what "black box" are you talking about? Caltech is a major research institution -- their invention was published and widely cited (according to Google, by 840+ papers), so others clearly benefitted from their work. It seems only major tech companies making billions by selling wireless gears didn't know about it.

Patent litigation is expensive but you would hardly need a billion-dollar war chest to monetize a portfolio with a successful litigation history. Between contingency fee arrangements and litigation financing, you could essentially print money with little downside risk, albeit with far less upside potential than if you paid sticker price to litigate.

"But that type of communication threatens patents..."

Patents, for all their downsides, are designed to get scientists to share results. Without patents you have enormous disincentive to sharing repeatable results: you can't charge money for it if anyone can go and copy you. With patents, you can freely share results and still charge money.

Should patents last as long as they do? Should there be a cap on how much companies can charge for patented medicine, since it's effectively a temporary government-granted monopoly on curing disease? All reasonable questions, but I don't think abolishing (non-software) patents would improve communication between scientists.