With that said, the amount of societal destruction caused by it, I do wish it hadn’t been invented.

I don’t disagree but I believe this is inevitable and unavoidable. We could never prevent pandora’s box to be opened.

I believe the main thing is the loss of trust that you are talking to actual human beings and that massive amounts of opinion and upvotes on social media are just manufactured. Economic power translates into supremacy of political opinion and influence.

But this already began before LLMs were good enough, from content farms producing slop articles to the russian web brigades and other state backed operations, to special interests like fossil fuel lobby, Trump’s rise to prominence during Obama’s term with the help of organized operations by Alt-Right and Gamergate that Steve Bannon weaponized, the simplified smartphones that allowed boomers easy access to social media. The political correct mandatory narrative about Ukraine.

Basically we had bots long before AI and the culture war was already lost.

I don’t know when or even if it happened, but discourse now is drastically different than it felt a decade earlier. You’re either with us or against us.

One stupid idea I have is “Only a good AI in a browser can stop a bad AI in the internet”. Not sure about the technical feasibility, but terniary models can pack 27b LLMs into a smartphone. Imagine we could actually filter out a lot of disinformation and brainworms. Or have an LLM parse all comments by a suspected bad faith actor and make an estimative judgement on it. Make the internet usable again lol. Maybe it can be used as a powerful tool for good. But fuckai cuts off any discussion about this. The immediate backlash to an AI model inside Firefox was near absolute.