logoalt Hacker News

fennecfoxy • today at 5:08 PM • 0 replies • view on HN

YES!

These were exactly my thoughts when they first added it and people were super annoyed. But even though they're not using beefy models for it (makes sense) they have tuned it enough to be right moooost of the time, especially for basic questions.

The biggest improvement that LLMs have given me so far in life is the democratisation of knowledge. No longer do I have to read some fake story about how some recipe came from someone's grandmother in Tuscany blah blah 10 paragraphs later I get the actual recipe. If I had wanted to read a cute story I would have chosen to. Same goes for websites littered with ads and other dark patterns, or even the ability to compare multiple sources in an instant to see what the common opinion is as well as (more importantly) use logic to discard or warn against aberrant information.

People mourn that humans will become dumber because of it; but I disagree. Smart humans will continue to be smart, our brains are plastic and I can just as easily forget how to do a low level search in the best way in favour of learning high level search skills and then re-learn those low level search skills again on a whim. There are people who don't have a strong ability to learn new things (or are simply apathetic) and those people existed before AI.