Comment by eszed
9 months ago
But, again, do they need to? It seems to me like "market share" is a metric relevant to companies pursuing VC-founded, unicorn lottery-ticket scale. If they generate enough revenue to pay competive wages, cover their operating costs, and make a reasonable (real-world, not VC-world) return on investment, they're a gosh-darn success. It's only within tech, where valuations and evaluations sailed off into ZIRP-ified bizarro-world, that people think of that as a failure of ambition or execution. I think it's time to re-assess our mental models.
> But, again, do they need to?
Why not? There are tens of millions of people who need/want a high quality search engine and can pay for it. Kagi deserves to be successful for having made a better search engine than Google. And their success can inspire other entrepreneurs to start delivering quality information products, so that maybe we can get out of this ad/scam fuelled quagmire once and for all.
Good products and ideas should be successful, that's progress.
That is a complete non-sequitur from the question of if Kagi will die if it doesn't grow.
This is orthogonal to questions of morality and justice.
> This is orthogonal to questions of morality and justice.
What?
If Kagi doesn't grow, I fully expect the owner to eventually shut it down and move on to more fruitful ventures. Nobody owes anybody to keep a business running. So yes, it would die.
3 replies →
> do they need to?
I don’t know what Kagi’s minimum sustainable size might be, but it’s probably bigger than what it is now. Particularly if they want to stay competitive with LLMs.
I suspect you're correct, and am rooting for them to hit sustainable, as soon as possible. I don't know what their maximum sustainable size is, either - that's equally important, though always a moving target. I only wanted to point out that neither inflection point, for a paid-service business model, has to do with "market share" - that's a VC thing, to which I'm increasingly allergic.
I use Kagi for finding things and LLM for asking questions. Two different use cases. I want them stay separate but I am probably in the minority.