IBM has elaborate incentive systems that encourage employees to patent every idea that comes into their heads, and few checks to weed out the bad ones. The plan, I suppose, is that it's hard to identify the important patents in advance, and it's worth it to spend millions a year on useless paperwork if every now and then it makes you 100x that.
Taking one of IBM's patents as revealing of their overall strategy is like watching a guy scratch his toe and concluding that he probably wants to amputate his leg. There's just no way to connect your observation to any evidence for (or against) the conclusion.
This is absolutely correct - at IBM, successful patent applications based on any idea you could come up with directly affected your reviews, ability to get promoted, and general perception of worker bees by the management chain.
Patents incentives were also roundly ridiculed by lower band and longterm employees at IBM. For individual contributor types, there was some expectation that patents were part of the job description.
I found the article rather sensationalistic, but one of the comments rings true with regard to globalization of knowledge work in general:
"...you end up with a barbell distribution – a bunch of those rare people who can work through systems they’ve never seen before and fix them at the top, and a bunch of people to act as interchangeable cogs at the bottom..."
If you cannot quickly add value to systems & technologies you've never seen before, you're probably a depreciating commodity. IBM is merely riding this wave.
Hopefully the 50+ year olds will manage to transfer the key piece of information to the 18-25 year olds - what they get paid and how to get that same salary.
Nathan Myrvold has a patent factory project that goes quite a bit further than this. There was an interesting business week article on it a while back:
Taking one of IBM's patents as revealing of their overall strategy is like watching a guy scratch his toe and concluding that he probably wants to amputate his leg. There's just no way to connect your observation to any evidence for (or against) the conclusion.