A friend of mine mentioned that Vitalik Buterin of Ethereum said that POC has a flaw and that someone can attack the network and cause a double spend. He said it was posted on Reddit. But I can't find it anywhere. Is this true? If so can someone post the link? ty
I read what he posted on https://www.cryptocoinsnews.com/cryptocurrency-burst-makes-smart-contracts-reality-happened-ethereum/ and normally I would say FUD but you have to look at who it is and think not but then again the article pokes butthurt at his invention so....
It would be nice for the BURST Dev to comment on this for the record.
Quote
Vitalik Buterin zash • a month ago
I'll copy my post from Reddit:
> Okay, I thought about POC for 15 minutes, and...
>
> wow, congrats, these guys have managed to create an algorithm which
simultaneously has a nothing at stake vulnerability AND is economically
inefficient.
> The problem is this. It seems as though the intent of the algorithm
is for the bulk of the cost to be storing the hard drive, and for the
hard drive to be able to "idle" after it scanned through the entire
plot. However, this means that once the plots are built, there is very
little cost to redoing the scan multiple times on multiple forks. Hence,
if an attacker could either (i) convince miners that its fork had a
>0.1%* chance of succeeding or (ii) bribe miners 0.001x the block
reward, miners all have the incentive to double-vote.
> * In order for the algo to be storage-bound and for a shortcut attack
involving recomputing everything not to exist, we need reading from the
hard drive to take less time than recomputing the data. But then we
want a 1000x safety margin if we want that condition to hold true
against potential ASIC implementations, hence reads need to be 1000x
cheaper than the plot computation step. Hence, reading more than one
time would have a marginal incremental cost of only 0.1%.
Things are this are why we didn't launch in Jan 2014
I'll copy my post from Reddit:
> Okay, I thought about POC for 15 minutes, and...
>
> wow, congrats, these guys have managed to create an algorithm which
simultaneously has a nothing at stake vulnerability AND is economically
inefficient.
> The problem is this. It seems as though the intent of the algorithm
is for the bulk of the cost to be storing the hard drive, and for the
hard drive to be able to "idle" after it scanned through the entire
plot. However, this means that once the plots are built, there is very
little cost to redoing the scan multiple times on multiple forks. Hence,
if an attacker could either (i) convince miners that its fork had a
>0.1%* chance of succeeding or (ii) bribe miners 0.001x the block
reward, miners all have the incentive to double-vote.
> * In order for the algo to be storage-bound and for a shortcut attack
involving recomputing everything not to exist, we need reading from the
hard drive to take less time than recomputing the data. But then we
want a 1000x safety margin if we want that condition to hold true
against potential ASIC implementations, hence reads need to be 1000x
cheaper than the plot computation step. Hence, reading more than one
time would have a marginal incremental cost of only 0.1%.
Things are this are why we didn't launch in Jan 2014
Same here. I think it would be nice to hear from the BURST dev.
Nothing at Stake(NaS) is a theoretical attack typically talked about for Proof of Stake(PoS) systems. The general idea stems from the fact that since PoS mining/staking requires negligible work, users can vote(mine) on as many chains as they want with their full voting/mining power, unlike with PoW where their mining power would have to be split to be used on multiple chains. Some people argue that this property weakens the system as smart miners should mine every chain instead of just the one they think is best, as it costs them nothing to mine on the extras and if the other one happens to win they stand to gain, or that someone wanting to attack a coin could pay miners to multi-vote. Users multi-voting this way would reduce the amount of hashpower required to do a 51% attack. As far as I know this has never caused problems to a coin, but it is an interesting and commonly discussed property. This applies to all PoS coins, and it also applies to Burst.
He is criticizing the fact Burst requires work to be done in it's mining, and NaS applies to it, saying it's taking an undesirable property each from PoW and PoS. He's not wrong and I've previously discussed NaS in this thread, however both of those undesirable properties do apply less then than they do in their normal implementations(the work while mining limits the amount of multi-voting you can do, and the work you do is far less than mining PoW coins). Personally I don't see this as a problem.
Not seriously though 
