> This is premature optimization at its finest.
Not really, this is an example (of many) where that quote is taken out of context and does not apply.
The premature optimization quote is about spending tons of time fine-tuning functions for optimal performance, shaving a millisecond here or there.
The choice of languages and/or frameworks to allow the product to reach your intended goals is a critical early architectural choice. One that is costly to revert later.
IOW, if you're building a niche SaaS app that you don't intend (or want) to market to a larger customer base then it makes sense to use a slow language if it does what you want conveniently.
OTOH, if your business plan is to scale to millions/billions of users as soon as you can get there, it is a mistake to paint yourself into a corner where you'll have to rewrite everything soon in the growth curve.
Now, perhaps, in the age of AI this matters less if you can use AI to rewrite everything when you get to that point. Although I haven't gone through an exercise of using AI to rewrite the whole company codebase at once, not sure yet how well that'll work.
The percentage of businesses that need to scale to even millions of users is miniscule.
> Although I haven't gone through an exercise of using AI to rewrite the whole company codebase at once, not sure yet how well that'll work.
Right, have you gone through exercise of having AI write an app in a language you don't know (inventing a whole web framework along the way)? Does that sound reasonable to you?