Actually, he is saying "what others find valuable is an important signal for research", and I don't see anything obviously wrong with taking into consideration

Actually, he is saying "what others find valuable is an important signal for research", and I don't see anything obviously wrong with taking into consideration what others find commercially valuable - because NSA is not a "fundamental sciences" research organization. I agree that NSF should pursue what scientists find interesting/important and disregard commercial concerns. NIH will often come ahead doing that as well. On the other hand, let's look at NSA - They develop technologies for spying on people. They write software & do a bunch of other applied work detailed in that PDF. I don't think they should run a business either, but NSAs primary aim is not science. If they go the extra mile beyond publishing a paper and want to license their code / tech to companies for a fee, I don't see why that's a problem. Often some exclusive access or license can spur further investment. I do think open access to all can & often also spurs innovation - but not always. Pharma for instance depends on a government-granted monopoly on the manufacture of a small molecule in the form of a patent, to ensure it can recoup its costs of R&D and indirectly bribing prescribing doctors & paying for ads.