Hacker News Reader

Comment by lrvick | Hacker News Reader

Could also be 100% if their Proof-Of-Stake game theory does not work as well with real funds worth stealing as it does on testnet. Full disclosure I hold some ETH but I would not dare put all my eggs in that brazenly "do it live" basket.

One party can hold >50% of all funds and we would have no way of knowing.

I wish people would look at less energy consuming proofs of work that still allow us real world evidence of decentralization.

We need more experiments like Chia with proof of wasted disk space instead of proof of wasted electricity.

View on HN

I'm confused as to the purpose of ETH. On one hand it sounds like an investment that you can "hold" as you say. On the other it is a computing platform. So what is it you "hold"? Compute resources? Why is it valuable to allow speculation on compute resources? Can a person ever really reason about the cost of a system running on ETH? How do you plan for those costs?

Is is a compute platform? A cryptocurrency? Something else? Both?

ETH is what you use to pay fees to the network to transact on it. i.e. if I want to transfer you some "tokens" on the Ethereum platform, I would have to pay a fee in ETH. This makes it a speculative investment, since if more people want to use the network, there is more demand for ETH as a currency. Since there is not an unlimited supply of it, there is a supply-demand dynamic on ETH as a currency.

It is a computation platform first and foremost. You can pay gas in ETH. Every opcode has a set gas price, so executing a compiled contract on the network costs you some ETH. Why would you want to run code on ETH?

Most notably, it requires no 3rd party trust to execute the code whereas traditional cloud architectures require you to trust that the execution environment is kosher e.g. amazon vm is not compromised, amazon hardware does not have side channel attack.

It is ideal to write a token or currency on ETH platform as state changes like balances are better left completed on trustless architectures

Ethereum is the platform and ETH its native currency. It's much like Bitcoin but with better scripting and the ability to store arbitrary state.

> One party can hold >50% of all funds and we would have no way of knowing.

One party can hold >50% of all mining power and I find that far more likely.

>I wish people would look at less energy consuming proofs of work that still allow us real world evidence of decentralization.

Are there any candidates for this? I can't imagine a PoW problem where you can't get ahead by throwing more computers at it.

It's called a zkSNARK. One proves they performed any computation with a constant size proof of the result.

But most of the energy consumption is in the work, not the proof.

Constant size or constant time?

Both, but 'constant' here needs some qualification.

Generally, it is 'constant' w.r.t. the witness you are proving you know, and something like logarithmic w.r.t. the statement about that witness you are proving.

Ethereum could use SNARKs to verify certain computations more efficiently, but the purpose of PoW miners isn't really to verify computations, it's to make a 51% attack expensive. CODA protocol uses SNARKs to verify the entire blockchain history, but they still need miners for consensus.

> Are there any candidates for this?

https://nano.org/en/about/

They are doing beacon chain (a distributed unbiasable verifiable random number generator, constructed using commit-reveal scheme + VDF to fix commit-reveal biasability) to choose sets of validators for each of the shards. It is not your typical PoS design. I'm not sure how exactly validators for single shard will form a block, but even PBFT (tendermint style) solution would work in situation when you don't have to worry about membership of validator set.

one party already holds >50%, don't forget that Ethereum is a pre-mine scam.

Reply on news.ycombinator.com