Bitcoin Is Less Secure Than Most People Think | Hacker News Reader

Bitcoin Is Less Secure Than Most People Think

marginalrevolution.com

The title is clickbait trash, but the actual point in the blogpost is actually really good. So I feel like this deserves at least some discussion.

Proposed Alternative Title: "Proof of Work 51% attacks are far cheaper than most people think"

General point: When a BTC Miner creates a 51% attack, they get rewarded the block. Which means they get the 12.5 BTC (or whatever coin) associated with the blockchain.

In effect, the rewards of the 51% attack are not only the double-spends, but also every coin that is associated in the blockchain. Mining rewards are therefore, granted to the double-spend attacker.

To but it into more concrete terms. If a 51% attacker were to attack Bitcoin after 1-week (1008 blocks), then the 51% attacker not only gets to double-spend everything over the week... but the attacker ALSO gains at least 1009 blocks worth of reward from the blockchain (At $3500 / BTC, that's $3,531,500).

----------

Once you account for the BTC that is mined by the 51% attacker, the effective profit margin of a 51% attack is far easier than what people seem to expect.

The mining reward is a special transaction in the newly mined block called the coinbase transaction. It's a transaction with no inputs, only one output pointing to an address of the miner.

Bitcoin does not allow these kind of transactions to be used as input for another transaction unless a certain number of confirmations has passed.

Currently this is set to 100 (https://github.com/bitcoin/bitcoin/blob/1d9d314573ee48f6f511...)

So roughly 1000 minutes need to have passed before the network will accept a transaction with a mining reward as input.

Thanks for the additional technical details. But I don't think it changes anything.

> Bitcoin does not allow these kind of transactions to be used as input for another transaction unless a certain number of confirmations has passed.

This means that in the hypothetical attack case, where 1008 blocks are replaced by 1009 blocks by the 51% attacker... then 909 x 12.5 BTC are immediately usable. While the last 100 x 12.5 BTC will be usable in 1000 minutes later.

In either case, the 51% attacker mines blocks faster than literally everyone else combined. So there's no period where of waiting that is safe against a 51% attacker.

The 51% attack is the end-all-be-all of a coin. If it happens, the coin is hopelessly lost. The entire cryptographical nature of the coin depends on the 51% attack remaining infeasible.

----------

In any case, the 51% attacker's 1009 blocks is longer than the 1008 "honest blockchain". So the honest miners (by default) will mine the 1009-long blockchain after the 51% attack.

So the 51% attacker's coins are safe and will be spendable in short order. Perhaps the 49% honest miners can "blacklist" his coins, but with 51% of the hashrate, the 51% attacker can always legitimize his own coins by using his own 51% hardware (since he can build blocks faster than everyone else).

Isn't Nano far better in terms of "One CPU- One vote"? The transactions are voted on by nodes, which are free/cheap to run.

That's also a problem, though, because an attacker just needs a large number of nodes to effectively control the currency, which is not challenging or necessarily costly (see: Mirai). Doesn't stop the hardware race that leads to centralization.

If the incentives promoted by cryptocurrencies inevitably lead to centralization, what's their benefit over the existing financial system?

I can't debate any specific point in this article. It seems well written and researched to me.

Here's what doesn't add up to me. If it's actually possible why hasn't it happened? It seems like there is all the monetary incentive in the world to get it done. It feels to me like the fact that it hasn't happened supports the idea that it isn't economically viable under the current conditions

Reply on news.ycombinator.com