The Mathematical Mesh

The talk starts from the regular stream of data breaches and asks why personal data is not simply encrypted. The speaker argues that encryption is the easy part. The hard part is getting encrypted data onto every device a person uses, sharing it with colleagues, including those who join later, and recovering it after a loss. He presents the Mathematical Mesh as an open specification with reference code that handles private key management, so that applications do not each have to solve it.

The demonstrations cover a sequence of scenarios. A second laptop is joined to an account by sending a connection request, comparing a short verification value on both sides, and granting only the privileges the device needs. Key generation is split between the two devices, so the final key stays secure unless both fail. Group decryption is shown next: a file encrypted to a group becomes readable to a newly added member and stops being readable when that member is removed. Service providers hold key shares that let them withdraw decryption ability without being able to decrypt anything. Contact assertions and trust tracking are presented as ways to keep a relationship stable when messaging providers or addresses change. A browser built on WebView2 shows access to one website being granted and revoked.

The closing section sets out a wider goal: open alternatives for messaging, bookmarks, social media and internet-connected devices that users own rather than rent from a vendor. The speaker asks for contributors and notes that anyone can run a service provider with a static address and a domain name. The code is published on GitHub under an MIT license. The main takeaway is that threshold cryptography can spread control of keys across several parties instead of concentrating trust in one.