4 ms·

Every once in a while I bother not ignoring a Google AI overview, then I waste some time fact-checking it and find out it's wrong. Most recently about a python

Every once in a while I bother not ignoring a Google AI overview, then I waste some time fact-checking it and find out it's wrong. Most recently about a python library (where it hallucinated a function that doesn't exist, complete with documentation and usage examples) and breaking news (where it authoritatively said [non-culture war, non-controversial, local] thing doesn't happen, above a dog-bites-man story from a conventional news source about how thing happened again)

ajross 1y ago
> Every once in a while Pretty much what I said, no? You don't use the product and when you do, do it through a filter[1] where you only remember the bugs. Do you use other AI search products and find that they don't show this behavior? [1] I mean, come on: framing it as "bother not ignoring" is just a dead giveaway that you aren't doing a product evaluation in good faith!
- bcoates 1y ago
  I am doing the review in good faith though--by default, I scroll past to the first result, then if it seems unsuitable and I'm desperate enough I check the LLM thingy. If it were providing any value it would sometimes be both novel and correct. Usually it's non-novel (correctly-harmelessly but unhelpfully, restating the web search results). When it's novel it's because it's wrong. I would remember the situation where reading the LLM thingy added any value if it ever happened. The weird little UI thing they do where they only show the LLM result if you wait for it to render makes this very easy, I have to scroll up to even consider it.
- chneu 1y ago
  I think you're also showing some bias. People are now changing how they search and gather information to use AI. You're automatically discounting that person's experience because they avoid using AI for valid reasons. A lot of times Google AI is outright wrong, not even close to correct. It makes sense to not rely on it and only going back to see if it's improved. I don't "ignore AI". I just haven't changed how I search and gather information because it's clearly not accurate yet. I still have to fact check it which negates the benefits(time saving). Occasionally I go back to see if it's improved and usually it hasn't. That's not bias.
danielrico 1y ago
My experience with Gemini in AI studio mirrors what the AI overview shows. An hallucinated libraries and their internal reasoning dialogue reinforcing the hallucination and saying "the user don't know how to search on pipy".

ajross 1y ago

> Every once in a while Pretty much what I said, no? You don't use the product and when you do, do it through a filter[1] where you only remember the bugs. Do you use other AI search products and find that they don't show this behavior? [1] I mean, come on: framing it as "bother not ignoring" is just a dead giveaway that you aren't doing a product evaluation in good faith!

bcoates 1y ago
I am doing the review in good faith though--by default, I scroll past to the first result, then if it seems unsuitable and I'm desperate enough I check the LLM thingy. If it were providing any value it would sometimes be both novel and correct. Usually it's non-novel (correctly-harmelessly but unhelpfully, restating the web search results). When it's novel it's because it's wrong. I would remember the situation where reading the LLM thingy added any value if it ever happened. The weird little UI thing they do where they only show the LLM result if you wait for it to render makes this very easy, I have to scroll up to even consider it.
chneu 1y ago
I think you're also showing some bias. People are now changing how they search and gather information to use AI. You're automatically discounting that person's experience because they avoid using AI for valid reasons. A lot of times Google AI is outright wrong, not even close to correct. It makes sense to not rely on it and only going back to see if it's improved. I don't "ignore AI". I just haven't changed how I search and gather information because it's clearly not accurate yet. I still have to fact check it which negates the benefits(time saving). Occasionally I go back to see if it's improved and usually it hasn't. That's not bias.

danielrico 1y ago

My experience with Gemini in AI studio mirrors what the AI overview shows. An hallucinated libraries and their internal reasoning dialogue reinforcing the hallucination and saying "the user don't know how to search on pipy".