Post

Is the reason to avoid AI summaries because of use of one's info / security / privacy? Or because of the real possibility of falsehoods / hallucinations in the responses? Not entirely clear to me from the article.

In my case I've found the LLM generated summaries are starting to get useful, and more recently I've converged on the approach of adjusting the query to help the LLM-based summarizer pick up on what I'm really looking for. Though my other rule is that the LLM must give references or it is ignored.

Switch engines. Fuck google

Just switch to Brave🦁

And often wrong

DuckDuckGo!

There's also this filter list for uBlock Origin: bsky.app/profile/stev... by @stevoisiak.bsky.social which blocks Google's AI summaries

perplexity.ai does a much better job than today's google (even aside from the obtrusive AI and sponsored results) for searches of relevant data and cites the orgs and academic papers and news sites it cites for the fact finding. No sponsored results.