1. Google clearly lost focus and is rotting. Sundar Pichai makes Steve Ballmer or Brian Krzanich look good by comparison. They're killing search to optimize metrics like "# of searches" that are harmful to long term user value of the platform [1].
Google search has been getting worse for ~5 years, and will not change direction until they get a new CEO. As long as Pichai is at the helm, they'll keep optimizing the KPIs that boost next year's bonus at the expense of making a good service.
2. Kagi does a lot right, but it's run by the kind of lunatic that will spend a significant fraction of the company's budget to buy a t-shirt factory to make t-shirts instead of, you know, making a search engine.
3. Bing exists simply to be a corporate alternative to a competitor. That doesn't make for a good service.
4. DDG <-> bing when it comes to the actual search results. Nothing disruptive in here.
5. LLM based search is nice, but it doesn't fill a lot of the core things people actually search for on a day to day basis. It's a value-add, not a replacement.
Couple of thoughts:
- With generative AI filling the web with noise at an increasing rate, it'll be more important to rely on actual trust of sources rather than the content that's in there. Network/graph based metrics should do better than parsing/embedding the content itself.
- If you're building a search engine service, maybe try to differentiate it for the 21st century? Start by targetting the kind of people that could switch - the same kind of people that already paid for Kagi.
Paying for search is a good idea; it's just that Kagi is run by a loonie. Companies run this way don't tend to last.
Kagi managed to get going with less than 5 engineers. You can bootstrap your project from commoncrawl these days.