[runeks]

With proof-of-stake, an attacker compromising a majority of private keys results in the attacker taking control of the blockchain. The attacker could publish only empty blocks using this stake majority, and there'd be no way to distinguish these malicious valid blocks from actually useful blocks.

This makes proof-of-stake inherently less secure than proof-of-work because for PoW blockchains, consensus is not affected by compromising private keys.

[bawolff]

Instead, it can be attacked simply by purchasing a lot of compute.

For a smaller chain, purchasing a lot of compute sounds easier than hacking everyone's keys. Not to mention if you can hack someone's private key, surely you can hack their server farm

(Fwiw, i think both have a similar level of risk, its just coming from different threats)

[runeks]

> Instead, it can be attacked simply by purchasing a lot of compute.

Firstly, no one is offering for sale enough compute to attack e.g. Bitcoin.

Secondly, and more importantly, this would be a temporary attack, during which the blockchain in question is rendered unusable. Once an honest majority of hashing power is reestablished, everything is back to normal. But a compromise of a majority stake is a permanent compromise of a chain. No honest actor can conjure up more stake to gain an advantage over the attacker who now sits on a majority.

[sph]

> Instead, it can be attacked simply by purchasing a lot of compute.

It’s easier to accumulate a lot of money than to accumulate a lot of hardware and building a nuclear plant to literally outcompete the rest of the world, no? And at that point you are no longer invisible, so the rest of the world could simply decide to ignore your monopolistic contributions to the blockchain altogether and fork.

[everfree]

I think it would be much easier for e.g. state actors to take control of a few of the world's PoW mining locations than to compromise a bunch of people's private keys.

[proxynoproxy]

You don’t really get how mining pools work. It’s not a monolithic thing. well OK MARA is a proprietary pool, but the rest are made up of people contributing hash. If a state actor took over a pool, we could spin up another quick smart. Hash is fluid. Try that with slashing and stake lockups and others nonsense PoS invents to do a worse job.

[everfree]

Bitcoin is not currently mined by "people", it is mined by large specialized companies with razor-thin margins in warehouses with access to bulk ASIC purchase agreements and incredibly cheap electricity.

The last time a "person" was able to mine Bitcoin at breakeven was around 2017 or so. Maybe 2020 or 2022, if you threw the dice on buying ASICs from shady companies, hired an electrician to install a new circuit, then filled your home office with 10,000W of equipment.

In contrast to stake as an individual, you don't need special access to electricity or agreements or anything. Just buy tokens and run them on a mini-PC that costs less than a grand. The minimums are low due to DVT.

> Try that with slashing and stake lockups

I don't see how slashing is relevant to the conversation.

As for lockups, it's much easier to wait a few days for stake to unlock than to try to sell a bitcoin mining ASIC. It's also less counterparty risk and a much more liquid market.

[proxynoproxy]

https://bitaxe.org/ I guess all those plebs are just imagining mining sats at the edge. Mining isn’t just farms. And farms are not pools. You are regurgitating ETH propaganda that doesn’t match reality.

what you are missing is that PoS coins are printed from nothing. The original sin. From nothing. Worth nothing. PoS is a Ponzi to dump worthless tokens on you. You can’t escape that.

PoS trends to zero. I can’t believe we are still arguing this with >2T market cap for BTC and ETH has actually shrunk since it moved to PoS. It’s going in the wrong direction…

[everfree]

One bitaxe costs $145 and runs at about 400GH/s. That's $0.02 per day in gross revenue, before electricity expenditure. It breaks even at $0.05 per kWh, which is an electricity rate so cheap it's impossible to get most places in the world. Mining 20 sats per day - which are worth a tenth of a cent each - is completely meaningless.

Even if you could somehow get completely free electricity, you need to buy 1700 bitaxes at a cost of $240,000 to earn a modest $1k per month, or 5% annual return initially, which dwindles quickly as ASIC technology improves year-over-year. You never reach breakeven.

So the bitaxe is a novelty item. Economically viable mining is indeed just farms now.

> what you are missing is that PoS coins are printed from nothing.

Satoshi's coins were printed from running some lines of C he wrote on a desktop computer. That's just as "nothing".

> The original sin. From nothing. Worth nothing.

That's not how economics works. See my other reply about how software also comes "from nothing" but is obviously not worth nothing. All intellectual property comes "from nothing", but intellectual property powers the modern economy.

> PoS trends to zero.

Opinion.

> I can’t believe we are still arguing this with >2T market cap for BTC and ETH has actually shrunk since it moved to PoS. It’s going in the wrong direction…

Cherry-picked timeline.

[proxynoproxy]

Cherry-picked timeline… It’s reality dude! It’s the timeline that reflects reality in 2025.

Of course if you imagine a reality where ETH PoS was the best and the coin was worth $100k/ETH post merge you could say that. But that’s fantasy. It’s worth less per unit than when it was POW.

[runeks]

> [...] than to compromise a bunch of people's private keys.

No reason to target many people's private keys. Just find a few insecure exchanges — maybe a single exchange is enough, even — that together hold a majority of stake.

[everfree]

The largest single staking entity, Coinbase, currently holds 7.6% of the ETH stake.

[someone235]

Btw, this is one of the reasons they wanted to revert the DaoHack - they didn't want a malicious entity to hold 5% of the stake