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.

So, I searched a bit and this github issue just give bad vibes: https://github.com/rustdesk/rustdesk/discussions/6444

This is also discussed on hn: https://news.ycombinator.com/item?id=39256493

I would also like to know. They give some sketch vibes, unfortunately (even compared to some alternatives), because otherwise the offering looks pretty good!