running nostr:nprofile1qyxhwumn8ghj7mn0wvhxcmmvqydhwumn8ghj7mn0wd68ytnzd96xxmmfdecxcetzwvhxgegqyz9lv2dn65v6p79g8yqn0fz9cr4j7hetf28dwy23m6ycq50gqph3xc9yvfs

Reply to this note

Please Login to reply.

Discussion

You can use @routstr in contex.st

Can the routstr spec be extended to add support for confidential computing? Would be nice to have a MapleAI alternative that uses a decentralized list of providers

yeah they can do confidential computing, but you must trust AWS. homomorphic encryption solves this but it is a ways off

mapleai can run a routstr node

I think Maple additionally encrypts the data sent to their server using a pubkey owned by the CC hardware, we'll have to add that to the routstr spec. Also, verifying the attestations that Maple does would also need to be a part of the spec I think

live debugging with the routstr devs sitting beside me :D

What are the prerequisites and cost/earning running it?

You have to load some ecash for it to spend on llm queries. That's it!

As for earning...well that's up to you to produce something worthwhile.

I meant earnings running it as a provider!

Amazing! ⚡️⚡️