Bitcoin mining power reaches 1000 petahashes per second | Hacker News Reader

Bitcoin mining power reaches 1000 petahashes per second

blockchain.info

12 pointsno_gravity14 comments

I really don't get why people are so hung up on these two issues -- "wasted" electricity and wasted computer power without actual numbers to put things into perspective.

Anybody has any real numbers on electricity used for bitcoin opperations vs. electricity used by banks?

From what I know about hashing algorithms and hardware, I'm pretty sure that the dedicated bitcoin hardware is not so capable for general computation stuff.

I'm more concearned about the pissed brain cycles that go to waste inside the big corporate oligarchies that we call banks.

All that electricity pissed away because people can't (won't) trust a banking authority.

If you think about it, there is a lot of energy and time pissed away because of trust issues. Reports, "cover your ass" emails and tasks, workflow limitations because of lack of trust, locks on doors, security systems, LOADS of bullshit jobs. Lack of trust shows in most aspects of our life. And, ironically, I think(without any concrete data to back up, it's just my own musing on the subject) that verification and "check systems" contribute greatly at the trust individuals show to each other.

But bitcoin miners have turned this into an energy-squandering engine with no upper limit. They've automated pissing away energy for imaginary points. That's not going to end well.

Don't you think there are market forces that will resolve this energy-squandering thing?

It sounds like an externality to me, so "no".

It's simple economics.

If you have a lottery with a single 25BTC prize, that can only be played by spending electricity to compute hashes, then expect 25BTC worth of cheap electricity to be spent (every 10 minutes that the lottery is held).

The increase to 1 exahash/sec is simply due to the latest generation ASICs having achieved 0.06 joules per gigahash.

At least this "waste" is reduced in half every 4 years.

All that compute power could have been put to so much use only if the inventor had thought of a better way to choose a nonce which works to create the hash.

What useful work do you imagine being possible with SHA256(SHA256(nonce+block)) by choosing the nonce differently?

Instead of having an arbitrary number as a nonce... let's say you chose the next prime number to create the hash. This itself will help us discover lots of prime numbers :)

It would also make bitcoin completely unusable, as the block time would be tied to the time it takes to find a new prime. So transactions would only go through every few years. So that would not work.

That was just an example of useful work. I am sure we could find something both useful and that is not time consuming.

I agree that this would be great, but this is not a new idea, people have been talking about these things since 2008-ish. Unless you have a firm proposal, I think we're at exactly the same place as we've always been: everyone agrees that it'd be nice, but noone has any idea how to make it happen[0].

The useful work also has to happen as a byproduct of the blockchain securing calculations, because if it's seperate, the blockchain will offer worse security properties than a competing blockchain with similar hashing power, so there's a disincentive to use it.

[0] There was PrimeCoin, which computed some series purported to be mathematically interesting. Math people I've talked to mostly shrugged their shoulders when I asked them whether that work had any real applications.

Hasn't Primecoin tried something similar and failed?

Reply on news.ycombinator.com