2026-05-14
I have noticed an interesting phenomenon with LLMs. There is something I have never quite understood.
From what I have gathered from people talking about LLMs in comments, and blog posts, and chat rooms, and Youtube videos, etc. One of the most common uses for them is looking up information. And I don't get it.
I do try to avoid using LLMs where I can, but this particular case is not aversion or resistance for me. It is categorization. I simply that I don't associate them with finding information. When I want to look something up, I go to Wikipedia[*], or Youtube[*]. I go to a search engine and look for blogs and forums.
Wikipedia
Youtube
LLMs aren't even in the list of things that I think about for learning. I am a little mystified by how they got onto that list for so many other people. There are many tools for this task already. Those tools can certainly be improved on—who has anything good to say about search engines in 2026? But for me, LLMs do not seem tool-shaped.
I suspect this may be a bigger divide between me and LLM users than even moral or economic considerations.
What tools do you use to find information, to learn? What makes something look and feel like a tool for that?
Wikipedia is, of course, patterened after encyclopedias, the default reference tools of the print era. Youtube videos vary, but often register either as lectures like you might find in schools, or practical demonstrations. Blog posts are (we always hope) experienced people sharing the lessons they have learned, as if we were listening to a mentor or colleague.
Moreover, when I am looking something up with one of those resources, I am usually looking for something specific. In that directed, information-hunting mindset, a conversation or chat is simply the wrong format.
Learning can and does happen in conversation—but it is reactive. It begins when a particular person makes a statement, either written or spoken. And perhaps you would like them to clarify a point you did not fully understand. You would like them to elaborate on an idea you found interesting. You would like to challenge them on something you disagree with. And in this way the conversation starts and may proceed to new information. But this does not map onto LLM conversations. The bot has nothing more to say than you do.
And for my part, I rarely have anything to say to actual people. Going to a website to talk at a bot in hopes of something productive is extremely artificial and unlikely.
In conclusion, yes, I am still using search engines to find things. That is what they are for. Or are supposed to be for. Tools for finding information and knowledge and experience.
Blog
Reply by email