It's different when you invalidate your own research.