Thank you for making a case for using a beefy single server instead of overengineering everything for the serverless cloud from the start. I agree 100% that it is better to thoroughly understand the challenges of scalability, how it applies to your own situation, or if it applies at all, instead of betting everything on the latest trending technology du jour that might lock you in and make you regret it down the road.
A single strong database and web server can go a long way.
I've been also making all the mistakes you mention in terms of software development. I'm 100% OK with customizing the heck out of my dev environment though :-D
My personal obsession with overthinking problems and trying to find the best approach was part unhealthy perfectionism, but also part the way I grew up with PCs and how I learned writing programs for them. I know what it means to write code for an under-powered platform where optimization was a necessity in order to get usable performance. There was little RAM, little (or no) storage. Updating the screen could be very slow unless you knew how to write directly to video RAM. Preferably in machine language. That became pretty much a non-issue over the years, but when smartphones, iOS and Android became a thing I felt thrown back to that time. Once again we were dealing with relatively weak hardware and its shortcomings that can be worked around with some trickery. I wrote a couple of graphics processing apps and hit the limits of Dalvik and the ARM CPU pretty quickly. Fixed point integer computation instead of floating point was necessary. Native code close to the metal instead of Java interpreted in Dalvik was the solution.
These are not examples for premature optimization though. It was necessary.
These examples are more of something that might have shaped my mindset over the years to look for ways to optimize things before I even run into issues just yet.
This, and also what Some Dood mentioned in his comment: our abstract methodical thinking were we are conditioned into generalizing things and finding ways to create reusable code. I also know that it can be incredibly appealing for developers to create reusable code for the sake of being reusable, as if that library was the product and not the application using the library. This is a trap I ran into time and time again. That's basically when I was caught up in programming land, and didn't care much about end user land.
In itself aiming for generalization and reusability is a good and useful practice. But like anything in life, too much of a good thing can turn into a bad thing.
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.