Probably, we'll end up with a new paradigm for web browsers that involves a lot of filtering on top of the morass of raw internet. Dunno exactly what it'll look like, but it'll probably involve multiple LLMs and some kind of abstraction on top of 'web pages'.
The end result with be a south park type of human centipede that few sane people will want to be a part of. I expect communities to become more segregated - as in real life where small communities isolate away from criminals - and more closed. It makes sense considering rampant theft, misinformation and ever increasing dubiousness that plagues the internet as a medium - in no small part thanks to procedural content generators that are schizoidly compared to human intelligence.
> even with massive amounts of noise it's still better than nothing
Nothing isn't the alternative. If the web literally vanished tomorrow, we still have other existing ways of communicating, researching, and recreating.
Is your friend running a phone app that's identifying the people around them and the topics they are interested in, all while earning a few dollars?
Where are you getting news from that's not just the firehose of falsehood? Do you really trust the people sending it to you, or are they profiting off of shoveling bullshit in to the feed?
For me, this is already about 80% true. The web is largely a vast wasteland at this point.
Obviously there ARE points to doing so, but it feels ultra-shitty that you're consigning yourself to having your work stolen if you're good enough.
In my experience, ChatGPT has a very distinctive writing style that's easy to identify, even if you "prompt engineer" to get something a little less generic. It also tends to not write anything of substance. The very few GPT-generated comments I've seen tend to get downvoted into oblivion.
That's already degrading the quality of all online discussions all by itself.
They weren't even bad comments. They fit in to the discussion, yet seeing it cheapened the experience.