Hacker News Reader

You have to launch the bad thing first and as fast as possible; there's no way around but through.

I agree. launch and improve iteratively. Don’t waste time on perfection. Launch only core features and build from there

What's your process been for deciding core features?