Exactly. IMO, this is a big problem in some areas of computational biology right now. Code to a famous project gets released, and instead of independently validating the results, other researchers (rationally) choose to just build on top of the famous project. It's faster, easier, and the "collaboration" leads to social prestige for everyone involved.

The end result is that empires are built, reproducibility is (paradoxically) harmed, and we're someday going to end up finding out that some big, high-profile projects were built on pillars of sand.