easily detectable thanks to pseudo-anonymity thus relies completely in trust in a single trusted party to refuse free money

It's a simple fact that purchasing power of coins is decided by the market value of those coins. Since on-chain incentives are made of the same coins, the value of coins determine value of incentives used by blockchains to guide behavior and security. This means the value of incentives on each chain including forks is decided by the owners of existing supply. The incentives include purchasing power of awarded rewards from producing blocks for miners. For coin stake holders on each fork, expecting increasing stake value creates incentives to purchase or hold coins, and expecting decreasing stake value creates incentives to avoid buying and selling coins. The importance of market value of coins is also shown by refusal of mining pools to collude to 51% attack their own chains by hurting their ability to recoup hardware losses and incentivising honest behavior.

Ethereum uses a block chain design that relies on trust in a single permissioned trusted party in charge of what happens to super-majority premine that could control value of incentives, security, and liveness of each chain after each fork. 1 trusted party is not trust minimized or decentralized.

It's a simple fact that decentralization (trust minimization) is exact opposite of trust in a single trusted permissioned party.

The entire point of 2009 Satoshi's invention of Proof of Work was to distribute control including the supply to everyone without a need for a trusted party. The design allowed anybody seeking the incentive of profit to compete for virtually all fractions of the supply - unlimited number of parties. The miners were also incentivised to sell coins and distribute the supply to others via markets getting the control over value of incentives into unlimited number of hands.

Unlike many altcoin developers that valued trust minimization over personal profit, Ethereum, BitconnectX, Onecoin, and many others failed to use even the most basic design known since 2009 for trust minimized distribution of control and supply while profiting from that choice.

A trusted party deciding over a large premine at each fork can effectively choose which chain to attack the incentives to mine, secure, or hold coins to guide miner and user behavior.

What are some examples the trusted party could do? No better place for examples than Ethereum. Ethereum developers have

  • created new hard fork just to edit state, censor user, and confiscate money (link)
  • push black lists (link, link, link)
  • choose on which fork value of incentives is safe and funded by premine (link)
  • choose on which fork value of incentives is attacked and gets no premine funding (link)
  • lie about premine significance (link)
  • design an unsafe fork putting funds at risk of replay attacks (link)
  • warn select users their attempts to use their chain will be compromised similarly, showing intent to attack as pre-emptive measure (link, link, link)
  • lied, deceived, or manipulated for profit (link, link, link, link)

It's a simple fact Ethereum is not decentralized or trust minimized by design.

Ethereum Foundation and anyone including Vitalik claiming any trust minimization after the premine and ICO demonstrate factually inaccurate representation of the properties of their design and incentives and suggest they don't understand the role of incentives in block chains.

Here's an example of a Ethereum "developer" who suggests control of coins (determine value of stake, incentives, security of each chain via markets) is "disjoint" and not relevant to Proof of Work security:

PoW is how decentralization is maintained. Ether holders and network security are two disjoint ideas. They seem similar so it would be an easy mistake to make though. Joseph Delong https://i.imgur.com/CFJ2Ymb.png

Strangely enough, Ethereum developers are aware of market based decisions calling it "market consensus" that's determined by coin holders which their design compromised from day 1: https://i.imgur.com/yhnAd46.png

It gets even more bizarre in Proof of Stake where stake determines block production directly instead of also via incentives. In fact, the incentives are so poorly thought out that slashing conditions used in Casper to combat "censorship" can not only be abused by hypothetical DDoS attacks, but even easier by large holders such as premining party or even cartels: https://i.imgur.com/kM9xWH2.png. Slashing has been strongly avoided by many proof of stake developers on purpose as they do not add significant improvement over actual stake market value being at stake while scare of those afraid of making mistakes and introduce possibility of abuse that matters less with higher centralization from premines or cartels: https://i.imgur.com/jqL2rdO.png.

Additionally, switching to Proof of Stake removes introduction of any new trustless distribution, requires permission from current coin holders and, thus, the trusted premine party to enter into consensus by selling you coins, and thus amplifies the dependence on trust in the Ethereum Foundation even further: https://i.imgur.com/HV9Espg.jpg

To claim premine does not matter for decentralized tech is to claim incentives do not matter for decentralized tech.

Creating an alternative to centralized trust maximized designs used by Banks, Paypal, Ethereum, Onecoin, BitconnectX are precisely why the invention of trust minimized design of Proof of Work for distribution from start was so significant and unique. It was not unique by using hash functions, cryptography, nodes, internet, or some limited proof of work that makes trust minimization and decentralization possible, but the combined implementation using all those tools in a unique manner. The unique approach was successfully utilized by Bitcoin and countless cryptocurrencies to create trust minimized networks since then. Ethereum's pre-Bitcoin era trust model's failure to meet even the minimum standard set 6 years prior for trust minimization is only their fault.

Simple fact is Ethereum developers consistently failed to design Ethereum to be trust minimized or decentralized, created the designs incentivising centralization and dependence on trust in a single party, while effectively mislead people while profiting.



No comments:

Post a Comment