I don't see how that would be effective without fundamentally changing the structure of the Internet.
For example, I have access to your HN comment history. I could easily start my own blog with insights Ctrl+C Ctrl+P'd from HN'ers comments histories, and sign it as if it were my own.
Unless we ditch graphical user interfaces and the HTTP/S protocol and revert to 80s computing, with a command line interface for everything.
> Unless we ditch graphical user interfaces and the HTTP/S protocol and revert to 80s computing, with a command line interface for everything.
I think we should do precisely that.
SSL is all about explicitly trusting server names and implicitly trusting the data they serve. The whiz-bang UI's of the modern web are predicated on blindly running whatever code those servers give you. That's why we have all of these asinine trainings on how not to click the malicious link.
It's time we started explicitly trusting people and implicitly trusting the data that they sign and let which server we're talking to fade into an implementation detail. If we trust the data it served for other reasons, it doesn't matter if we trust the server. We can just ignore whatever malware showed up because it's not signed by someone we trust.
Besides, imagine what we could get done if we didn't have to stop to rebuild the UI for maximal engagement every few months. We could, I dunno, compete on merit.
For example, I have access to your HN comment history. I could easily start my own blog with insights Ctrl+C Ctrl+P'd from HN'ers comments histories, and sign it as if it were my own.
Unless we ditch graphical user interfaces and the HTTP/S protocol and revert to 80s computing, with a command line interface for everything.