I disagree. SQLite is a 'database of first resort'. It doesn't scale to the same heights as server-based approaches, but those heights are quite a way off for the majority of systems and it's performance is good when you stay within it's scaling limits. The docs are also up-front about when you *don't* want to use it.
Using sqlite in a solution simplifies the solution compared to server-based RDBMSs (fewer moving parts and fewer deployment/management issues) and so reduces the costs and time spent dealing with it.
So unless you:
- already have existing server-based database infrastructure
- know in advance your expected load is approaching the limits of sqlite
- or particularly need a feature one of the server-based RDBMSs provides
I'd say that choosing MySQL or postgres over sqlite is premature optimisation.