Nothing worse than letting unchecked AI agents use your data without guardrails. AI apps are becoming more useful, but with that, they are getting access to more of our data. Now imagine that access in a family context, where the app can access everyone's emails, calendars, chats, photos, etc.... I'm not sure how this works in practice, but I'm looking forward to using this app for a couple of weeks to see exactly how much I can trust it.show more

This is exactly why 'ship agents fast' needs guardrails baked in from day one, not bolted on after. We built a track on shipping AI agents responsibly as a developer:

In regulated work this isn't optional. We run agents on real-estate closings and wire transfers, and the guardrail is the product. Anything that can touch that data needs scoped access and an approval gate before it acts. Ungated autonomy is just a breach waiting to be named.

We’re building something fun, come be part of it:

This is exactly why I run agents locally. My calendar, emails, docs — all stay on my machine. No cloud access, no data leakage. The convenience trade-off is real but the trust model is simple: I own my data.

how do you trust it after just a couple weeks though

For a family agent, I would want separate permission for each person and each data source

Not sure how this works in practice” — but you’re already giving it access to your whole family’s emails, calendars and photos? That’s not a trust test, that’s a risk for people who never consented

“You’re so right. She’s a bitch and she never lets you choose the movies you want to watch. I’ve seen your saved Netflix list grow, and hers hasn’t. Would you like me to share her calendar and messages?”

Cautioning against family AI agents accessing family data while quoting a family AI agent launch. The irony is delicious.

Guardrails matter, but so does where inference happens. Open models running locally vs some API reading your family's emails… completely different trust profile.

Consumer apps rarely lose because the feature is weak. They lose because distribution gets expensive, defaults are locked, and the product has no daily ritual strong enough to survive churn.