The idea that non-game software doesn’t involve creativity or spit balling or iteration is ridiculous. But from what I’ve seen it does involve a lot more waiting for consensus and thinking too far down the road, which are political activities aimed at being right (as measured by vice presidents) rather than productive activities aimed at getting something done or making something cool (as measured by your own name in credits of a completed work offered to the public).

I think the key difference is what the goal is. With non-game software, there's usually a goal of we want something that achieves X - let's create, spit-ball and iterate until we achieve that. X is a measurable outcome - it requires some creativity, spitballing and iteration, but it's easy to see if/when you've succeeded.

With games, things are a lot more subjective. The goal is create, spitball and iterate until you have something that people find enjoyable. You just keep going until you recognise that you've got something worthwhile. It's a "you'll know it when you see it" situation, rather than something you can track your progress towards. Sometimes you can follow a formula/template and iterate on another games' mechanics/systems and people will like it; sometimes you can do that and people will call it a soulless copycat instead. Sometimes games are technically good but just don't feel enjoyable; sometimes they're enjoyable despite any technical issues they might have.

Amazon and Google's issues stemmed from treating game development like any other software development.