If you make the same query twice, relatively close together, they give you back the same results.. this is intentional, since it won't consume another search request, and it allows you to return to your search if you didn't find the results you wanted on the first pass through...

They don't save the search results as history, this is simply a short-term cache.

Kagi is different from Google because they have a financial incentive to not serve you AI junk websites with optimized SEO. They're not targeted by SEO companies / optimizers.. and even if they were, unlike Google, you can simply ban the domains. If that domain pops up a lot because it's just an SEO optimized content stealing site, you never have to see it again.

I expect kagi will eventually care about the age of domains and the amount of people that are marking these as junk domains, as part of their search results.. because they have a financial incentive to make us happy, whereas Google has a financial incentive to make us click through multiple websites and revisit their search page, and make additional searches with different words, so that they can serve more advertisements