I completely understand the point about Snowflake having been created for use in such scenarios.

Your comment raised a couple of interesting points though.

1. If governments are able to identify SSL-VPN traffic, then the VPN technology isn’t working as expected. That’s a failing from the VPN’s side in that case. One way in which I could clearly see such traffic being compromised is by logging the hops that traffic takes, and realising that everything is going through a single point, I.e. a VPN. But if one were to use a revolving VPN, that shouldn’t be a problem in theory.
2. My original question was more in the line of “what tech does snowflake use that distinguishes itself from SSL-VPNs in terms of masking traffic?”
3. Since you raised the point about snowflake connections being ephemeral, I’ll assume that snowflake connections are automatically rotated across available peers without the user having to set it up? However, just like with the rest of TOR, most governments can rent a bunch of cloud infrastructure and deal with deanonymising this part of the chain too.

And let’s be honest here: TOR isn’t exactly the most private network on the planet. It’s well known that TOR devs collaborate with the 5 eyes and have backdoors built in, alongside the American agencies having access to a lot of the traffic on TOR to be able to mathematically deduce origin and destination of traffic including up to the point of clearnet IP addresses.