The “conventional view” of academia equates researcher’s worth with their publication record. The more “enlightened” among us don’t like that very much, but either accept it as the best out of bad options, or attempt to supplement it with additional metrics, e.g. related to teaching, open-source development, community engagement etc. Both of these views feel quite alien to me.

Don’t get me wrong - some of the most brilliant scientists I’ve met have a stellar publication record and a clear ability to turn rough ideas into scientific gold. However, in my experience, for every brilliant scientist who built many brilliant things and published many first-author Nature papers, there are ten equally brilliant scientists who built many brilliant things, but published next to nothing anywhere.

When I think of PhD projects in my own field - experimental quantum computing - I think the basic mechanics works something like this. The PI comes up with a rough project and gets it funded. Different projects vary across a number of important dimensions:

  • Max payout. If successful, some of the projects promise “a couple of nice PRAs”, while others have potential to turn into a “Nature factory”.
  • Difficulty. Projects range from “obvious next steps” to “extremely wacky”. Often the difficulty is uncorrelated - or even inversely correlated - with max payout. One way to look at project difficulty is as the number of major unexpected challenges the grad student will encounter on the way to success, which can range from “a couple” to “dozens and dozens” in a few-year project.
  • Resourcing. Depending on the grant and the funding agency, you might have to “make stuff” where others could just “buy stuff”.
  • Support. In some circumstances, a new PhD student may be surrounded by a supportive PI, two experienced postdocs, three staff engineers, and a state-of-the-art cleanroom. In others, they’ll be alone with a absent PI and “a machine that the previous grad student almost finished building as he graduated”.

All in all, for every smart and motivated PhD student on a feasible well-resourced high-payout project with good support network, there are ten on wacky under-resourced lonely ones.

Yet, those projects produce some of the best scientists. Their quality is not obvious for the CV at all, but fairly evident even from a short conversation. Everyone would rather buy a laser than build a laser during their PhD, but turns out that the latter gives your some pretty insane skillz. And the more of these annoying problems you had to solve by yourself, the more OP you become.

It’s also important to remember the soft skills. Research is hard on everyone, but there is a special kind of suffering associated with toiling away on a nearly impossible project, with nobody to tell you what to do. Projects like this filter for grit and perseverance - skills you definitely want your employees or coworkers to have.

A publication list is like a book cover. Some books come in fancy hardback with blurbs from ex-presidents. Others are less flashy, but just as interesting. But to really know if you like a book, you must open it, right?