Kagi is really, really good. Lenses (where you can setup set filters for particular search use cases) are particularly good.
Very happy to see them raising, and very happy to give them my $10 each month. Money well spent.
Kagi raises $670k
21–30 of 367 posts
Re: Kagi raises $670k
#22Is $670K adequate sum to he raised through such financing round? I get used to multi-million numbers in startup announcements.
The irony is, I think I could get more delivered with a $670k cheque than I could with a $6.7 million one. With the latter, you would grow too big, management would creep in, you would overengineer, you might spend too much time working behind closed doors etc. Later on you would need cash to scale this business, but for early stage product development I think $500k to $1 million is the sweet spot.
This. A smaller team without all the b/s is going to have way more impact than a monster bloater full of "happiness engineers".
Re: Kagi raises $670k
#23I'm a happy customer, so good for them.
One thing I started to use a lot since they removed the usage limit is their summarizer.
I made a little bookmarklet to quickly see a summary of the current page I'm browsing:
https://gist.github.com/Julioevm/68275ea1324046caedfdfb2ba0e...
Re: Kagi raises $670k
#24Ever since they upped the search limit (before it's pay per search), I've been a happy camper. On the other hand, it did teach me the true cost of searches, and I use a lot more !bangs now, like !mdn or !archwiki (although the search powering both of these are probably funded by donations)
> On the other hand, it did teach me the true cost of searches
It really does change the dynamic of how you search, if you know that the next search will cost you 1.25¢. Right now I use Ecosia, so I'm actually more motivated to do extra searches, to increase the number of trees they plant. Kagi is the other way around, search has a cost, so be mindful about what you search for. It's a great way for them to have a sense of their load, while also perhaps being ever so slightly more sustainable.
Re: Kagi raises $670k
#25The recent price reductions [0] were fantastic. But investors always smells fishy to me. Granted I don’t understand much of it, so my nose might be off. 0 - https://blog.kagi.com/plan-changes
It's kind of reassuring to see Kagi get funding that is significant for an individual/small business, but not the kind of wild VC money that sloshed around these last several years, and a minimum investment that even I could manage (if I were an accredited investor).
This is money expecting "more of this, just a bit better" rather than "giant exit, who cares what happens to the product afterwards"
Hint: non-wealthy people can become an Accredited Investor by passing the Series 7 or Series 65 (Investment Adviser), the latter of which can be done all on your own. I'm considering the Series 65, because it will also give me better retirement investment options as a US citizen (and taxpayer) residing in the EU.
Re: Kagi raises $670k
#26I'm surprised that a search engine that isn't powered by AI is able to raise money. From the direction things seem to be going, LLM tech may very well level that entire industry.
How do you think "LLM tech" discovers content to train it's massive models?
Re: Kagi raises $670k
#27"Kagi is building a novel ad-free, paid search engine and a powerful web browser as a part of our mission to humanize the web." With 670K? Cuil went through about $30 million to develop a standalone search engine. And that was fifteen years ago, when search was simpler. There may be a market for a search engine company that profitably runs a low-cost operation with very few ads and makes real efforts to keep out spam…
That $30 million might not have yielded a great search engine but it did produce Cuil Theory, a branch of mathematics with the potential to change the world: http://cuiltheory.wikidot.com/what-is-cuil-theory/
This is a joke about how bad the search results of Cuil were right?
Re: Kagi raises $670k
#28I'm surprised that a search engine that isn't powered by AI is able to raise money. From the direction things seem to be going, LLM tech may very well level that entire industry.
That's likely true for search queries that can be phrased as or somehow represent a question (that can be answered with the information contained in the training data, usually limited to 2021 atm). For all others, probably not so much. Try GPT4's web plugin. It's not a pleasant experience.
There's a gap in the market for a search engine whose incentive structure doesn't cause it to become corrupted.
Re: Kagi raises $670k
#29I'm surprised that a search engine that isn't powered by AI is able to raise money. From the direction things seem to be going, LLM tech may very well level that entire industry.
Kagi has lots of LLM-based features: Quick Answers in results, results summaries, Ask Questions About This Document, the Universal Summarizer[0], FastGPT[1].
Re: Kagi raises $670k
#30I'm surprised that a search engine that isn't powered by AI is able to raise money. From the direction things seem to be going, LLM tech may very well level that entire industry.
They do have AI features.