Internet centralization and the original sin of NAT
Points and comments are a snapshot, not live.
Network Address Translation (NAT) broke the Internet's original peer-to-peer design and enabled centralization.
RFC 1631 (1994) proposed NAT as a short-term fix for IPv4 address depletion, mapping multiple private IPs to one public address. This breaks inbound connections, requiring workarounds like port forwarding, UPnP, STUN (hole punching), TURN (relaying), and ICE (trial of all methods). IPv6 was meant to fix this, but adoption stalls, and ISPs often apply NAT to IPv6. The author argues NAT normalized client-server thinking and made self-hosting difficult, as services like file sharing, email, or running a server now need complex configuration or external infrastructure.
What commenters are saying
Commenters broadly agree with the article's critique. One camp highlights that IPv6 adoption may finally eliminate NAT, though others note ISPs still firewall IPv6. Several discuss Tailscale as a practical VPN that traverses NAT, enabling secure remote access without open ports. A specific counterpoint argues that pre-NAT, running servers was easy but invited hacks; NAT's firewall-like properties reduced attacks, while IPv6 will require deliberate firewall rules. Another commenter defends Tailscale's security: listening on a private Tailscale IP exposes you only to Tailscale's network, not the entire Internet.