Fair enough, I am also not attached to kagi, mostly I want companies with good business models to succeed in tech. I want to see ad-revenue based companies (and all the connected industry) to crumble. A man can dream…

But yandex is useful for those who search in Russian. The low utilization probably comes from a mostly US/EU customer base, but when it is used, it is useful. I would disagree with disabling it. The best would be letting people decide what back ends to use, but that requires a whole rewrite of the search logic on their side, so it’s not happening any time soon…

BTW in EU we still use a lot of gas and oil from Russia, so it’s quite difficult to avoid giving them money (especially because we don’t know where energy came from for every product we buy).

I think their customer base is basically 90% made of people that - like me - would quit in a second.

Good thing is that there is no vendor lock, it would be a shame, but changing search engine is quite simple.

Yeah, I agree. In general I will personally try to evaluate if the good that comes from a company succeeding outweighs the fact it’s a US company. I won’t use a dogmatic approach, but I will definitely be careful to choose even more carefully than before.

At least in Europe that’s still quite impossible, who knows what their gas and oil is used to produce. Which means you might buy some european product and also give them money. Anyway, everyone has their lines and I respect that.

I think most people are unaffected from the actual data, unless they search in russian, which is useful for me as a Russian language learner for example. I mostly search grammar stuff.

I don’t know the details, so maybe there is a reason, but I am not part of the “outraged” crowd. I think kagi use case is neat and innovative, bot protection is meh

Technically you could extend that reasoning to plenty of EU countries that also send aid to Israel (e.g., Germany, where Hetzner is located, or tuta, etc.).

At some point one has to make compromises, and everyone can place the line where they wish. Considering 1000 searches per month, the price is going to be between $0.20 and $3.84 (synchronous). So let’s say $2, which is probably an order of magnitude more than the real cost. Of that 2$, the margin is maybe 1$? That 1$ becomes profit for some Kazakh company, which ultimately means $0.2 in taxes. If this was in Russia, that would be $0.018 to the federal government, but let’s say that it doesn’t matter. Of that, 40% goes in weapons, making it $0.08/month. In 1 year, that’s $0.96.

Now, as I said I wouldn’t be surprised if this was an overestimation of 10x or more, it also assumes that absolutely nothing goes to Kazakh government, which is fully used to bypass sanctions, and a 50% margin for the company. It also assumes 1000 searches (the average was around 300 if I recall correctly) and that yandex is used for each one of them.

Every cent count, absolutely, but it’s objectively such a tiny amount that a one-time donation to UA army or some humanitarian relief org will offset you for like 15 years.

Since they implemented privacy pass, there is now something stopping them from doing both. See https://help.kagi.com/kagi/privacy/privacy-pass.html

Obviously with it you trade the need to trust them for your own personalization (as they can’t know it was you searching).

Sure, but they don’t (their privacy policy is exemplary). They have a whole shpiel about their business model. Just few weeks back they released a feature that makes it technically impossible for them to see who did searches, so no trust is needed anymore. They implemented a very novel protocol, quite cool.

I have doubts considering they are an american company, but I want to see them succeed. Plus, they are remote, so at least a good chunk of the income taxes from salaries are going outside the US.

Why would they do what Google etc. do, but much worse? It makes sense that they do scrape what google etc. most likely miss (and that’s what their index is about). Even a company with Microsoft resources tried and failed to scrape the web as a whole (failed in the sense results are worse).