That's kind of the point of the article's argument, which is to say that if you compare your trading to an "index stock" then the index is likely to win, by a big margin.

So, good luck with ML, but make sure you reality test against the "set and forget" solution of indexes.