This is every bad idea about "how to fix the Internet" of the last 15 years, restated.<p>In fact, we don't have to change anything about TCP/IP to create fully authenticated Internet services. Without any changes to Internet hardware or programming languages, we can use public key cryptography to allow people to prove their identity to a third party and use that identity to access other services. In places where anonymity is problematic, those services can simply disallow unauthenticated users.<p>The reason we don't have services like this isn't technological, has little to do with money, and nothing to do with privacy concerns. The problem is that not enough people want this service. If they did, it would exist already.<p>Meanwhile, the idea that fundamental changes to Internet protocols would alleviate security problems is as old as IPSEC: it was literally the pitch Robert Stratton gave to assembled hackers at Summercon in 1995. The problem with the pitch is that Internet protocols and hardware have practically nothing to do with security; rather, the core security problem is that all programs have bugs, and bugs can be assembled into levers for unexpected behavior. Take it up with Edsgar Djikstra, not Vint Cerf.