4 ms·

Changing from 1 MB every ten minutes to 8 MB every ten minutes is nothing

treyd 3y ago
That's a huge change and putting it in domain specific terms obscures the implications. It changes the chain growth rate from up to ~52 gigabytes/yr to up to ~410 gigabytes/yr, it also means that the network takes 8 times as long to propagate blocks, which makes selfish mining attacks more viable, which weakens the game theory and adds a mining centralization force. Also consider that network bandwidth is a much more scarce resource for people in developing countries than it is for us in our first world country, and having a relatively-trustworthy payment system has a much greater relative utility for them than it would for us. And consider that full nodes don't only download blocks, they have to rebroadcast them to (ideally) at least 2 other nodes in order for blocks to even propagate at a decent rate. This is data that has to be kept by somebody forever, even if particular users run pruned nodes, and the best way to ensure that it's readily available for anyone to inspect is to limit how much of it there is. Since Bitcoin doesn't have state commitments, it's much less safe to use light clients, and increasing the bandwidth also shifts the balance more towards running light clients, compromising users' trust in the network. Making it more difficult to run fully verifying nodes like this is another centralization force. Also, you should read about induced demand. https://en.wikipedia.org/wiki/Induced_demand https://en.wikipedia.org/wiki/Induced_demand
- JCharante 3y ago
  1. People in developing countries aren't running nodes nor do they need to 2. Most people, even today, use light wallets. > Since Bitcoin doesn't have state commitments, it's much less safe to use light clients, and increasing the bandwidth also shifts the balance more towards running light clients, compromising users' trust in the network 3. Everybody except whales already uses light wallets. Even downloading the ledger in 2014 took a long long time. > And consider that full nodes don't only download blocks, they have to rebroadcast them to (ideally) at least 2 other nodes in order for blocks to even propagate at a decent rate. 4. Rebroadcasting 8MB is nothing > Making it more difficult to run fully verifying nodes like this is another centralization force. 5. Verification nodes aren't as important as miners and mining basically requires a lot of money put into ASICs so the system is already designed against the hobbyist enthusiast contributing to the network on their bedroom. > Also, you should read about induced demand. 6. What's the point of a system if it can't scale?
  - treyd 3y ago
    The core misunderstanding you have here is about the role of miners. The game theory that makes blockchains safe to use in a untrusted distributed fashion is that anyone can just look at the chain and be convinced it's correct under its own logic. If most users are just relying on third parties to tell them about the blockchain then they can't be sure of its correctness. And that gives power to miners to collude between themselves and service providers to just change the consensus rules as they see fit, including in ways that are detrimental to the broader user. And why shouldn't people in developing countries run full nodes? Are you saying it's fine for them to be dependent on service providers in rich countries that can then control/surveil their activity?

That's a huge change and putting it in domain specific terms obscures the implications. It changes the chain growth rate from up to ~52 gigabytes/yr to up to ~410 gigabytes/yr, it also means that the network takes 8 times as long to propagate blocks, which makes selfish mining attacks more viable, which weakens the game theory and adds a mining centralization force. Also consider that network bandwidth is a much more scarce resource for people in developing countries than it is for us in our first world country, and having a relatively-trustworthy payment system has a much greater relative utility for them than it would for us. And consider that full nodes don't only download blocks, they have to rebroadcast them to (ideally) at least 2 other nodes in order for blocks to even propagate at a decent rate. This is data that has to be kept by somebody forever, even if particular users run pruned nodes, and the best way to ensure that it's readily available for anyone to inspect is to limit how much of it there is. Since Bitcoin doesn't have state commitments, it's much less safe to use light clients, and increasing the bandwidth also shifts the balance more towards running light clients, compromising users' trust in the network. Making it more difficult to run fully verifying nodes like this is another centralization force. Also, you should read about induced demand. https://en.wikipedia.org/wiki/Induced_demand https://en.wikipedia.org/wiki/Induced_demand

JCharante 3y ago
1. People in developing countries aren't running nodes nor do they need to 2. Most people, even today, use light wallets. > Since Bitcoin doesn't have state commitments, it's much less safe to use light clients, and increasing the bandwidth also shifts the balance more towards running light clients, compromising users' trust in the network 3. Everybody except whales already uses light wallets. Even downloading the ledger in 2014 took a long long time. > And consider that full nodes don't only download blocks, they have to rebroadcast them to (ideally) at least 2 other nodes in order for blocks to even propagate at a decent rate. 4. Rebroadcasting 8MB is nothing > Making it more difficult to run fully verifying nodes like this is another centralization force. 5. Verification nodes aren't as important as miners and mining basically requires a lot of money put into ASICs so the system is already designed against the hobbyist enthusiast contributing to the network on their bedroom. > Also, you should read about induced demand. 6. What's the point of a system if it can't scale?
- treyd 3y ago
  The core misunderstanding you have here is about the role of miners. The game theory that makes blockchains safe to use in a untrusted distributed fashion is that anyone can just look at the chain and be convinced it's correct under its own logic. If most users are just relying on third parties to tell them about the blockchain then they can't be sure of its correctness. And that gives power to miners to collude between themselves and service providers to just change the consensus rules as they see fit, including in ways that are detrimental to the broader user. And why shouldn't people in developing countries run full nodes? Are you saying it's fine for them to be dependent on service providers in rich countries that can then control/surveil their activity?