Live data from Hacker News

The Age of PageRank Is Over

blog.kagi.com

61–70 of 380 posts

Re: The Age of PageRank Is Over

#61
post #43

Earlier quoted context omitted.

Maybe. They can hold batches and smuggle them in various ways.

If they have an auto-complete feature, the keypresses WILL be seen in not even in a smuggled way. They'll be directly in the URL to the search function!

You could probably hide that from anything but a deep dive by chucking the data through a websocket. Unless you see and recognise the initial socket connection you may miss it entirely.

Re: The Age of PageRank Is Over

#63
The author quotes Page and Brin's original paper:

> “Currently, the predominant business model for commercial search engines is advertising. The goals of the advertising business model do not always correspond to providing quality search to users. For example, in our prototype search engine one of the top results for cellular phone is “The Effect of Cellular Phone Use Upon Driver Attention”, a study which explains in great detail the distractions and risk associated with conversing on a cell phone while driving. This search result came up first because of its high importance as judged by the PageRank algorithm, an approximation of citation importance on the web [Page, 98].

A search for cellular phone now returns a page of links to plans, and one link to Wikipedia.

Re: The Age of PageRank Is Over

#64

> it costs us about $1 to process 80 searches.[0] I am genuinely curious how did Kagi arrive at this cost. [0]: https://kagi.com/pricing

I saw that too. Seems pretty high. I also think their plan of hoping future users search less on average is probably bad. I would want users to use my product more and I definitely wouldn't want to have an incentive to make users use it less. If the company is basically default dead unless usage drops... That seems really bad.

If I were them I would be focusing on: How do we get to 8k searches for a dollar?

Re: The Age of PageRank Is Over

#65
post #60
post #50

Earlier quoted context omitted.

How would you really blackbox a search ranking, though? A website owner can always just search for themselves, and see how their ranking changes. When there’s so much money at stake, people will go to great lengths to reverse-engineer the things that put them higher in the page, no matter how you try to hide the levers that make the rankings work.

Personalized results. If you Google yourself, you just see what you see. But you can never know what others will see.

> Personalized results. If you Google yourself, you just see what you see. But you can never know what others will see.

That just turns it into a statistical problem.

I don't think you can escape the problem of a search engine being used as an oracle.

Re: The Age of PageRank Is Over

#66

> it costs us about $1 to process 80 searches.[0] I am genuinely curious how did Kagi arrive at this cost. [0]: https://kagi.com/pricing

I’m guessing they’re running NLP inference on some GPU server that costs X per day and can process Y pages. I would guess that it’s the indexing that incurs the most of the costs, not the actual queries performed by the users - though I could be wrong.

Re: The Age of PageRank Is Over

#67

I've been using Kagi on and off for the last few months. I often forget about it since I can't set it as my default browser in Safari. But in general, especially when looking for specific phrases or trying to find discussions related to a topic, it's working extremely well. Whenever I used it, I was happy with it, and the search result quality was better than in Google. It's good to note though that iirc Kagi does ac…

They have a Safari extension. Have at it!

Safari mobile too: https://kagi.com/faq#How-do-I-set-Kagi-as-my-default-search-...

Re: The Age of PageRank Is Over

#69
post #51

Why does everything in the future-web have a price tag attached? If anything I want less business middlemen on the web period. I don't understand why funding through direct sales will give a better product than funding through advertising: "And yes, the non-zero price point will mean you have to budget it with your other costs. But faster access to higher quality information will make you much more competitive global…

Doing things requires hardware, which uses electricity and needs software which costs man hours.

Re: The Age of PageRank Is Over

#70
post #20

Can the technology behind GPT-3 be used to find relevant articles given a query? E.g. instead of training it to generate sentences given a prompt, train it to generate URLs given a query (?)

I would hope that GPT-4 or 5 will obviate the need for this entirely. Just ask the LLM your questions and it will answer.
Post reply on HN