People think of privacy concerns as a tax.

A common pattern today in AI frameworks: come up with a clever agent-based LLM approach with composition and then later figure out how to shoehorn privacy principles onto it, a kind of bummer of an asterisk to the system.

An alternative: lead with a new privacy paradigm allowing composition first, and then add LLMs to that.

This could create a system where a copernican shift of data provides leverage for whole new types of experiences to be possible that weren't imaginable before.

Privacy not as a bummer afterthought but as a fundamental game-changing enabler.