It's an exaggeration of both the effectiveness of web searches (which has gone downhill to the point of unusable in many cases) and the frequency of LLM hallucination. If you've had bad experiences with them, don't use them, but clearly millions of people find them effective - ThreadSky

About ThreadSky

chrisgrace.com • 3 days ago

It's an exaggeration of both the effectiveness of web searches (which has gone downhill to the point of unusable in many cases) and the frequency of LLM hallucination. If you've had bad experiences with them, don't use them, but clearly millions of people find them effective

Comments

pradlin.bsky.social•3 days ago

Thank you! I concede exaggeration of the Generative AI example. I had taken artistic liberties with my recollection of this post:

https://bsky.app/profile/jasonschreier.bsky.social/post/3leezrzlvrk2m

chrisgrace.com•3 days ago

this is why I find these comparisons disingenuous. Here's ChatGPT's answer, compared to Google search's answer. The third result (buried under other links after you scroll down a page) is a fandom where someone pitched the sequel, so this is one of the results you'd get from the query without AI.

chrisgrace.com•3 days ago

correction, the fake info is the SECOND result in the Google search (again, not even visible on the first page of results because of all the irrelevant links). I don't see how you could argue the product Google is providing is better than ChatGPT.

chrisgrace.com•3 days ago

Given this is a comparison, ChatGPT will win as a product easily (also note that the direct sources are linked from the chat response so you can click through and verify the sources it is composing its answer from).

Posting Rules

Be respectful to others
No spam or self-promotion
Stay on topic
Follow Bluesky's terms of service

Comments

Posting Rules

Reply