[margorczynski]

Open source is really interesting in the regard that for many projects it starts being optimized not by top-down design but by evolution. You get many variations on the same idea that live, die, merge into one or mutate step by step into something else. Similar thing with Linux distros.

And it shows the same upsides and downsides as evolution vs e.g. gradient descent - you avoid getting stuck in local optima as people will always try out new stuff just to create something interesting (in business without clear evidence this will bring money it's hard) but finding an optimal solution is slow.

It is much easier to attain certain goals by having a well managed team who report to you directly or indirectly and you can order them basically what to do and create a coherent vision for what they produce.