Comment by MeetingsBrowser
2 hours ago
If the AI overview produces plausible sound answers, how do you know that it messes up very rarely?
If I search topics I am knowledge about, the overviews are almost always at least slightly wrong.
Not all websites are correct sources of information, but I am generally aware of which websites are trustworthy and can cross check.
When it comes to programming, it’s very easy to tell whether it’s right or wrong. Just run the code.
The small/fast model used for the overview isn’t smart, but it’s still pretty great at helping me find the function I need and syntax. Best of all, clearly and plainly, unlike many documentation pages.