What about validating self-produced scores? Is it supposed some actor in the middleware should be responsible to verify those scores?
Discussion
At the moment I dont need this. But let's say someone wants to pay 200k sats for some work. I could probably finish a decent task in that time, like to make the prototype of an app. There will be a market for this, so that if the payer is satisfied they'll buy more, or increase the rate. If they are not satisfied they will go elsewhere. The system will evolve, especially as it will include humans and machines doing tasks, and machines can validate work. Humans can validate the work of machines. This is indeed a vital part of the workflow. Or marked zaps can replace validations.
Lets say I want to pay 100k sats for labeling of 1000 nostr notes. I send sats to you somehow since you hosting the app, correct?
Then your app show this order as “ready for work” or something. And then participants can take that order in work. They can constantly stream their progress and once they reach 1000 they get automatically paid
The thing is, if my assumption is correct that you get sats because you host the app, then you must validate workers results to maintain your reputation. There are probably infinite ways on how to validate, simplest are like:
1) KYC workers on sign up to check their skills etc and if customers complain, figure it out with workers what was wrong
2) once worker submitted score 1000 validate actual results for instance by manually scanning random 100 records and penalize worker if errors found
And in this case its a very interesting system however the main killer feature part is validation. Do you have any specific ideas on validation flows or you have different priorities for now?
Forgot, you must validate results because there can be a bot that listens for events and automatically submit max scores right when job is ready for work