While I agree with others that AI can take away from a products core vision, I’ve been very happy with Kagi’s path and roadmap. I feel like the AI products that you guys have released have served well as complements to search, and hope the trend continues.
Hopefully this helps with indexing while offering a cool service to small creators!
Edit: I forgot to say, the change where a `?` appended to a search triggering the quick answers was an amazing change. I would love to see more features that can be invoked by appending or prepending to the search query.
How will you prevent someone from connecting Sidekick to a website that appears at first to be a small website but eventually fills with LLM-generated SEO keywords and ads? You might be able to manually review sites when they first sign up for Sidekick, but once the channel is open for them to inject content into Kagi's index, what's to stop them from abusing their privileged position?
In parallel, we are developing LLM-content detector technology to be more efficient at detecting such content regardless of how it is monetised (and we will offer this as an API once developed).
(Source: have created such schemes, although would generally not recommend them to my customers nowadays)
This could also be an attack vector for adversaries looking to pollute Kagi's search results and/or force you to divert resources to policing it.
In fact, I would rather they not get penalized for it, since low-quality SEO content is a good way to show up in certain other search engines (Google), and every business wants to show up in Google, making that content quite common even from reputable businesses making a quality product.
So, we shouldn’t penalize low quality, SEO, spam because of people’s wants? I do want them to penalize those sites because they are a disservice and more often than not crappy, unsecured WordPress that drowns out those that are not spam.
Thank you Kagi team! A shame how far Google’s results have fallen.
Edit: also SEO is one of the more seedier parts of the software industry. Tons of unaware small businesses conned into these awful, low quality sites. I literally quit because it was so morally bankrupt.
> they are a disservice and more often than not crappy, unsecured WordPress that drowns out those that are not spam
What does that have to do with "reputable businesses making a quality product" as my comment said?
I have my doubts about the effectiveness of this method and realistically, it won't make any difference because the bad actors will just use an LLM that doesn't snitch on them, so you're technically correct.
The problem is that people keep consuming the samey low quality content instead of skipping it (think superhero movies and Netflix series that are all indistinguishable from each other). As long as they're satisfied with that, they'll fall for fake product reviews too.
Someone did something like that to identify HN authors (as in correlating similar writing styles between pseudonyms) a few years back, for example: https://news.ycombinator.com/item?id=33755016
Or a study applying similar analysis to LLMs: https://arxiv.org/pdf/2308.07305
Of course, LLM output can be tweaked to evade these, just like humans can alter their writing style or handwriting to better evade detection. But it's one approach.
Combine that with that guy who boasted about his 'SEO heist', I think it's a very valid concern.
And we…want to discourage writing useful web pages, even though articles on understanding TypeScript's type system aren't all that closely related to their main product…? What am I missing?
I don't see the problem with what you're describing. It seems like one of the most contributory ways to market well.
Another decent one would be linux sysadmin info from Digital Ocean and the likes.
But for every joelonsoftware there are 99999 sites that have all copy/pasted the same tutorial about something basic and try to push some random product or just ads.
I've also found that type of developer marketing valuable many times in the past. It's sometimes obvious its going to end in a pitch for the product, but often it does a good job summarizing the key problems in the space, mentioning or showing other solutions / offerings, and pitching which tradeoffs they made for their own product and how they solved issues.
Even if you don't go with the ad, you can quickly pivot to other named players or get a better understanding of the terminology or jargon to start searching more.
I think kagi sidekick would be very well received in the bioinformatics space. Lots of complex docs that require end users to digest large complex data.
Can it be tuned to only point users to the docs and not answer questions?
Seems like the problem in [2] is a few entities controlling the majority of spaces other than search, to me. Shame we don't have any real laws against anti-competitive behavior (just the way YC likes it).
inb4 flagged
Do you know that "sidekick" is aibó in Japanese (相棒)?
Notice the "AI" in it?
aite
愛犬 aiken: beloved dog
相手 aite: companion, other party, opponent.
藍色 aiiro: indigo blue; 濃藍 koai: deep indigo.
...
According to Wiktionary, ō is used as a long vowel in two Latvian languages, and in Swedish as a hand-written form of ö (not always a long vowel). Also in Silesian, a language in a region of what is now Poland.
Since I'm originally from Slovakia, I used my Slovak IME, which is convenient for me.
Incidentally, Slovak has its own romanization of Japanese, which uses almost nothing but Slovak letters and diacritics (plus "w").
Wikipedia pages about 九州 and 四国:
https://sk.wikipedia.org/wiki/Kjúšu
https://sk.wikipedia.org/wiki/Šikoku
The Hepburn system is an Americanism. Americans don't own the Roman alphabet or the way it should be applied to Japanese. Though Hepburn has official status in that it is taught in Japan, and used for the benefit of visitors. E.g. signs giving names of train stations or government buildings or what have you.
I guess this assumes that you aren't already doing that when they click one option over another for a certain search term.
Just thinking about searching through some documentation sites and you get a dumb result you weren't looking for at the top, and would want to deprioritize that result.
For a random example, there's the Baldur's Gate 3 wiki (https://bg3.wiki), which has upwards of 8,000 pages often with pretty dense text (see https://bg3.wiki/wiki/D%26D_5e_rule_changes for an example) and is funded entirely off donations.
If it worked in a shell script or similar old-school unix architectural style on my bespoke static site generator, which is a slow-motion train wreck of a weekend hack-fest being ported from python/staticjinja to rust/minijinja.
Is kagi competing with whoogle?
Whoogle gives me the old-school, seemingly linear algebra of pagerank, hauntology that I expect.
Keep up the great work, you have an incredible product.
Because "AI" increases the shareholder's chances of winning the lottery.
If you offered such subscription, it would motivate me very highly to buy a subscription.
With the monthly price, can't really afford it.
Why does this exist?
It would be an efficient way for us to build and expand our own index. Assuming users of this would be Kagi users, we would expand our index by tens of thousands of high quality? personal websites, hobby projects, startups, documentation websites, etc., also helping to surface them in our results (where relevant, like we already do with Kagi Small Web initiative [1]). It is a win-win for both our users and Kagi.
It would also be a way for Kagi to get some exposure outside of kagi.com (provided the search widget has some branding on it).
This is why it makes sense to offer it for free for smaller sites/projects.
And crowdsourcing index is completely opposite direction of one that causes deterioration of web search results in ad-supported search where few entities control the majority of space [2], so we like it.
That is the plan - and since this is a "Labs" project, we are open to it crashing and burning. Know we do not, until we try. Try and try again, we must.
[1] https://kagi.com/smallweb
[2] https://detailed.com/google-control/