>So if you're building something right now, maybe stop and ask yourself "ok, so how does this all look if we end up taking over the world 25 years from now, and this thing is a gatekeeper between a lot of people and a lot of revenue?"
This is a fair concern, but I'd say that if you're building something right now, I think is better to worry more about getting your startup in a position to take over the world first and then deal with those concerns. Worrying about these big problems when you're small I would say is a bit of premature optimization.
Same here, I was running around like a headless chicken and wondering what I was doing wrong when elasticbeanstalk just wouldn't deploy anything. The fact that we doubt ourselves rather than thinking aws is down is a good damn indicator how reliable they are, luckily it happened during daytime (EST anyways)