Your data is worth more now

AI agents are only as good as the context they can access.

Right now personal data is fragmented across 20+ apps. Health data in one place, tasks in another, notes somewhere else, calendar in a third. None of it talks to each other. The person who unifies their data into one queryable layer gets dramatically more value from AI than someone whose data is scattered. This is Palantir’s insight applied to individuals: the value isn’t in the model, it’s in the data integration.

Personal data stores have been tried before and failed. Solid (Tim Berners-Lee), Hub of All Things, MyData, Meeco. Same thesis, no traction. People said they wanted data ownership but didn’t act on it.

The difference now is agents. Previous attempts failed because humans had to do the work of unifying their data — build connectors, maintain schemas, do the integration manually. That was a human bottleneck. LLMs change this. They can read a CSV, parse an email export, interpret a health app’s JSON dump. The cost of building a connector dropped from days to minutes. Agents create the demand that was missing before.

Previous unification attempts also tried to define one schema for everything — vCard, iCal. Too simplistic or too complex. You don’t need the world to agree on a contact schema. You need your agent to understand your schema. Define your own types, start loose, add structure as patterns emerge. A slightly messy schema that an LLM can interpret is more valuable than a perfect standard nobody adopts.

The obvious counter: Apple and Google are better positioned. They already have health data, calendar, email, notes, and now AI connecting them. They might win. But their data is still siloed within their own standards — it can’t easily cross-link or be queried across domains. And they might sell your data, or lock it to their ecosystem. Local-first is the hedge.