4 ms·

To address a couple of concerns: - "running your own blockchain based messaging/mai server would be a nightmare because of spam". I think spam can actually be

To address a couple of concerns: - "running your own blockchain based messaging/mai server would be a nightmare because of spam". I think spam can actually be solved by using blockchains: given the low conversion rates in spamming, putting a small cost for sending a message can discourage spammers. Take this even further: put a variable cost to it - a higher cost for the first messages which would drop exponentially when the recipient engages in the discussion, taking it toward 0. I've never heard of this before, maybe I'll make an ICO to implement this. If you're doing it before me, I'll gladly accept some of your issued coins (my non blockchain email is in my profile) - "blockchain apps are not scalable". Indeed, given the current algorithms they're not scalable. But it's about web scale scalable. How many apps need to be web scale level scalable and what about the rest 99% of the others? Also when we'll have a functional implementation of lightning network this might be a solved problem.

runeks 9y ago
> Also when we'll have a functional implementation of lightning network this might be a solved problem. This is unrealistic. With 1MB blocks, around 2 million people can deposit/withdraw into payment channels once per week. Even using the best-performing/most-insecure layer 2 protocol ("deposit your bitcoins with me and I'll send them anywhere"), no more than 2M users would be able to join/leave the payment network per week. The Lightning Network is much more restrictive than this, since they aim for trustlessness, which further decreases performance (particularly for merchants, who want to receive large sums from a lot of people). Unless LN aims to create a separate cryptocurrency (where you can deposit into the network but not withdraw), there's no way their trustless model can perform better than a trustful one. And if they aim to create an altcurrency, then it's not a Bitcoin payment network, and there's little reason to do it on top of a blockchain in the first place. In that case it will be like a Bitcoin side-chain that you can't withdraw from, so you're basically converting your bitcoins into this other coin, without being able to go back. Fundamentally, the Lightning Network doesn't scale because it models global payments as anyone-to-anyone (like Bitcoin), when in fact payments flow from billions of consumers to millions of retail merchants to tens of millions of producers who pay out wages to their employees (consumers), after which the cycle repeats. Any payment network on top of Bitcoin needs to take this into account, in order to scale, since Bitcoin's throughput is so limited to begin with. As an example, imagine you're a merchant with a $100k monthly turnover. Using the Lightning Network you, as the merchant, would need to deposit $100k into a payment channel at the beginning of the month, in order to be able to receive this amount, and this is only if the LN is highly centralized (one hop between you and the payer). For every additional node between you and the payer, an additional $100k would need to be deposited, sitting in a payment channel ready to send. This just doesn't scale. LN is a decentralized network whose performance decreases for every node that joins it. For every additional hop your payment crosses, someone needs to deposit bitcoins into a payment channel, which both costs capital and increases the risk that this channel is exhausted and you need to touch the blockchain (expensive). A 10M-node LN would be able to send around a total of 2.1 bitcoins (remember, all nodes need to deposit bitcoins into a payment channel, and there will only ever be 21M bitcoins), and crossing 10 hops would require, on average, touching the blockchain every time 0.21 bitcoins are sent.

runeks 9y ago

> Also when we'll have a functional implementation of lightning network this might be a solved problem. This is unrealistic. With 1MB blocks, around 2 million people can deposit/withdraw into payment channels once per week. Even using the best-performing/most-insecure layer 2 protocol ("deposit your bitcoins with me and I'll send them anywhere"), no more than 2M users would be able to join/leave the payment network per week. The Lightning Network is much more restrictive than this, since they aim for trustlessness, which further decreases performance (particularly for merchants, who want to receive large sums from a lot of people). Unless LN aims to create a separate cryptocurrency (where you can deposit into the network but not withdraw), there's no way their trustless model can perform better than a trustful one. And if they aim to create an altcurrency, then it's not a Bitcoin payment network, and there's little reason to do it on top of a blockchain in the first place. In that case it will be like a Bitcoin side-chain that you can't withdraw from, so you're basically converting your bitcoins into this other coin, without being able to go back. Fundamentally, the Lightning Network doesn't scale because it models global payments as anyone-to-anyone (like Bitcoin), when in fact payments flow from billions of consumers to millions of retail merchants to tens of millions of producers who pay out wages to their employees (consumers), after which the cycle repeats. Any payment network on top of Bitcoin needs to take this into account, in order to scale, since Bitcoin's throughput is so limited to begin with. As an example, imagine you're a merchant with a $100k monthly turnover. Using the Lightning Network you, as the merchant, would need to deposit $100k into a payment channel at the beginning of the month, in order to be able to receive this amount, and this is only if the LN is highly centralized (one hop between you and the payer). For every additional node between you and the payer, an additional $100k would need to be deposited, sitting in a payment channel ready to send. This just doesn't scale. LN is a decentralized network whose performance decreases for every node that joins it. For every additional hop your payment crosses, someone needs to deposit bitcoins into a payment channel, which both costs capital and increases the risk that this channel is exhausted and you need to touch the blockchain (expensive). A 10M-node LN would be able to send around a total of 2.1 bitcoins (remember, all nodes need to deposit bitcoins into a payment channel, and there will only ever be 21M bitcoins), and crossing 10 hops would require, on average, touching the blockchain every time 0.21 bitcoins are sent.