[Pxtl]

Right. I'm not saying that real identity should be mandatory throughout the internet, but that providers should be able to confirm real identity before allowing people to contribute content to their boards.

For reference, I'm rate-limited here on Hacker News because I'm politically opinionated and don't shy away from angry arguments. Would this be a better site if I went and created a new account every time Dang got annoyed with me?

Otherwise any open text entry box will get flooded by AIs that are completely indistinguishable from human content. Basically the existing problems with troll-farms and spam-bots and hackers will get magnified order of magnitude. Repeatedly, as the tech becomes better and cheaper.

[barrysteve]

Well your rate-limit comment raises an interesting point. My thought is that if you allowed back-and-forth conversations on public forums the comment sections would become uncomfortably long to get through, which already is kind of a doom-scrolling problem anyway.

Maybe public forums could hold the form of introductory areas that cause new conversations and relationships to shoot off into their own bubbles.

I'm an non-US empirical republican so most of my political views are bland to discuss in an American context. I do believe that more discussion builds better structures and reveals better ideas.

With human moderation, text input that carries meaning to the conversation can be allowed, regardless of it's machine-or-human source. In the big-picture sense, all information is generated by humans, so someone using ChatGPT to parrot old data into hackernews doesn't necessarily spell doom unless it's meaningless spam, like you say.

If we gave up "all" control of our computers an external web authority, you could validate if the user had copy&pasted from ChatGPT. But since we're in this odd middle-area of desiring no oversight on the PC, and high quality filtering on the web service, then you need humans to do the work of interpreting....