For a philosophy professor, I really don't get the thesis of this post. Is the point that engineers like to build stuff and most stuff they build won't be a successful startup? This seems extremely trivial and doesn't solely apply to internal tools (in fact, it applies to everything one builds). Is the argument that the signal is worse when it comes to internal tools rather than non-internal ones? Could be true (I dunno'), but there's no supporting evidence.

Genuinely confused.