Even assuming point 4 does not hold up, the GP is describing a system that promotes arbitrarily long development cycles. i.e. if it takes 5 years to develop the next generation of a drug, but the patent expires after 20 years, you are basically rewarding the whole industry to do no more than 25-30% of technically feasible R&D.