We Killed the Open Internet Because You Couldn’t Stop Using xhost +

I remember 1996. The university had these Sun workstations we called “cpu servers.” At home, we ran Linux or Windows 3.1 with Exceed on a 486. Everything had public IPs. Cleartext telnet was fine. And we ran X11 apps from across the building without a second thought. It was glorious.

And then we broke it. Not because the technology failed — but because we lost trust.

We didn’t replace X11’s network transparency because it was broken. We replaced it because we broke trust.

Every modern remote desktop protocol — VNC, RDP, Wayland’s pipewire — is a monument to our failure. They exist because we couldn’t be bothered to secure the thing that already worked. Instead of building proper access control for X11, we just stopped using it. The industry decided that convenience was dangerous, so they threw the baby out with the display.

You know the pattern. A developer needs to show an X11 app on another machine. They Google it. They find xhost +. It works. They move on. That single command turns off all host-based access control. Every client from every network can connect to your display. No cookie. No password. Nothing.

I’ve seen production servers running xhost + because “it’s just a single-user machine.” On a box listening on TCP, that means any machine on the same network can read your keystrokes, capture your screen, inject events. It’s the digital equivalent of leaving your front door open with a sign that says ‘come in.’

And yes, the narrower xhost +local: still trusts every UID on the box. One compromised user, and your display is theirs.

So we abandoned X11 networking. We piled on SSH tunnels, VNC from behind NAT, RDP gateways. We traded a feature that worked beautifully for a mess of workarounds, because we couldn’t be trusted to configure it safely.

But here’s the thing: the problem isn’t X11. The problem is that we never built the right tool for the job. We needed something that gives you the raw power of network-transparent X11 without the hand-grenade safety of xhost +.

Enter FamilyWild.

FamilyWild is a small, opinionated tool that does exactly one thing: it lets you share your X11 display with specific remote hosts without opening the floodgates. It’s not a tunnel. It’s not a proxy. It’s a thin layer of access control that sits between your X server and the network, enforcing a whitelist of allowed clients. No wildcards. No “just make it work.” Just explicit, verifiable trust.

Think of it as the bouncer at the door of your display. You tell it who’s allowed in, and it kicks everyone else out. Simple. Effective. Exactly what we should have built 25 years ago.

I’ve been using FamilyWild for a few weeks now. I run a remote X11 app on my laptop from a Raspberry Pi in the lab. No SSH. No xhost. No hesitation. It works so well that I forgot I was even doing it — until I remembered the old days. And then I smiled.

This isn’t about nostalgia. It’s about reclaiming what we lost. The industry told us we couldn’t have nice things. FamilyWild proves we can — if we’re willing to build responsible access control instead of shrugging and saying “it’s too hard.”

We didn’t need to kill network-transparent X11. We needed to grow up and secure it.

Go read the original article. It’s short, it’s clear, and it will change how you think about remote displays. Then try FamilyWild. Because the open internet isn’t dead — it’s just waiting for us to trust each other again.