Every datapoint reduces the number of people it could belong to. CPU + browser + browser version + OS + major OS version can narrow it down by a lot.
Then add resolution, IP address location (which VPN they use is also a datapoint), which time they are active at, etc. and you can get a good almost-unique identifier.
Then add resolution, IP address location (which VPN they use is also a datapoint), which time they are active at, etc. and you can get a good almost-unique identifier.