Nostr Web Client

I've been playing around with some numbers regarding the security of 12 word seed phrases in a hyperbitcoinized future. The Fugaku Supercomputer is 542 petaflops and could test ~322 billion different seeds per second.

If all eight billion humans have a 12 word seed phrase, it would take this computer ~67 million years on average to find a single one. Obviously the incentives are massively misaligned for anything like this to make sense. But what if Moore's Law continues as the same pace for the next 70 years as it has for the previous 70?

This malicious actor with a hypothetical future supercomputer would be able to find one of these eight billion (assuming little population growth) roughly every 17 hours.

There are MASSIVE assumptions made in this super rough math. But generally, I'm curious how the bitcoin protocol might need to adapt to prevent this kind of issue if Moore's Law does continue?

Reply to this note

Please Login to reply.

Discussion

Poppz 2y ago

Multisig, passphrase, 24 words fix this

Michael R. Sullivan 2y ago

24 makes it harder, but the same problem would occur father out into the future then.

I imagine all intelligent actors will continue to upgrade their custody strategies, but old, lost, or forgotten wallet would be vulnerable.

Poppz 2y ago

Would it make sense for that supercomputer to waste years just to crack a Bitcoin key ? What about the waste in capital allocation?

Michael R. Sullivan 2y ago

The incentives are wildly misaligned for them to do this any time in the foreseeable. But given a powerful enough computer, there are likely many honeypot lost or old wallets out there that actually could hold massive amounts of BTC. (Satoshi's hypothetical) wallets for example)

Poppz 2y ago

Silicon also has a max potential that we are close to reaching it , the exponential growth won't be at the same pace as it is now .

They say quantum is the next step , but I doubt it , and quantum computers are not good at simple functions but helping with complex functions, and gusseting a key is a simple function.

Michael R. Sullivan 2y ago

I've heard that as well about silicon.

This exercise has honestly made me much more appreciative of bitcoins resilience. I'm mostly just curious is someone much more knowledgable around bitcoin's future security than myself has thought or written about this problem.

Poppz 2y ago

Thank you for doing the math for us ! I think it falls on the same category as quantum computers , it is so far on the future it's not something to worry about that now , and would make more economic sense to use this supercomputer to help you mine Bitcoin than to try to crack keys .

Poppz 2y ago

But in reality, neither would make economic sense.

Michael R. Sullivan 2y ago

Don't trust, verify ❤️

I work as an engineer and in that world there is often an emphasis on not solving for hypothetical future problems years before they happen. I think you are right that things like quantum are nowhere near close to even looking like a problem.

zach 2y ago

Nice note and interesting thought experiment.

Seems that given enough time a future supercomputer will crack anything.

Michael R. Sullivan 2y ago

That's what I'm curious about. Anything increasing that exponentially is overpowered given enough time. It's a super hypothetical problem, just curious if anyone smarter than me has given it more thought.

nobody 2y ago

Appreciate this rough math though.

I did a fun one on calculating the rough approximation of Bitcoins *actual* inflation rate.

I think it was about 1.5-2.4%... depending on which valves and screw you wanted to turn.

I later read a much more detailed math breakdown and it was almost the same which was validating. Lol

Fun stuff.

Michael R. Sullivan 2y ago

Thank you! Doing a little bit of this math actual made me much more appreciative for the power of bitcoin's security. The numbers are completely wild and the incentives beautifully designed. Gotta do a bit of math myself just to make sure that I'm verifying, not trusting 😄

nobody 2y ago

Yes. To be fair the math we're doing is like... pre school compared to the actual Cyptography hahaha

Nik⚡ 2y ago

#[0]