From the wikipedia article that was referred to:
Prior to the enactment of Bayh-Dole, the U.S. government had accumulated 28,000 patents, but fewer than 5% of those patents were commercially licensed
Not saying I understand the complexities but an outright repeal doesn't seem like it would address this underlying systemic problem. Maybe times have changed.
They haven't changed. Some government agencies excel at basic research (NIH, DOE, NASA, etc.), but for a multitude of reasons, they don't do product development. (And having worked at one of these agencies for several years, I think it would be a terrible mistake to change this - even though I personally have very mixed feelings about how Bayh-Dole has worked out.)
Some large fraction of patents are useless and will never be commercially licensed because it would never make any sense to pay a dime for them. A further fraction would demonstrate market failure if they were licensed - there's something much better that isn't patented, so if people are obliged to do the worse patented thing AND pay commercially for a license then we've made everything worse. So "more than 5%" might be a bad sign anyway.
This is what happened to audio codecs for years for example, several alternatives are superior to MPEG Audio Layer III (today you should use Opus), but the MP3 patent holders could expect their revenue to keep flowing for their inferior alternative anyway.
A certain kind of person really wants a patent, the same way someone might really want to win themselves a Hugo. I guess employer can think of it as a relatively expensive employee retention benefit, OK, we'll pay for your useless patent filing and we'll get you a laminated front page you can pin up in your cubicle. Go you.
Not saying I understand the complexities but an outright repeal doesn't seem like it would address this underlying systemic problem. Maybe times have changed.