The Genius Tax

Having spent considerable time in different labs in academia, I have noticed some interesting behaviors of people in power [I]My experience is mostly from labs that work in computational neuroscience or machine learning.. You have a principal investigator (PI) that is unquestionably really really smart, one of the best people in their field. They see things before others and have a drive to bring projects to their conclusion. However, some of them can be just awful to work with. For example, they can be dismissive of ideas they don't like, they are harsh in group meetings, and they can make you feel like you are not good enough.

And this system is still celebrated, even after all these years and clear indications about the harms this can bring to junior people. There is still this unspoken understanding that this is just how it works. You are learning from the best, so what did you expect?

For me personally, I expected them to be good at mentoring. Good at instilling confidence and curiosity. Good at bringing students up in a way that makes them want to continue doing science.

The argument for brilliance

Now obviously these people are still doing science and still publishing papers. Why? Because the argument for brilliance (seemingly) trumps all others. These people just deliver faster and better than others, because, for one, they can go through students much faster. And I admit this argument isn't crazy, and there are arguably cases in which one person's vision accelerated science by many years.

What's much harder to track and quantify though, is the amount of people leaving science due to "genius" PIs? Not because they were not good enough, but because they internalized a message of how science should work from a person who shouldn't share it. Many of these people who left academia were themselves capable of extraordinary work but we will never know what they would have done because they are not around anymore. In the same vein as a genius PI can accelerate the field by five years, the three students that leave their lab can represent a delay of ten.

There is a massive survivorship bias in how stories of science are being shared. We look at the people who made it through these labs and went on to do great science ("The system works. The harshness was formative") but this just looks at the survivors as the others are invisible to the narrative.

What to do about it

I had the chance to work with many talented researchers, and I can say that the best environments are intense, but not cruel. Intensity means high standards, honest feedback, and a rewarding intellectual challenge. Cruelty means making people feel small so the PI can feel big.

We can't just hope that toxic PIs suddenly become better people. We have to change the system's incentives. Here are some ideas:

• Tie tenure to human impact: Feedback from lab members should be formally evaluated by HR or a neutral committee, and it should directly affect tenure decisions. If a PI consistently breaks their team, they shouldn't get a lifetime appointment.
• Exit interviews: When a student leaves a lab or drops out of a program entirely, someone outside the department needs to ask why.
• Make mentorship a metric: Right now, we promote people based on publications, grant money, and h-indices. We need to add mentorship quality as a formal, heavily weighted criterion for hiring and promotion.

I am not arguing against high standards. I am arguing against the lazy idea that cruelty is a necessary byproduct of them. It isn't. The people who claim otherwise are mostly the ones being cruel, or the ones who survived it and now have a psychological need to believe it was necessary.

The best science I've seen comes from environments where people feel safe enough to take intellectual risks, challenged enough to grow, and respected enough to stay. And believe me, these PIs exist.[II]Just DM me if you want to know more.