Often times it does not make sense to serve free users who are often causing more hassle than it's worth. Having a free trial and then either converting users or not makes sense, but serving a user who uses it for a few queries every month ("I check every few months and only do a few searches at most") potentially does not exactly make sense business wise or even warrant the time building that free tier for an audience that's not willing to pay anyway.
But as I read the OP it is that he objects to the barrier of entry. He would prefer (possibly very harsh) rate limiting over the hassle of registrering an account. Maybe combined with a weak "nag" screen.
It might be hard implementing in a bulletproof way as IP restrictions are easy to circumvent. But it might be "good enough" to drive more adoption.
I'm a bit on the fence. It would be an interesting experiment.
If you want to hear from a happy Kagi user, I can say that I used Google thrice in the last two years, and it didn’t bring better results than Kagi.