Let's say company X spends $14M to discover a much better production method, and starts using it to sell cheaper/better goods in the market. Big company Y decides to spend tens of millions to discover it themselves, and at great speed, hiring many scientists, and conducting lots of trials simultaneously. It looks like it'll take Y only 60 weeks to make the discovery, and 20 weeks to bring it to production.
Within that time, X realizes they won't be able to make $14M profit. So they decide to sell the discovery to Y for $12–14M on the condition that Y waits 40 weeks before launching the product. Y agrees, because they save time and money. During those 40 weeks, X makes $6M profit, bringing their total to $18–20M. The $4–6M is a more than excellent return for the time X spent discovering it.
Now that both X and Y have the discovery, they can either compete against each other, in which case all is good, or they can act as one and fix prices. In the latter case, big company Z decides to spend 10s of millions to discover it themselves and beat XY's pricing, in which case the story repeats.
The use of exact numbers makes it seem like this example is attempting to fool you somehow, but the only essential part is another company having the capability to rediscover the method. They won't even have to try, Y can just offer to buy it, with the implication that they will go all out trying to rediscover it unless X agrees to sell.
Thus no human effort is actually wasted discovering the same thing multiple times. Now things might not work out this way all the time, but in a free market, it will most of the time, and that's enough. Remember, the alternative we're trying to prevent is one company being able to charge whatever price they want for a potentially must-have discovery, for decades.
