Best part in this is that this is without the upcoming persistent caching. With persistent caching it gets below 10 seconds for small changes on this particular application that has 70.000+ modules (70K)
β² ~/ πππ‘π πππππ --πππππππππ 6 minutes β 70 seconds (π».π·πΊπ‘!) for cold vercel-on-vercel prod builds next up: faster (incremental caching)