I'm also a critic of certain Agile principles. First and foremost estimation, which do not see work.

However, you say "Determine what needs to be built. Then build it, however long it takes".

I think that the exact opposite should be the case. The best way to figure out what needs to be build is not via extensive planning and "predicting" but by writing as many proof-of-concepts as possible, see which ones get the most traction, then trow the rest away and make the one that's working well stable.