3 ms·

Yes, I hate the idea of future public lotteries, published ahead of time. The entire idea is to be resistant to a nation-state attack; lotteries seem pretty far

Yes, I hate the idea of future public lotteries, published ahead of time. The entire idea is to be resistant to a nation-state attack; lotteries seem pretty far down the chain. If you're in love with lotteries, you could do something like order all lotteries in the world in a published list, and then choose them based on the last digits of bitcoin block hashes over a certain period. I'm not sure what bitrate you could get through sunspots, but I agree we shouldn't worry about subversion from that angle.

baby 11y ago
It's based on lotteries from different countries all around the world. No way ONE state adversary could rig this. It's also set in the future, you need to commit to the time frame where you'll be doing that. I think their plan was to tweet it and sing it on soundcloud (among other things)
- api 11y ago
  > No way ONE state adversary could rig this. What about multiple state adversaries? It would be hard as hell but it could be done. I'd still opt for an astronomical source. If God wants to backdoor our crypto we're SOL, but other than that it would be pretty solid. I also don't get what's wrong with nothing up my sleeve constants. Hashing the word "YOLO" is mathematically unlikely to be a rigged constant to almost "probability of falling through the floor due to random quantum fluctuations" degrees.
  - baby 11y ago
    > What about multiple state adversaries? It would be hard as hell but it could be done. Highly unlikely, there is too much money in play, too many processes to corrupt... Cryptography is usually made solid enough so that there are easier ways to break the system. http://cryptoexperts.github.io/million-dollar-curve/#lotteries http://cryptoexperts.github.io/million-dollar-curve/#lotteri... > what's wrong with nothing up my sleeve constants It's not unpredictable. There are a million constants you could choose from if you wanted to make a backdoor. I think the badass paper is about that.
tlrobinson 11y ago
Why not just use Bitcoin blocks directly? A SHA-256 hash of 100 consecutive blocks should be sufficient.
- paulajohnson 11y ago
  I wondered about that, but Bitcoin mining is increasingly dominated by a small number of big players, and they also have the hardware to do exhaustive searches on hash codes to find inputs that give the answer they want. I'm not saying its necessarily a bad idea, but its not immediately obvious that this is going to be harder to subvert than a selection of national lotteries.

baby 11y ago

It's based on lotteries from different countries all around the world. No way ONE state adversary could rig this. It's also set in the future, you need to commit to the time frame where you'll be doing that. I think their plan was to tweet it and sing it on soundcloud (among other things)

api 11y ago
> No way ONE state adversary could rig this. What about multiple state adversaries? It would be hard as hell but it could be done. I'd still opt for an astronomical source. If God wants to backdoor our crypto we're SOL, but other than that it would be pretty solid. I also don't get what's wrong with nothing up my sleeve constants. Hashing the word "YOLO" is mathematically unlikely to be a rigged constant to almost "probability of falling through the floor due to random quantum fluctuations" degrees.
- baby 11y ago
  > What about multiple state adversaries? It would be hard as hell but it could be done. Highly unlikely, there is too much money in play, too many processes to corrupt... Cryptography is usually made solid enough so that there are easier ways to break the system. http://cryptoexperts.github.io/million-dollar-curve/#lotteries http://cryptoexperts.github.io/million-dollar-curve/#lotteri... > what's wrong with nothing up my sleeve constants It's not unpredictable. There are a million constants you could choose from if you wanted to make a backdoor. I think the badass paper is about that.

tlrobinson 11y ago

Why not just use Bitcoin blocks directly? A SHA-256 hash of 100 consecutive blocks should be sufficient.

paulajohnson 11y ago
I wondered about that, but Bitcoin mining is increasingly dominated by a small number of big players, and they also have the hardware to do exhaustive searches on hash codes to find inputs that give the answer they want. I'm not saying its necessarily a bad idea, but its not immediately obvious that this is going to be harder to subvert than a selection of national lotteries.