This makes 3 months project delivered in 1 month and in the end the client usually benefits that.
> with the argument that one day he could just scale
> it up even by rewriting it.
Which in practice of course never happens. Instead, the old code gets patched up until everything fails completely, and only then a rewrite happens :)
Pokemon Go is a bit of a singularity in growth. It's apparent to everyone on the street that (atleast, for the moment) this thing is huge.
When something is so obviously on fire with growth, over engineering might not be so bad. Look at any of the popular PoGo apps.. they have all dealt with big numbers.. the landscape is just a bit insane right now for that game.
I like making side projects, and i definitely agree with you not to over engineer.. but i feel like this is one case where it may not be the end of the world. With PoGo, if you don't balance the two (over engineering vs mvp) you'll likely end up sad, either way.
I also tend to disagree a bit with the article. For every situation like this were early scalability is important their are a 1000 MVP apps that are prematurely optimized or over engineered. At the end of the day the chance of anyone building an app that will get over 100,000+ in a week (and keep those users coming back) is very very very slim.