Just caught an interesting take from Xiao Feng on why AI's real commercial breakthrough might not come from privacy tech alone. He was discussing how hospitals sit on massive amounts of data that could transform drug development and treatment planning, but they're stuck in a dilemma: share the data and lose control, or keep it locked and miss out on the potential value.

Here's where it gets compelling. Privacy computing tech like fully homomorphic encryption and zero-knowledge proofs solve the 'data usable but not visible' problem pretty elegantly. But according to Xiao Feng, that's still just Web2 dressed up differently. You still need negotiations, licensing agreements, payments, the whole bureaucratic dance.

The real shift Xiao Feng is pointing to? Combine that privacy tech with blockchain. Transform medical data into permissionless, automatically billable AI tokens. Hospitals wouldn't need to know who's accessing their data anymore. Someone uses it? Revenue flows automatically. No middleman, no endless negotiations. That's the kind of trustless model that could actually unlock the value trapped in all that dormant data.

It's a pretty different way to think about the Web3 opportunity. Instead of just decentralizing platforms, you're fundamentally changing how data monetization works. Xiao Feng's arguing this combo of privacy computing plus blockchain could expand what's actually possible in AI applications way beyond what either technology does alone. Worth thinking about if you're watching how these sectors evolve.