I just started reading a piece in Nature Reviews Neuroscience entitled, "Choices in Neuroscience Careers", and by the second paragraph I was totally enraged. Some huge-ass PI dude named Tamas Bartfai had the following advice for grad students and post-docs concerning choosing an area of research and laboratory for their training:
Rather than choosing a 'hot topic', it is much more important to find a great mentor: somebody who has made a significant contribution to science, who has name recognition and whose laboratory attracts the brightest young people who will provide both real camaraderie and competition. It is clearly better to be in a laboratory with a well-known scientist with a depth of experience, who has little time but who provides a stimulating environment with many competing postdoctoral researchers who will be your peers during the decades to come, than it is to work in the laboratory of a young scientist who might themselves still be struggling and might even change their affiliation or direction during the time that you are training with them.
This is pernicious nonsense. Just as there are pluses and minuses to being in a big-ass lab filled with trainees where you hardly see the PI, but she is a big macher, there are pluses and minuses to being in a start-up lab with hardly any trainees where the PI is around all the time, but nobody has any clue who the fuck she is.
Big famous labs have the benefits of being, well, big and famous: lots of resources, name recognition of PI, and lots of other trainees to interact with. On the downside, trainees may flounder for months, or even years, going in fruitless directions, because the PI doesn't have time to pay attention to any one trainee, and doesn't have any incentive to give a flying fuck about the success of any one trainee. For such a PI, trainee success is a statistical issue, and so long as some non-zero percentage of trainees don't crash and burn, everything is copacetic. Also, there is probably intense competition among trainees for the interesting projects, and you may get stuck with some boring shit.
Small start-up labs have the detriments of being, well, small and start-up: there may be limited resources, the PI does not have power, and there may not be tons of other trainees to interact with. The plus side is, however, substantial.
You will have tons of time to interact with the PI, and the PI cares a fuckload about you and your success, because trainee success is not a statistical matter: every trainee in a start-up lab has to succeed. You will be working directly with the PI on an ongoing--even daily--basis in planning, interpreting, troubleshooting, and moving forward your project. You will have the pick of interesting projects. And you will have experienced what it takes to start and manage a new lab, invaluable when you start your own. Also, your success as a trainee in a young lab provides a lot more evidence to job search comittees of your future promise as an independent investigator than success in a big-ass lab that squeezes out a couple fucking Cell/Nature/Science papers every year with ten authors each.
I started my own post-doc in my mentor's lab the day she opened the doors as a totally green assistant professor. We worked together side-by-side on a daily basis designing, interpreting, troubleshooting, and anaylzing experiments, as well as doing both short-term and long-term planning of her research program. Can there be any question how fucking valuable that was as a learning/mentoring experience for me? And I saw both the smart and not-so-smart decisions she made--and how she made them--that determined the future course of her career.
I managed three first-author papers in 3.5 years as a post-doc with her, one in Cell/Science/Nature, and the other two in >7 impact-factor journals. I was successful in my own job search, and am now a junior faculty member at an extremely well-regarded private medical school. I have multiple R01s, and my lab is publishing papers at a nice clip. Sounds like a pretty fucking decent outcome, no?
Based on many discussions concerning my own faculty search, and in the context of faculty search committees I have served on, I can promise you that it impresses the shit out of people when a trainee has tremendous success in the laboratory of a junior faculty member. And talk to anyone who has ever served on a faculty search committee, and they will regale you with stories of two- or three-C/N/S-paper post-doc from big-ass laboratory who sucked total shit once they got out from under the protective umbrella of Professor Fucking Big-Ass.
Are there risks to joining a new lab? Of course. The PI could suck, she could get fired, etc, etc, etc. But there is a tremendous potential up-side, if you choose wisely, and hitch your star to that of a junior faculty member who is, herself, a rising star.
And not that trainees should give a flying fuck about this, but junior faculty do need to have good trainees in their labs so that they can, you know, succeed and get promoted and tenured and become a huge-ass swinging-dick macher like Bartfai, who goes on to tell trainees not to join the labs of his junior colleagues!
Note to Tamas (who I don't know at all): Way to simultaneously be totally full of shit and kick your junior faculty colleagues right in the fucking teeth!
(I know there is more patently false pernicious bullshit in Bartfai's comment I quoted than I have analyzed here, so I will leave further exegesis as an exercise for our readers. Go to it in the comments!)
UPDATE: Just to clarify, I agree completely with the first sentence of what I quoted, so long as "significant contribution" and "name recognition" are construed broadly enough to encompass a highly successful post-doc who has just started her own lab.