Except we know as software engineers that solutions to problems at scale are never that simple. The "just increase the block size limit bro" solution doesn't consider the externalities that increasing it would have on the network performance and decentralization. This is the same line of reasoning about this whole thread, compromising decentralization compromises the ability to use a credibly peer-to-peer (cash) system.
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
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?
I ran a Bitcoin node on and off for 7 years. I remember the day it cost more to bootstrap an entire node than make a transaction. Your argument about "developing countries" and "internet costs" in other comments is completely disingenuous for this reason alone.
Bitcoin's runway was cut incredibly short by not raising the blocksize. The excitement around decentralization died because it was strangled.
> "We know as software engineers that solutions to problems at scale are never that simple"
Yes, L2s are needed - But it's insane to suggest 10x-ing the throughput at the bottleneck for a rounding error cost is a bad idea.
In what units?
