If you never fail, you need harder things to do.

On the other hand, if you never publicly fail you may just be prudent. Or paranoid. Or depriving yourself of some beneficial pressure. But it's not quite as clear-cut as you make it seem.

IMHO, ambition is not publicity. Setting high ambitions doesn't mean make big announcements like "make python 5x faster" without any evidence that the approach is promising. OTOH the pypy guys, for example, have kept a much lower profile but much higher ambitions, like write a completely new compiler toolchain with a novel architecture and then implement python on top of it. I wouldn't say they didn't have "harder things to do".

> I tend to prefer underannouncing and overdelivering...

Yes, assuming that you hear about the project at all. Overannouncement is lying, but it does get you more publicity.

"I tend to prefer underannouncing and overdelivering..."

An example of this would be Mike Pall and his LuaJIT project.