Researcher argues academic incentives reward papers over correctness
The post opens with a line it attributes to Charlie Munger: "Show me the incentive and I will show you the outcome". The author's opinion is that the outcomes of academic research have drifted very far from its original goal, which the author takes to be the advancement of scientific understanding and the training of a new generation of researchers. Research, the author writes, was meant to break new ground, keep to honesty and good scientific conduct, be upfront about uncertainties, errors and mistakes, and pass those principles on.
The author says they have recently run into multiple cases where the correctness of results, the honesty of the people writing them, or the lack of curiosity once told the results are wrong has been unsatisfying. The central complaint is that once a paper is published and the promotions and fame have arrived, it no longer matters that the results are wrong, even if the authors know it. It becomes someone else's problem.
The main example: the author challenged the authors of a paper considered state-of-the-art, whose authors knew its evaluation was incorrect. One of them acknowledged the issues but refused to retract the paper, and asked instead whether it was bothering the author in publishing their own paper. The post says the right response would be to retract the incorrect claims or tell the community about the known serious issues.
The author then turns to PhD students, saying they see more and more who are less and less interested in correctness, precision and what the author considers proper scientific conduct, having learned from those successful in the field what does and does not matter. In one recent case, a PhD student had written a tool meant to perform well in certain contexts. When the author explained that the approach to the evaluation was incorrect, the student did not follow up at all. The author followed up later, demonstrating that the evaluation was not careful enough and that the tool is less useful than the paper's evaluation suggests. The author thinks the student was surprised by that follow-up.
The author contrasts this with their own past reactions. When someone told them their approach to strongly connected components discovery was wrong, they went home that day and immediately fixed the tool to use Tarjan's algorithm, feeling deep shame at having had no clue. When a reviewer of their dissertation challenged a graph, they were stressed for a week before working out that the reviewer had misread it, because the author had not labelled it clearly. And when they accidentally left out restarts as a major advancement in the history of SAT solvers during a presentation, an audience member rightly pointed it out, and the author felt embarrassed.
The post ends without a solution. The author writes that they are not sure how to fix the problem, that researchers do not seem incentivised to care about correctness, precision and scientific curiosity, and wonders whether the system has become so damaged, with so many PhD students having grown up to be professors in this environment, that much of what they wrote seems alien, even repulsive, to many. The author says this makes them sad.
Key facts
- The author argues that academic research outcomes have drifted far from the goal of advancing scientific understanding and training new researchers, opening with Charlie Munger's line about incentives.
- In one case, an author of a paper considered state-of-the-art acknowledged its evaluation issues but refused to retract it, asking whether it was bothering the post's author in publishing their own paper.
- A PhD student who wrote a tool did not follow up when told the evaluation was incorrect; the author later showed the evaluation was not careful enough and the tool is less useful than the paper suggests.
- The author contrasts this with their own past: fixing a strongly connected components approach to use Tarjan's algorithm the same day, and a week of stress over a graph in a dissertation review.
- The author offers no fix, writing that they are not sure how to fix the problem.
Why it matters
The post is a personal opinion about how academic careers reward publication, promotion and fame rather than correctness, and about what that teaches the next generation. Its sharpest claim is that a paper known by its authors to have an incorrect evaluation can remain in the literature and keep its state-of-the-art standing, because retracting it brings the authors nothing. The author frames the damage as passing to others who do not know the results are flawed. The story is a view from the inside, not a study.
Who it affects
The author points first at researchers who publish results they know to be partly wrong, and at PhD students who, in the author's account, learn from successful people in the field what matters and what does not. It also touches anyone who relies on published evaluations, since the author says a tool can be less useful than its paper suggests. The author also worries that many PhD students have since become professors, so the same habits are passed on to the next group.
How to use it
The post is not a tool or a method, and it gives no proposal. What it does offer is a standard the author holds up: when someone shows you a mistake, go and check it, fix it, and retract or flag known serious issues. The author's own examples are fixing the strongly connected components approach the same day it was challenged, and taking a reviewer's graph challenge seriously until the cause was found.
How solid is it
This is a first-person opinion essay, and the claims rest on the author's personal experience. No statistics or data are given on how common the behaviour is. The paper, its authors, the PhD student, the tool and the institutions involved are not named, and the specific incorrect evaluation in the state-of-the-art paper is not described. No response from the criticised authors or students beyond the reported paraphrase is included. The author's name is not given in the supplied text.
Risks and caveats
Read it as one researcher's account, not as evidence of a general trend. The author says they believe the correctness and honesty problems exist and calls the drift away from the original goal their opinion. The examples are unverifiable from the post itself because nothing identifies them. The author is also candid that they have no fix, and wonders whether their own views now seem alien to many in the field.
“Show me the incentive and I will show you the outcome”
— Charlie Munger, as cited by the post's author