Nostr Web Client

Again, how are these going to be created? Who puts all the work to download all the data and index them? If you only care about like a 1000 blogs, then fine you can always fetch latest updates and read them all like RSS works.. but that has nothing to do with search or algorithmic feeds... These by definition require unbounded number of data you have to crawl through to find the best of what you are looking for

Analogue Dog 2mo ago 💬 1

I guess its a trade-off?

I consume a lot of video and audio content, but I'd get by with a transcription of podcasts and informational videos.

The info that's really important to me is predominantly text: Health related stuff, techniclal/build guides, nutrition, general encyclopaedia stuff. Again, most of it is blogs, podcasts, journals. I don't think it's unrealistic (in terms of resources) to build and maintain a bounded index of all of this and share signed versions with peers in Nostro/Vostro directory structure on a public key addressed home server. There could be an economic incentive for sharing perhaps?

This obvsly won't compete with huge centralised search engines, but I think it's probably true that after a certain volume of content, there's a diminishing return (in terms of value) the more that's indexed.

Appreciate your thoughts.

Reply to this note

Please Login to reply.

Discussion

Nuh 2mo ago 💬 1

The web once was mostly text, and it still had only handful then only one search engine... The web is huge you can't crawl it cheaply even if it was only text.

Yes someone can do the indexing and sign it and share it, but who cares, how is that different from visiting a website, the index remains centralized and thus the Indexer will both comply with censorship AND look for ways to make money to offset the cost... That is also ignoring personalisation and suggestions entirely... Just a dumb index.

Again, someone has to download the data to discover it ... And that can't be made cheap

Nuh 2mo ago

Best possible option is to let millions of people crawl small parts of the web, index them and generate a ZK proof of that and get paid for their work _somehow_ then share these parts with each other in some sort of a DHT.

However, a centralized indexer will always be 1000000x efficient because zk proofs are expensive as fuck.

Even then I don't think this will work well at all.

It is much easier and cheaper to just pay someone for a high quality aggregated feed