Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Miners Aren’t Friends

blog.keep.network

Post reply on HN
Print view View original

61–70 of 256 posts

Re: Miners Aren’t Friends

#61

Post by prestwich »

scyclow wrote: post #21

It seems like a lot of these problems are due specifically to using proof of work. Do any of these problems go away with other consensus algorithms? I would at least expect that miners wouldn't have as much incentive to arbitrarily cause errors if they were using something like proof of stake. Also, how easy is it to detect that miners are doing these things? If it's not terribly difficult, then maybe a federated sys…

Hello, author here. I haven't put a lot of thought into this yet, but my first guess is that this could become more of a problem with staking. Stakers still need to optimize, but all the easy ways are gone. Can't build better ASICs for staking.

Detection is hard. Like with selfish mining, it looks just like normal network operation in most cases. Most obvious mitigations don't work well either. E.g. we could try to establish canonical tx ordering in the tx pool. But then we would have to come to consensus about what the txpool is. Which is a bit of a chicken/egg problem.

Re: Miners Aren’t Friends

#62

Post by barbegal »

mhluongo wrote: post #54
barbegal wrote: post #37

Earlier quoted context omitted.

Proof of stake doesn't work in practice unless we can find a source of randomness that is globally available, non-influencable and everyone can agree on [1]. At the moment, that source of randomness is completely elusive. It might be possible to create a random source from some form of binary astronomical event which can be widely verified but it would have an incredibly low bit rate of new randomness. One possible c…

Publicly verifiable randomness is possible without relying on astronomical events. Check out threshold relay by the Dfinity team (which we're porting to Ethereum), or RandHound and RandHerd [1]. 1 - https://eprint.iacr.org/2016/1067.pdf

That paper assumes that at most f out of at least 3f+1 participants are dishonest. In practice one can't make that assumption since sybil attacks are easy.

Re: Miners Aren’t Friends

#63

Post by konschubert »

barbegal wrote: post #37
nspassov wrote: post #2

To my limited understanding of Ethereum and Bitcoin, the network already has rules in place to punish nodes that misbehave or cheat. In the case of Bitcoin, that is part of layer 1 so it is independent of any application layers and thus it should always be at the core of the network. With PoS cheating is likely to become a bigger issue [1], since with PoW the miners who get punished will have wasted their time and el…

Proof of stake doesn't work in practice unless we can find a source of randomness that is globally available, non-influencable and everyone can agree on [1]. At the moment, that source of randomness is completely elusive. It might be possible to create a random source from some form of binary astronomical event which can be widely verified but it would have an incredibly low bit rate of new randomness. One possible c…

> does this apply both to blockchain-style proof of stake and BFT - style proof of stake?

Re: Miners Aren’t Friends

#64

Post by bfsg »

So many false assumptions in this post. As if only 1 miner is trying to mine a block. There is simply no way to know whether you or someone else is going to mine that specific block, it's totally random.

> There are two ways to increase mining profitability: generate more revenue, or reduce expenses.

No. There are more ways. Mining the right coin at the right time and selling it at the right time for example.

Mining has a lot of common ground with investing in the stock market. It's not that simple.

Re: Miners Aren’t Friends

#65

Post by ShabbosGoy »

Looks like the author of this piece is not aware that proof-of-work is antiquated and essentially obsolete.

There are way better methods to achieve distributed consensus, like Hashgraph or even proof-of-stake.

Edit: The critique that if the blockchain is slow, your Dapp will be slow is fallacious as well. You can mitigate that using side chains.

Re: Miners Aren’t Friends

#66

Post by MrBingley »

Can someone explain to me how on earth Ethereum is still a thing? People have lost hundreds of millions of dollars in the Ethereum ecosystem, many due to basic flaws in the "Solidity" language. For example, last November $280 million was frozen in Parity wallets because function visibility defaults to external rather than internal (see the previous HN discussion here[0]). Just for fun, I was looking through the Security Considerations[1] in the Ethereum docs, and here are two so-called "minor details":

- In `for (var i = 0; i - The `constant` keyword for functions is currently not enforced by the compiler. Furthermore, it is not enforced by the EVM, so a contract function that "claims" to be constant might still cause changes to the state.

WTF?! Trivial loops won't terminate to due (unchecked) integer overflow, and the `constant` keyword does nothing? It's a complete gong show. I'm not touching anything to do with this language with a 100 foot pool. Is the hype bubble just that big?

[0] https://news.ycombinator.com/item?id=14807779

[1] https://solidity.readthedocs.io/en/develop/security-consider...

Edit: Oops, it looks like I'm mixing up my multi-million dollar screw-ups. First in July $30 million was stolen due to the "public is default" debacle, and then the fix for that problem introduced another bug (the developers forget to call the `initWallet` function), which lead to the $280 million being frozen in November[2]. Holy crap.

[2] https://news.ycombinator.com/item?id=15642856

Re: Miners Aren’t Friends

#67

Post by barbegal »

konschubert wrote: post #63
barbegal wrote: post #37

Earlier quoted context omitted.

Proof of stake doesn't work in practice unless we can find a source of randomness that is globally available, non-influencable and everyone can agree on [1]. At the moment, that source of randomness is completely elusive. It might be possible to create a random source from some form of binary astronomical event which can be widely verified but it would have an incredibly low bit rate of new randomness. One possible c…

> does this apply both to blockchain-style proof of stake and BFT - style proof of stake?

Both styles since both reward the creators of a block and both select the creators of a block through some form of randomness. So in both schemes there is an incentive to bias that randomness in your favour.

Re: Miners Aren’t Friends

#68

Post by ConnorLeet »

corford wrote: post #47
alexktz wrote: post #38

Earlier quoted context omitted.

£6 per card per day (1080ti) is quite appealing. 60p electricity costs. £48 per day minus £4.80 costs. Every day? For just leaving something switched on? It is quite appealing. I got my 1080ti's for £650 a pop. ROI is 108 days. That's why people, myself included, are doing it.

I still don't get it :( That's 108 days before you _start_ to make a profit and future returns are guaranteed to diminish as difficulty goes up. Versus £5,200 capital to invest immediately in a basket of safe (as much as crypto currencies can be), liquid coins that will likely x3 to x8 over a 12 month period. To each their own I guess but it's not the bet for me.

If the whole crypto ecosystem crashes (unlikely in my opinion) graphics cards can be resold and recoup losses.

Re: Miners Aren’t Friends

#69

Post by konschubert »

barbegal wrote: post #67
konschubert wrote: post #63

Earlier quoted context omitted.

> does this apply both to blockchain-style proof of stake and BFT - style proof of stake?

Both styles since both reward the creators of a block and both select the creators of a block through some form of randomness. So in both schemes there is an incentive to bias that randomness in your favour.

Doesn't BFT reward anyone that's validating a block? Meaning that everybody who validates the right block gets the same reward?

Or is there an extra reward for the creator?

Re: Miners Aren’t Friends

#70

Post by Bromskloss »

MrBingley wrote: post #66

Can someone explain to me how on earth Ethereum is still a thing? People have lost hundreds of millions of dollars in the Ethereum ecosystem, many due to basic flaws in the "Solidity" language. For example, last November $280 million was frozen in Parity wallets because function visibility defaults to external rather than internal (see the previous HN discussion here[0]). Just for fun, I was looking through the Secur…

Does Ethereum's existence hinge on Solidity? Isn't Solidity just a language that one can compile to the lower-level language actually built into Ethereum?
Post reply on HN