I'm glad to see that several other commenters here are picking up on the fact that many PhDs don't actually make novel contributions, contrary to the suggestions of the article [0][1][2].
To add my own 2 cents, I think many people with PhDs believe they pushed the boundaries of knowledge, but many actually did not. As dahart indicated, the boundaries of knowledge look different to different people. Given that, one frequently "reinvents the wheel" because they think they're at the boundary, but they are not. There's this idea that science is incremental, but I don't think it is. It's more of a random walk. And given that much (if not most) research is conducted by relatively ignorant people (PhD students) who do not have much experience in their fields, we should expect people to be starting far from the boundary and "pushing" into areas already explored. The advisor is supposed to provide more knowledge, moving people closer to the boundary, but again, I don't see that happening as often as it should. Too frequently I see exploring previously seen territory while thinking it's unexplored or just accepting ideas about previously explored territory that are wrong.
Reinventing the wheel is becoming increasingly likely in my view as the number of published documents increases dramatically, but search technologies, strategies, and habits (most PhDs don't do a comprehensive search or use good search strategies) aren't keeping up.
While I do think I had some actually novel contributions during my PhD, I view myself as more of an "intellectual janitor", cleaning up the mess/technical debt that previous researchers left behind. Unfortunately this sort of work is perceived by many researchers as not research. My own PhD advisor was not a fan of this aspect of my work.
I tried to be aware of the literature far more than most PhD students. Many people seem to be offended when I point out that their "novel" idea isn't actually novel. In my view, novelty should not be required to get a PhD or publish a paper. One should never be offended that their ideas aren't novel. Disappointed, sure, but offended, no. The only way to have high confidence that an idea is novel/pushes the boundaries is to do a comprehensive search, and even then you can't be 100% sure that you're not missing something. Most PhDs spend far too little time on this aspect.
0xB31B1B got it right: A PhD is training. Because PhD students are so inexperienced, they should not be expected to produce novel research or even recognize what novel research looks like. I think the US should greatly decrease the amount of funding available for PhD positions while increasing the amount of funding for permanent research positions. Total researcher headcount would decrease, sure, but a smaller number of more experienced researchers with stable positions would do a better job overall in my view.
Very true. Regrettably, janitorial work is extremely hard to publish, as I had to find out. Mostly because reproducing other people's work, or comparing methods, is bound to find unfavorable conditions where a method does not work as well as published, or error measures that highlight common problematic behavior. At least in my area of research, such findings were unpopular with reviewers, to say the least. Probably not least because the reviewers tended to be the very people whose work was reproduced.
Which is sad, because this sort of work is well-suited to get PhD students acquainted with a field, and provides real scientific information.
To add my own 2 cents, I think many people with PhDs believe they pushed the boundaries of knowledge, but many actually did not. As dahart indicated, the boundaries of knowledge look different to different people. Given that, one frequently "reinvents the wheel" because they think they're at the boundary, but they are not. There's this idea that science is incremental, but I don't think it is. It's more of a random walk. And given that much (if not most) research is conducted by relatively ignorant people (PhD students) who do not have much experience in their fields, we should expect people to be starting far from the boundary and "pushing" into areas already explored. The advisor is supposed to provide more knowledge, moving people closer to the boundary, but again, I don't see that happening as often as it should. Too frequently I see exploring previously seen territory while thinking it's unexplored or just accepting ideas about previously explored territory that are wrong.
Reinventing the wheel is becoming increasingly likely in my view as the number of published documents increases dramatically, but search technologies, strategies, and habits (most PhDs don't do a comprehensive search or use good search strategies) aren't keeping up.
While I do think I had some actually novel contributions during my PhD, I view myself as more of an "intellectual janitor", cleaning up the mess/technical debt that previous researchers left behind. Unfortunately this sort of work is perceived by many researchers as not research. My own PhD advisor was not a fan of this aspect of my work.
I tried to be aware of the literature far more than most PhD students. Many people seem to be offended when I point out that their "novel" idea isn't actually novel. In my view, novelty should not be required to get a PhD or publish a paper. One should never be offended that their ideas aren't novel. Disappointed, sure, but offended, no. The only way to have high confidence that an idea is novel/pushes the boundaries is to do a comprehensive search, and even then you can't be 100% sure that you're not missing something. Most PhDs spend far too little time on this aspect.
0xB31B1B got it right: A PhD is training. Because PhD students are so inexperienced, they should not be expected to produce novel research or even recognize what novel research looks like. I think the US should greatly decrease the amount of funding available for PhD positions while increasing the amount of funding for permanent research positions. Total researcher headcount would decrease, sure, but a smaller number of more experienced researchers with stable positions would do a better job overall in my view.
[0] https://news.ycombinator.com/item?id=29139842
[1] https://news.ycombinator.com/item?id=29139567
[2] https://news.ycombinator.com/item?id=29139621