Many colleagues are struggling with The Hiring Question.
For junior folks, The Hiring Question is "what should I do to get a job in the LLM-era." They worry whether they should get in on the roboticized proofs gold rush, and how this interacts with the fear of getting scooped. There is much uncertainty about how journals or hiring committees would value their work, if the committee knew that the applicant didn't come up with their own proofs.
For senior folks, The Hiring Question is the opposite side of this coin. They are unsure how to define research excellence for their institution, leading to uncertainty about how to evaluate candidates. Advisors are also unsure how to help their students.
I am here to tell you, that Europe has solved all these problems. The next meta[1] has already been invented, and once American institutions catch up to this strategy then the career game theory will work itself out to a comfortable place for everyone.
== Three Papers Every respectable institution knows that bibliometrics are cringe. Research assesment cannot be based on counting somebody's number of papers, nor can it be based on seeing which journal or conference those papers appeared in. This is a broadly supported norm, even if some people still struggle. When I have been on hiring committees, there was always some laggard still counting papers and looking at CORE rankings. But when you remind them of the institution's commitment to DORA, these people are quickly embarrassed into silence.
What do we do instead? Every job applicant is asked to attach 3 papers to their application. Often these will be their 3 most prominent papers, though they can also select a subset that best conveys their breadth of work. In any case, these will be 3 papers that the applicant is proud of. These 3 papers are in addition to a CV (including publication list) and research statement.
The committee will look at the attached papers to evaluate the quality of the candidate's past work, and look at the research statement to estimate the quality of their future work. The role of the full publication list is merely to support the case that the candidate has the required experience to credibly execute on their promised future plans.
== The Meta for Institutions
This phenomenon, of a limited number of papers considered for evaluation, is widespread but not universal. In the UK this principle is part of the national evaluation of universities. Both the German granting agency DFG and the Dutch NWO evaluate a candidate's track record based on the few publications they may attach (Europe's ERC counts all the way up to 10). Every hiring committee I have been on has followed the same 3 paper format, but only 1 or 2 jobs that I have applied to did.
We see that the '3 papers only' principle is nowhere near universal. But it is common enough that it shapes your career strategy as an early career researcher. If, as an applicant, you want to have a shot at getting these jobs or grants, then all you care about is your top 3 papers. That is because papers 4 up through infinity do not count.
For institutions, following this evaluation meta allows you to hire responsible researchers at the top of their field. Those backwards institutions who still count papers and venues become less attractive places to work, because nobody wants to live under the leadership of bean counters. Such institutions get a bad reputation and are left to hire the salami-slicers among the talent pool.
== The Meta for Applicants
If you have only 3 papers to show off your skills as a researcher, it would be bad strategy to use 1 of those slots on your ChatGPT weekend project. A paper like that neither demonstrates your capabilities, nor does it illustrate what you spend your time on.
A good paper, one which increases your chances of getting hired, is one that represents a large fraction of your research time, reflects your standard of work, and conveys which subjects you consider interesting enough to spend effort on. In your research statement, you describe what type of work you have done, what you value, and what you plan to do. Your three papers you choose to best support your story.
It is up to the hiring committee to then look at everyone's stories, including the implicit or explicit messages about LLM use, and select applicants based on the committee's priorities. That is where pluralism reigns: there are no priorities that all scientists or all committees would agree. As Feyerabend would say, 'anything goes.' Focus on presenting yourself in a way that you are proud of, and by applying to jobs you will eventually find employment and colleagues that mesh with how you approach mathematical work.
== Why is This Good?
In the age of robots, the European meta is more desireable than ever before. The game theory for junior researchers is clear: any time you spend effort on a new paper, it should at least beat your 3rd best ever paper in quality. This encourages slow science and careful work, discouraging minimum-publishable-units and other degrading practices.
Under the three-paper meta there is no incentive to post flag-staking AI slop, even when you fear you may get scooped if you wait. You cannot attach slop to your application, because the hiring committee will actually look at your attached papers.
The three-paper meta has been favored for years already at all respectable institutions. In today's mathematically uncertain times these hiring strategies are more desireable than ever before.
== Conclusion
The three-paper meta already exists, it is prevalent enough that young researchers can build their careers on this. I personally did so. I publish ~1 paper per year and have never worried about pressure to publish more. That strategy has let me build a body of work that I am proud of, and has also led to stable employment. The same has worked for many of my friends and colleagues.
If we wish to reduce the phenomenon of AI slop flag planting, the main thing we need is for the alternative strategy (publishing lots of low-effort papers) to further lose out for young people. Hence, it is high time for bean-counters to step back and let their deparments enter the 21st century.
| [1] | In video game culture, the meta refers to the current dominant strategy for playing. Partly the meta is the current Nash equilibirium, partly it is the current state-of-the-art knowledge of optimal play. A game's meta can evolve based on outside circumstances (rules changes, new content being released) or inside circumstances (new strategies are found and disseminated, new counterplay is found). |