pull down to refresh
896 sats \ 5 replies \ @k00b 8 Sep 2022 \ on: How to Disrupt Google bitcoin
Max and I discussed this when it was still a draft. It immediately got me thinking. Mostly about how
- trust is hard to decentralize
- the incentives in a trustless system are hard to perfect, and
- bootstrapping would have to be done in phases
Recording value transfer on Lightning is hard. It is intentionally designed to be private. You’d need to record “receipts” (ie records of satoshi depth as Max calls it) for value transfers and it’d have to be hard/expensive to fake such a receipt. If transferring money is cheap (or the benefit of transferring it exceeds the cost of transfer), people will transfer as much money as they have to to get the effect they want.
Thus, creating a receipt either has to be expensive or you need to determine whether users generate trustworthy receipts. Either is hard even in a centralized system, but (today at least) trust is particularly hard to do in a decentralized system.
Further, if we expect people to upvote content in search results, they’ll need to be incentivized to do so. The 1% rule, while typically applied to "engagers" in forums, could probably be applied to any intrinsically motivated activity, like the "givers" in these value for value contexts. To bootstrap a large community of givers then, it would probably help (and might even be necessary) to extrinsically motivate them.
How I’d begin building this today:
- Start with a page rank-like search engine
- Allow users to "tip" content in results sending part of the payment to
- the site (using something like a meta tag with a lightning address in it)
- the search engine
- the higher the amount sent to the search engine the more "signal" or satoshi depth established by the tip (because it's a more "expensive" transfer)
- Let the satoshi depth influence results
- To bootstrap, reward tippers finding good results - somehow, someway
There's a ton to be figured out in the design obviously. What role does trust play? How to disincentivize whales from paying for higher results? But this is how I'd start. I agree with Max that there's likely something there.
I love this initial roadmap! If anyone is interested in experimenting with it please, let me know :)
My hunch is that roughly 1% of tipping users should be enough to generate signal. It will likely be a higher % if there are rewards beyond pure altruism.
I think you can generate user rewards via other income streams - e.g. a jobs board like Stacker News'. Or maybe a premium subscription with additional privacy or enterprise features? Or maybe tied to another revenue generating biz like @TheBTCManual described with Ahrefs and Yep.com?
And yes the trust model piece needs a lot of experimentation. Perhaps it's established with LSATs, DIDs, or could maybe even be tied into Nostr as someone mentioned to me this morning. Lots to explore there...
reply
And why not make the whole thing community-based. Users could, for a certain cost, submit search results for certain search queries. Those would then be displayed to future users of the search engine. The users who submitted the search result could then be rewarded when someone clicks their search result or upvotes it. Needs thought how to make the whole thing sybil resistant though.
reply
I'm definitely interested. No idea how you'd bootstrap a search engine though.
You could also pay sats for certain keywords and could buy the top search result that way (similar to ads) as another way to generate user rewards.
Check out "Presearch", they did something similar - just unfortunately with their own shitcoin.
reply
Thanks for heads up!! Lmk if you start experimenting :)
reply
Maybe a hybrid of a page rank search engine and stacker.news? Each page (or domain) has paid upvotes and a discussion attached? If done at the level of domains it's easier to centralize upvotes, regardless of the search query, than at the page level
reply