> The whole explanation of the recent “we want to stay afloat” was weird
I didn't find that weird at all: They were overly optimistic, both on how many people they'd be able to recruit for a paid service, and on the economics of how many searches these people do.
I'd be willing to bet that those skew heavily in the direction of power users and knowledge workers using search professionally, so they probably do a lot of searches. I was really surprised when I read my own numbers. In the survey they asked me to fill out, I wrote that I'd expect I'd be doing maybe five searches a day. Turns out the number is actually in the range of 30-70 most days.
One data point that seems to support that is that Kagi previously didn't take any outside investment, with the whole thing being bankrolled by the founder who had had a prior successful exit. But at the same time when they announced the pricing, they also announced they'd be willing to take on board small investments from enthusiast users. So it's looking like they really are starting to get worried about money.
I also indicated in the survey, that, for a really good search engine, the amount of money I'd be willing to pay would be a multiple of what their actual pricing is now. A good search engine is an investment that could pay phenomenal dividends for me, allowing me to make better decisions for my business by having better information as input. It allows me to waste less time, when searching for solutions to technical problems. It allows me to connect to the right people. So, I don't want to think of search as a commodity, like electricity. I think that, as a knowledge worker, I'm really only ever as good as the search engine I'm using, and I just want to hand somebody a pile of money to make me the Ferrari of search engines so I can have a competitive advantage against all those people still riding the bus. (The bus is full of billboards and paid for by people who never want the bus to reach its destination, so they can keep the passengers' eyeballs focused on the advertising).
> [...] A good search engine is an investment that could pay phenomenal dividends for me, allowing me to make better decisions for my business by having better information as input. [...]
It sounds a bit like the use case of ChatGPT/BingAI. Depending on your specific use case it might be the wrong fit - especially if you want to look up PI - however it is excellent as a rubberduck/idea provider.
It's precisely the antithesis of a use case of ChatGPT/BingAI. I want sourced information, and the opportunity to read the source in its original form, make a critical judgment about it, and draw my own conclusions. Giving me the opportunity to ask a question and get an answer, when I have no way of knowing whether the answer is a hallucination is less than useless. It's a waste of time.
At least BingAI does provide its sources and ChatGPT/BingAI are made to be conversational, so you get the opportunity to ask actual questions and get an answer. As with everything on the internet take the answers with a couple of grains of salt.
You can also directly ask for the sources btw. In the end LLMs with access to the internet can be an abstraction to search engines. If it is good can only be assessed once you actually tried as it heavily depends on your use case and your approach.
The other thing I'm wondering, even though this is veering quite far off the original topic, but: I don't just want all the sources that the AI used in responding to my question, I also want all the other sources that seemed relevant to my question but were for some reason not used.
I want completeness (i.e. if source material is being skipped, I want to be the one to make the decision that it can safely be skipped), and I want consistency (i.e. I want to be able to double-check that the same set of sources wouldn't also support drawing conclusions that contradict those drawn by the AI).
In other words: I'd still want to use search as a product, and I'd still want someone to solve the search problem for me, even if question answering were a solved problem.
I didn't find that weird at all: They were overly optimistic, both on how many people they'd be able to recruit for a paid service, and on the economics of how many searches these people do.
I'd be willing to bet that those skew heavily in the direction of power users and knowledge workers using search professionally, so they probably do a lot of searches. I was really surprised when I read my own numbers. In the survey they asked me to fill out, I wrote that I'd expect I'd be doing maybe five searches a day. Turns out the number is actually in the range of 30-70 most days.
One data point that seems to support that is that Kagi previously didn't take any outside investment, with the whole thing being bankrolled by the founder who had had a prior successful exit. But at the same time when they announced the pricing, they also announced they'd be willing to take on board small investments from enthusiast users. So it's looking like they really are starting to get worried about money.
I also indicated in the survey, that, for a really good search engine, the amount of money I'd be willing to pay would be a multiple of what their actual pricing is now. A good search engine is an investment that could pay phenomenal dividends for me, allowing me to make better decisions for my business by having better information as input. It allows me to waste less time, when searching for solutions to technical problems. It allows me to connect to the right people. So, I don't want to think of search as a commodity, like electricity. I think that, as a knowledge worker, I'm really only ever as good as the search engine I'm using, and I just want to hand somebody a pile of money to make me the Ferrari of search engines so I can have a competitive advantage against all those people still riding the bus. (The bus is full of billboards and paid for by people who never want the bus to reach its destination, so they can keep the passengers' eyeballs focused on the advertising).