Hacker News Reader

> The problem we now face, which is being made much worse by AI, is filtering things, as we get blasted by thousands of “relevent” search results

Especially with a malicious-trained model, there can just be so much content that seems about right that you can't really trust anything.

Kinda close to revisiting Usenet now without a killfile, specially on the sci.* and alt.asci* hierarchies.

What's really happening is that AI is amplifying a much more fundamental problem: Search Engine Optimization. We're exposed to so much information on a daily basis that we don't know what's real, and Google can't help us. As SEO makes Google searches progressively more useless, the web has become a morass of nonsensical data, blurring the line between fact and fiction. AI is now allowing us to generate even more information at an accelerating rate, essentially weaponizing disinformation. We have absolutely no way to deal with this, because we've invented the informational equivalent of gunpowder and decided it would be a good idea to give it to everyone.

This could be fixed to some degree if Google assigned less weight to old sites or authority domains. What happens is as soon as a site gains some traction in the rankings, this creates an incentive to produce tons of low-quality ad-filled content, even about stuff unrelated to the original theme of the site. This happened with rap genius, Investopedia, and many others. These ad-filled, low-quality pages fill up Google results. Ai generated content is just the latest iteration of the content farm.

Welcome to the Age of Bullshit