Don't forget second-order effects. Your proposal encourages people to post quickly, to try to get in before the costs start kicking in. Encouraging people to crank posts out faster is probably not a net win.

I've noticed that trying to engineer communities isn't entirely unlike trying to secure software. You're not defending against straight-line attacks, you're defending against a hostile adversary with more time, resources, and collective brainpower than you, who will actively react to any move you make. The miracle is that any community ever works at all/any software is ever secured at all.