that's about the only use case I can think of for this approach, as well as where to find the few hundred people doing it.
that said, wearing walmart pants and hoodies, a face mask, and using a $200 laptop you got off of gumtree or kijiji or craigslist or FB marketplace at coffeeshop probably works just as well. bragging and laziness gets you busted, not a lack of local LLMs to check lexicon.
- a client (company I do contract work for) sees a different source address that is different than the source address I use for casual browsing+posting.
- two SaaS used for purposes of servicing agreement with "client" don't see the same source IP address as used for other clients.
- a bank I use, and PayPal, always sees the same source IP address dedicated to my VPN account only and for this purpose.
- the tunnel (VPN) provider I use for casual browsing+posting does not see the destination IP address of my client's VPN.
- whatever first-hop ISP I use sees one single Wireguard tunnel and nothing else ever.
- the first-hop Wireguard tunnel is paid for with a pre-paid debit card, but any outbound TOR traffic is encapsulated by a secondary tunnel paid for with crypto.
- the TOR circuit used for browsing purpose A is not also shared by browsing purpose B.
- any arbitrary outbound tunnel is specific to the container or VM I intended to use but doesn't carry, nor has any risk of carrying, any of my other traffic.
Tor is important to me because I have a right to read.There is no crime within Common Law for any of the above. Nor is there any violation of any statute for which any of the above is, per doctrine of minimum contact (such as with a pre-paid debit card), within jurisdiction of statute.
Perhaps some users do operate with some concern of being "busted", but most users that do outbound network path management do not operate with this concern.
Its like saying, not good enough for video editing 8k.
Qubes makes outbound network path management easy enough but it's not too hard to do on Linux and FreeBSD, so it can also be done on machines with modest compute resources (which may or may not be subject to the machine being an older machine) as well.