This is how I previously thought, but I've since reconsidered. The reason is that you can't split your attention between multiple projects without a loss in ove

This is how I previously thought, but I've since reconsidered. The reason is that you can't split your attention between multiple projects without a loss in overall efficiency. The amount of attention given to each of n projects isn't 1/n, but rather 1 / (n^x), where x is some measure of how bad you are at multitasking. For me, I think x is probably somewhere around 1.5, so if I have three projects, I end up giving each an average of 20% of my full attention. The remaining 40% is lost to the ether due to distraction, task-switching costs, and not having one project constantly be the main thing on my mind. As a result, increasing diversification results in a lower overall expected value of your projects, because you're losing more and more of your productivity to efficiency losses.