It still seems that my users on phones and browsers would need to trust me in that case, right?
So one example I’d give - every team I’ve worked on has had to build a local development environment with several nodes to easily spin up with a clean slate for deterministic testing. Teams get sucked into tools like Infura to set these up and then it’s so easy to do the same for deployment they do just that. I think there’s tons of room for Blockchain-as-a-Service tools to improve development and testing processes without forcing centralization on main net deployments.
Usually, you still have the “server is selectively lying” problem; unless the users are talking to each other, how do the consensus rules help with this?
(Related to the https://en.wikipedia.org/wiki/Byzantine_fault problem, though that's about forming consensus rather than determining trustworthiness.)
If you're submitting txs to a node that doesn't communicate to the mainnet (they're isolated from it) then any txs that go to it would be void. You could just use that Eth on the proper mainnet as it wouldn't be on the chain. If the node decided to then come onto the mainnet it's chain would be vetoed by the other nodes states and would fork back onto the main chain. Ethereal has Byzantine-Fault Tolerance up to 50% and you don't gain anything by running an isolated node to try trick people.
The big Blockchain-as-a-Services shut down - both IBM and Azure are gone.
As is hopefully obvious, I am totally naive here; my questions are genuine. Thanks!
It's way simpler to just connect to Wikipedia.org and download the pages you want to read instead of downloading the whole Wikipedia.org database and keeping it stored and updated on your devices. Same principle.
Leave infrastructure work for other people.
You mean something like AWS, but that allows me to quickly setup an server containing a node?