Indie games on shoestring budgets are also the games that can least afford to pay employees to learn the "better" tool set on the job. Hiring devs that are experienced in Unreal or Unity means your onboarding is just about teaching them your studio's stuff, and the demands of your game. Budget is a zero sum game - if something like Expedition 33 (UE5) did it "right" instead of doing it "easy", they might not have been able to afford or produce the phenomenal mocap/VA/soundtrack/environments in the game.
Godot continues to mature, and some relatively big names in the indie space are publicly dumping Unity for it (like Mega Crit with Slay the Spire 2). But "pushing" smaller devs to ignore the onboarding problem isn't the way. It's the smaller devs that benefit most from engines with "good enough" defaults - bigger studios can afford to pay someone to "do the lighting".
Picking an engine (including the option of rolling your own shit) has to be a decision made very early in the game development cycle, like "before you hire anybody" early, and it's a really hard one to change your mind on later. For a lot of studios, the right decision isn't the "best, most capable, free-est" one. Hell, for Balatro the dev chose LOVE, which is usually used for VNs, because he didn't need all the other features he'd get out of something like Unity or Godot.
Do you not consider Expedition 33 a well-made game?
The comment I replied to says "we should push smaller devs to try engines like Godot now for that as Unity and UE got too big for their boots."