Is the argument that entire categories of software should not be used because they could be backdoored? That'd seemingly rule out messenging apps, as well as th

Is the argument that entire categories of software should not be used because they could be backdoored? That'd seemingly rule out messenging apps, as well as the examples you mentioned (VPNs, remote viewing tools) I think a more reasonable approach is to understand your threat model and act accordingly. If you need to remotely access a PC and are concerned about TLAs, maybe audit the source code directly. Or like another commenter noted, do direct IP connection between two devices on a Tailscale network.