Here are the top posts LessWrong showed me when I looked at it today.

This is the Recent display. It has tools to adjust what content you're seeing by up or downweighting tags. Here's mine. It's a pretty predictable set of interests.[1]

I have AI set to 0.01, the lowest the system will allow me to move something.
Currently I'm getting shown
Oh, plus the Best Of LessWrong banner is an AI post. ("Pay Risk Evaluators in Cash, Not Equity.")
This isn't an unusual range of topics to see when I show up on LessWrong. I haven't been taking screenshots like the above at regular intervals, but it feels like this is a typical spread.
Two years ago I said there was a lot of content about AI on LessWrong, I was mostly tuning it out with filtering tools[2] and I wanted some way to spend more time in the subsections I was interested in. I proposed some kind of subdomain setup to make the category boundaries stronger. I thought AI was smothering other topics.
That problem has become worse.
Today I swapped the AI tag to Hidden. That didn't work. I still see the curated posts ("I Did This Without Your Knowledge" and "GLM-5.3 Sometimes Considers How Train-Time Actions Shape Future Behavior") and I still see "Some Memories Of Earth Anthropic" and "Subtly Wrong Advice Misaligns Models More Than Obviously Wrong Advice" (which I guess aren't tagged at all?)
I don't know what to do to make LessWrong as a website useful for me anymore. Is there a way to filter out all untagged posts? Unless something changes, I think my best move is to just check my favourite authors from time to time and otherwise completely ignore the frontpage. If this is a tagging problem, I know I can propose new tags on a post do norms allow me to just go down my Recent page adding the AI tag to all of this stuff so it stops showing up for me? Can I pay the site ~$50 bucks to implement an auto-AI tagger? ($ 50 isn't worth the human dev time to them, but on the other hand it would be like one Claude request to implement a dumb email spam filter level system. Or maybe $50 would be enough to run every title and first three paragraphs through Haiku and ask "should this have the AI tag, Y/N?")
Is anyone else having this problem?
(World Modeling, Practical, and Personal Blog are on there with default weights because once in a while I want to shove those up or down if something's happening, but most of the time they're fine where they are.)
I read approximately everything Zvi M, Scott A, or Eliezer Y. publish on LessWrong or Substack plus miscellaneous things from elsewhere! I'm not putting my head in the sand and ignoring the entire subject here. But I want something other than AI talk when I come to LessWrong, I want the self improvement and the better decisionmaking and the social bonds and the meetup ideas. I'm here for the craft and the community in other words.
Proud to be the author of the only NOT AI post. Don't worry, my next post is about AI :)
I feel myself pulled in two different directions on this subject. On the one hand, I miss more of the classic rationality posts, or posts about rationality applied to various subjects.
On the other hand, if you really believe there's a serious chance you and everyone you love will be dead in the next ~5 years because of AI, it's inherently irrational to spend time thinking about the rationality of online dating, or whatever.
I think this community talks about AI nonstop, ... (read more)