#askNostr #DVM

When making a DVM job request, you can add relevant inputs in the `i` tag, but also in the `param`. How to decide which kind of inputs go where...?

For Github actions over Nostr I only make references to the input data, like repository and a .yaml file within that repo. Should that data be in the `i` tag or just `param`s?

Bonus question:

What do you think would be the ideal way to charge a customer for jobs that have wildy unpredictable execution times?

Some ideas i have:

- Include a payment larger than the expected cost, whatever's left will be returned to the customer

- Similar (but different) to the one above: Add cashu to a new NPUB and share the nsec with the DVM. DVM takes the money as it runs and leaves the rest. User reclaims the remainder.

- Some kind of NWC magic so the dvm can charge that

- ... your idea....

Reply to this note

Please Login to reply.

Discussion

Not an idea but more complexity. The pricing also has to reflect resource consumption over time to be competitive. Initially this could be sort of an instance size, but still to get the best offer from multiple vendors they need to be comparable.

You're absolutely right! I'm planning to have a similar pricing model to cloud vendors. Price per minute, 'billed' by the second. For various combinations of vCpu/ram/storage