Hey! New PI! Yeah, you! PhysioProf is talkin' to YOU!
If you act like an asshole to the trainees in your new lab, you dramatically decrease the likelihood that you will achieve a sustained upward trajectory for your research program. And without that, guess what? You got no more job.
Look, you are being thrown suddenly into a management position, when it is likely that you have essentially no relevant experience. (Supervising a technician when you were a post-doc is not relevant experience.) So what do you do?
You emulate what you think you are seeing around you. All the big 100-million-dollar corporate CEOs are admiringly described in the popular press as "tough bastards". And, closer to home, you hear stories about big-time PIs, even Nobel Prize winners, acting like arrogant jerks in their own labs. You hear about famous accomplished jerks all the time because they are rare and thus stand out. These people are probably exceedingly successful in spite of being assholes, and not as a consequence. In any intense professional pursuit, there a lot more successful people who treat others with compassion and respect than there are assholes. The former, however, don't make for salacious gossip or media attention.
Now that you are no longer in someone else's shadow, you are feeling free, feeling your oats, and figure you'll take an "uncompromising, tough" stance in your lab like you've heard about. This is a really, really stupid thing to do, to behave like an overt asshole. If you act like an asshole, you poison the atmosphere in your lab, science becomes no more fun, no one is motivated to produce, no one new wants to join your lab because post-docs and grad students gossip about you with other post-docs and students including ones at other institutions, and guess where your lab is headed.
The only effective way to motivate people for sustained long-term productivity is to create a rigorous, but humane and professional, environment in which people want to be productive, because they are excited by the science and eager to please and impress the PI and other lab members. But you also need to incorporate a very zealous system of unbridled critical analysis of data, hypotheses, and scientific conclusions that leaves no stone unturned in the hunt for unexamined assumptions and sloppy thinking.
This can be frightening for some trainees to feel like they are in an environment where at any moment of any day someone could come strolling over and start challenging them on their data, experimental design, etc. And some have trouble distinguishing between "That reasoning is completely fallacious" and "You are lazy and stupid". But it is key to running an efficient lab, avoiding or compensating for experimental mistakes and detours sooner rather than later. As a new PI, you should certainly be doing this with your trainees in public view of others in the lab, to lead and encourage by example.
So how do you get reticent trainees to relax and trust in this rigorous critical atmosphere, and not be too frightened to enjoy achieving mastery of some area of science? You make it clear through your leadership style that you trust your trainees as people, and that they should also trust each other as people. That trust allows trainees to feel safe enough that they can emotionally endure, and even come to spiritedly enjoy, hypercritical analytical discussions of their work and their scientific reasoning.
There are multiple contexts in which you can earn and foster this kind of trust in your lab. One of the simplest to implement, and one which has its own intrinsic benefits, is to adopt the attitude that so long as trainees are productive in the lab, whatever they're doing outside the lab is none of your damn business (yes, even blogging!), just like what you do outside the lab is none of their damn business. And what hours and days they work and what days they take off to be productive should also be of no concern to you. After all, your hours and days and vacations are of no concern to your department chair and dean, so long as you are productive.
It is an attractive feature of academia to be free to control your own time, and it is a good management move when dealing with highly creative people like your trainees to allow them this same freedom. It makes science more fun, and the more fun it is, the more motivated scientists are to GENERATE COOL DATA! Trainees who cannot handle this type of structure--such as "clock-punchers"--do not belong in an academic research laboratory, and should find a more suitable structured work environment.
This mutual respect for personal/professional boundaries creates a very good context for effective mentoring, because it provides a very stable safe interpersonal space for really getting down-and-dirty with the science itself in a completely unhindered way. You need for the people in your lab to feel free to be totally straight with you about everything, including, most importantly, data. This freedom of criticality enabled by a foundation of mutual personal respect is a great preventative of fraud, and a great encourager of good exciting vibrant science.
Now one clarification, just to be sure you don't misread me. It is very, very good to give trainees personal freedom and respect, as I described. It is also very good to give them scientific freedom and respect, appropriate to their stage of development, both for your benefit and for theirs. But they need guidance, they thirst for guidance. If you surf around post-doc blogs or talk to post-docs in your department, you will find that the single most common post-doc complaint about mentoring is that the PI does not give sufficient guidance.
It's never, "Dr. Pompous Pi is always coming over to my bench to discuss my experimental data, what it means, and where it points to next. Damn, that is so frigging annoying! I wish she would stop doing that." The single overwhelmingly most efficient task you perform as a new PI--in terms of ultimate benefit to your success as an independent scientist per unit time--is discussing data and experimental design and analysis with your trainees. Every minute you spend engaged with your trainees pays off many-fold in better data, more effectively controlled experiments, more novel ideas, more papers, more grants, promotion, tenure, your trainees themselves gaining independence, etc. If you leave your trainees without sufficient guidance, the vast majority will flounder and, ultimately, fail. And their failure is your failure.
Bottom line: Plenty of consistent ongoing rigorous critical guidance, coupled with a mutual respect for personal/professional boundaries and the freedom of trainees' time (for those that can handle it), make for a fecund scientific environment in the lab of the new PI.