This is precisely the point that the threat model bares its fangs. You can ignore it, but you should be aware that you're putting all your faith in that service.
A hypothetical Good VPN doesn't exist in China, for example, because they're legally not allowed to do what you suggest. Many of us don't live in China, but some do. Even outside of China, is it really true that a VPN service will simply give LEO the finger when they ask "Who was downloading child porn off your servers?" I'm skeptical they can.
That's the thing though, not all legal requests get the same weight or priority. I wouldn't trust a VPN to not roll over for your example cp case, but I think a middle finger equivalent isn't unreasonable for the less severe and more common case of receiving a complaint that the IP was observed as part of a swarm seeding copyrighted material. For lesser things, where does DO sit on the spectrum of will rat you out (which ISPs do), and likely won't rat you out (basically any paid VPN because their reputation depends on not doing so)?
This is precisely the point that the threat model bares its fangs. You can ignore it, but you should be aware that you're putting all your faith in that service.
A hypothetical Good VPN doesn't exist in China, for example, because they're legally not allowed to do what you suggest. Many of us don't live in China, but some do. Even outside of China, is it really true that a VPN service will simply give LEO the finger when they ask "Who was downloading child porn off your servers?" I'm skeptical they can.