That is an entirely valid concern, and I see where you're coming from with that. It would be short-sighted to introduce something revolutionary, only to open the floodgates for everyone else to start implementing it two decades later.

I was thinking of using patents more along the lines of "throw shit at the wall and see what sticks." Rather than trying to come up with every conceivable method for tracking user attention and patenting those, the hypothetical patent troll would create and patent hundreds of different smaller, novel processes that may or may not be needed as part of a larger system for tracking user attention. The overall goal being to make it likely enough for one or more of those patents to be violated that a company would consider it too risky to go anywhere near the idea of commercializing attention tracking software/hardware.