Some frames are faster than others. I've ridden enough bikes to say that's absolutely true. A stiff frame accelerates faster and climbs better. However, once a frame is stiff enough so that I don't flex it noticeably, no matter what, I don't see how stiffer than that would be advantageous for me.

The very cool thing about carbon frames is that their stiffness or compliance can be modified in specific areas. When creating a new frame and/or layup, test riders can take out trial frames and see what they do, both in performance and subjective feel. The layup can be altered and test ridden again. Etc.

Racers ride what's fastest for them unless they have a sponsor contract, and sometimes even then they use rebadged or blacked out frames. There's obviously huge competition between frame manufacturers. If bamboo were faster, that's what would be prevalent in the stores.

Jan Heine has a theory that certain flexy frames are faster because the springyness contributes power to the pedal stroke. He calls it "planing." OTOH he seems to represent a tiny minority advocating flexy frames. I'd go with democracy and market forces on this one. People vote with their dollars. If people really think flexy is faster, there are plenty of old steel frames still out there. They could shock everyone and prove their point at their next road race.