Hacker News Reader

A probably unacceptably large portion of the population DOES think they’re infallible, or at least close to it.

Totally. I get screenshots from my 79yo mother now that are the Gemini response to her search query.

Whatever that says is hard fact as she's concerned. And she's no dummy -- she just has no clue how these things work. Oh, and Google told her so.

That may be true, but the underlying problem is not that the LLMs are capable of accurately reporting information that is published in a single person's blog article. The underlying problem is that a portion of the population believes they are infallible.

They believe so because we have spent decades using the term AI for another category of symbolic methods (search-based chess engines, theorem provers, planners). In the areas where they were successful, these methods _were_ infallible (of course, compared to humans and modulo programming bugs).

Meanwhile, neural techniques have flown under the public consciousness radar until relatively recent times, when they had a huge explosion in popularity. But the term "AI" had retained that old aura of superhuman precision and correctness.

Reply on news.ycombinator.com