Agreed, if something is obviously silly or makes no sense no one will do it, and if someon does, you take on the economic loss of your time, labour and any money you put in.

But if there is something that has promise, try it on the ass option, thats what an MVP is, and if the MVP can work out its kinks fine. Testnet is obviously one way to go, but if some feature matures to a point where people want to try using real money or want open testing the ass-chain does make sense as a testing ground

Builders don't always know where something goes and crowd sourcing ideas only happens when people can F around with something and find out