Although I have not been a postdoc (you couldn't pay me enough... ha, ha) I did get to play along for five years while Dr. S enjoyed a passive-aggressive, backstabbing, lying advisor who liked to take people down in front of others.
He and I both got, and followed, lots of advice before choosing our various labs. He talked to many people who knew this person (including his grad advisor, who had known her both personally and professionally for 30 years). None of them told him the truth, which was: Run, don't walk.
So here's my unsolicited advice for anyone choosing a new lab: Do all the things people say - talk to the oldest grad student, talk to the postdocs, discuss possible projects and degrees of autonomy. Then find the last 5-10 people who have left the lab and gone elsewhere, and call them up on the phone and ask them what they really think of the lab. "What were the five best things and the five worst things, and what advice would you give someone thinking of joining this lab?" is a nice neutral way to ask. People with permanent jobs already (who no longer need the advisor's goodwill) have the least to lose by telling the unvarnished truth. Also, if there's rumors that someone left in a fit of bitterness, especially talk to that person. Their reasons may or may not be good ones, but you should hear them.
And if you are a naive young grad student who thinks "But it will be different for me! I will work extra-special-hard/ graduate in four years/ totally be the teacher's pet/ never be as bitter as that!"... well, don't say I didn't warn you.