Thread by @guywuolletjr

I still think ZK is under hyped. Let's talk about why, and a couple analogies from web2

We often talk about crypto's "iPhone moment" or Eternal September, but what finally tipped developer and user adoption over the edge?

The iPhone presented three new developer tools that hadn't existed together before. It combined a 1) programmable mobile computer, 2) GPS, and 3) a high quality camera.

From the camera emerged applications that were focused on photo sharing like Instagram. As @cdixon put it, users would "come for the tool (photo filters) and stay for the network (a social network)"

cdixon.org/2015/01/31/com…

From the GPS and camera combined we got apps like Uber and Doordash. These networks provided heavy subsidies to begin with and today have significant market power

I think something similar could happen with ZK, particularly zkVMs like Jolt. Verifiable compute is a new developer primitive, and I'm excited by a couple of use cases

I think DeFi protocols should adopt risk modeling in ZK and use more complex pricing functions that run off chain and are proven trustlessly onchain

Blockchains and ZK proofs could uniquely enable humans to own censorship resistant AI companions, an idea the great @packyM talked about his his Love in the Time of Replika post

notboring.co/p/love-in-the-…

Finally I think verifiable compute ought to be used to prove simple machine learning (linear regression or small NNs, not LLMs) models.

TLDR;

You could use ZK to build a more complex DeFi protocol, a censorship resistant AI companion, or onchain ML.

DM me if you do