Yes, but if someone gets that key, the worst they can do is encrypt encrypted data.(I think…I may be misinterpreting this).
Where this gets i is, I think, other libraries they have involve some basic multiparty compute. So you can apply basic logic without having to see the data. I saw a demo one…it was like 3 years ago. Looked like cool tech.
Just looking at the code - they need the original private key for round #2.
Bob is the only person who does not need it. Translation: our middle-man canister proxies would need it.
i.e. Node provider gets the key used in 1 and in 2 so they can undo the whole thread back to original text.
Anyone know people at Dmail?
‘VRF (measurable random function) + BLS (non-interactive threshold signature) + Container.’
https://dmail.ai/Dmail_litepaper.pdf look in Distributed storage.
Or Dmail Network: Decentralized EMail for the Web 3 Generation in DMAIL Message.
They were thinking about using a token to do this.
Hey! This was a little easter egg that I placed on the roadmap - there was a way to ‘decrypt’ if you somehow collaborated with the page, but indeed it looked confusing. I made the text ‘normal’ now, so you can read. This is a feature currently in discussion - will start a forum thread in the coming days 
I copy and pasted it to read it. What was the desired way to decrypt?
Also - super excited for this feature as it is a major blocker to b2b coming to the IC with $$ for devs.
This is great to hear! we are super interested to know what have been the main blockers to could be solved with this, so keep them coming. Looking forward to kicking off the discussion!
Sharing for visibility 