the point is that if some screwdriver is objectively better (doesn't snap and takes your eye off when you make a mistake) then the decision to use it shouldn't be affected by who recommends it.
Better for what? That is a huge and broad question to begin with.
For example, let's say Rust is the best language in the world, provably safe.
1. it is in practice? You will use unsafe somewhere: C interfaces, for example, or some abstractions.
2. even if it is still theoretically safer: can I finish my work with it? Ecosystem, libraries, etc.
3. do I have trained people for it?
4. what is the cost of writing software in it compared to other languages?
5. is my system really, really safety-critical? What are the consequences of my program not working well?
If it was really that easy, all of us would code proofs with something like this: https://dafny.org/ and would use languages like Rust.
That is not the case at all and there are a ton of variables, from which cost, and I mean short vs long term cost where short term or middle term dominates for which it makes just more sense to choose the "worse" tool. Why? Because the "worse" tool will make that piece of software exist. The "better" one will probably make it never exist, because of other costs.
Whoever ignores this when doing anything, from coding to other activities, then that's ignoring reality. If human lives are at stake, yes, increase the cost a lot, make it the safest you can, but, even in that case, probably achieving 100% is impossible, or maybe achieving 99.999999% is 10 times cheaper than achieving 99.99999999999999999%.
By this measure, we would not have a lot of the imperfect inventions that improved over time in many areas.
