If any of this was done to combat viewbotting, then any disruption to token calculation would prevent views from being registered - not videos from being downloaded.

From my perspective both problems are effectively the same. I want to count unique users by checking for asset downloads and correlating unique session IDs. People can request the static assets directly, leading to view booting and waste of egress bandwidth.

The solution: have clients prove they are a legitimate client by running some computationally intensive JS that interacts with DOM APIs, etc. (which is not in any way unique to big tech, see Anubis/CreepJS etc.)

The impact on the hobbyist use case is, to them, just collateral damage.