Why do random forests work? They are self-regularizing adaptive smoothers

Reading the abstract alone, I have no idea whether it's talking about algorithmic trees or, like, the big brown things with small green bits.

I had the exact same reaction: biology or computers?

The only hint I can see anywhere on the page is "Statistics > Machine Learning" above the abstract title.

I really want it to be about actual biological trees being studied on the scale of forests growing with smooth edges over long periods of time, but I suspect that's not what it is about.

Biological trees don’t make predictions. Second or third sentence contains the phrase “randomized tree ensembles not only make predictions.”