I'd consider it if they had some of the features Kagi has like raising/lowering/pinning/excluding certain websites from results, but every time I try it it still feels very light on features.
Why not use a bicycle instead of a car? Because it's 10 miles to work.
One search engine is a hobby project, the other ones competes with Google and gives better results and ability to filter them and prioritize them as you want.
I don't get the impression that Kagi intends to compete with major search engines. It is clearly marketed toward privacy-focused, tech-minded individuals. You can take that one of two ways. Either you are frustrated with the erosion of search engine quality due to advertising, or you disagree with the predatory practices such as data mining that comes along with such advertising. In both cases, the only real way to signal to major search engines that you disagree with these practices is to stop using their services (including their APIs).
For example, I have been using DuckDuckGo for decades. At first, I had to compromise search result quality, but now it has enough users and support that results are on-par with the likes of Google.
I do not think that Kagi is bad or that people should not use it. It simply isn't for me, because it does not actually address the reasons I do not use search engines like Google.
On the other hand, it doesn't really matter so much anymore.
LLM is the new search. I can ask it the actual question I have and it will give me the answer. If it's not exactly what I need I can ask it to specify further.
Contrast that with a search engine that just gives me a ton of bookmarks to sift through to see if they actually might answer my question or are just clickbait.
Of course there's still some times when you need search, like when you need to find an actual website, or when you need a source reference. But really the need for me is greatly reduced now.
Be careful relying on LLMs for "searching". I'm speaking from experience here - getting actually accurate results from the current generation of LLMs, even with RAG, is difficult. You might get accurate results most of the time (even 80% or more), but it can be difficult to identify the inaccurate results due to the confidence models present their output with when hallucinating.
Also, if your LLM isn't doing retrieval-augmented generation (RAG), then it isn't actually a search and won't find results more recent than the data it was trained off of.
I know. But I'm often not really looking for accuracy. I just need to know something for myself. Most of the stuff I look up is absolutely not critically important. It's not like I'm trying to write a PhD dissertation or something.
I know it can be inaccurate but I can verify the results (and they usually are totally fine).
I think it's not that complicated -- Kagi's search results are just far more useful. I think it's marketed at people who want good search results, not anything dealing with privacy (although, Kagi doesn't log your searches, so it's fully private for most everyday definitions) -- your viewpoint for you makes perfect sense to me and sure I respect it, but I don't think it's right to say that people are linking their credit cards to do a have-to-be-logged-in-first search on Kagi chiefly for reasons of privacy focus.
(I just tried the same experiment Doctorow tried, of searching for something that I'd been unable to find through Google, and Kagi did the same thing for me that it did for him (i.e. found it). That's actually not important enough for me to pay for Kagi, but "Google is shit now" is no fringe opinion and it's pretty easy to verify that Kagi does in practice work markedly better.)
So do DDG and a lot of other search engines. In addition to the time and cost of running a spider and maintaining a database (for little to no technological benefit these days), a lot of server admins will block crawlers that aren't googlebot or msnbot/bingbot.
That seems like a crutch instead of a real feature. I hate even just thinking about having to manage that. What if you want info from sites you do not already know about? Seems like finding new things through search is basically dead anymore.
Websites I've never heard of come up very frequently. The feature I find highly useful involves a) hiding shitty disinformation websites as I encounter them, or blocking sites like pintrest, and b) elevating results from websites I like, like having letterboxd results rank higher than IMDB when I search for movies or whatEVER.
Being able to customize my own results to favour what I'm actually looking for is such a crutch.
Where can I find that setting? I don't see anything like that anywhere in the UI. If it's in the config files and not in the UI, that isn't particularly useful to me.
It's in the UI on the Engines tab. However, you can only see it there, you still have to use the config file to actually change it, sorry. That's not hard at all though.
Oh, I think this is different to what I'm talking about. Seems like the weight in SearXNG impacts search engines used, where what I'm talking about is in regards to the actual results. So I could prioritize or deprioritize websites which tend to produce good or bad results, block domains, pin favourites, etc. The example I used elsewhere is like having letterboxd results rank higher than IMDB when I search for movies
Ah I see. I didn't really understand the requirement. That would indeed be a nice one though pretty hard to configure for general search because the results can come from so many sources.
As well as that, for special-purpose things like movies it does in fact have a ranking for those by querying common sites like IMDB directly as an engine. So in that case you can use the weighting system to show preference. It doesn't seem to support letterboxd as a source but it does some others:
That's cool, but I'd still prefer not to be wrangling a dozen or so engines with bangs and editing config files when I could do it all through the UI in a more intuitive way. While I don't mind initial setup, I have very little appetite for endless tweaking of config files. So while I'll keep an eye on SearXNG and occasionally pull it onto my server, Kagi can continue being my daily driver for now.
I have the big SearXNG portal bookmarked ( https://searx.space/ ) but I don't find that I ever reach for it that often. Not being able to cull lower quality sites is just a little bit of extra toil I'm happy to pay to go away.
You do have to host it yourself or run your own personal instance to get the power of SearXNG; if you've not tried this, please do not write it off.
If hosting it yourself or even running it locally in a container on your machine at home is too technical for you; nobody is going to bane you for that. In fact there's several guides and videos out there that might help you if you're inclined to learn.
If not; you're also free to continue consuming as you do.