Conversation

Replying to and
For context, I built a health tracker (Track-well, now shut down) and the data is a huge pain. Extracting the data is an additional pain. For example, in my back end the data was normalized across many tables in the database. Extract was a complex query I never got around to.
1
1
Replying to and
Yeah I built a thing (trailmeme) at Xerox where we did offer data downloads as xml, but I appreciate that health data is both higher volume and higher sensitivity’s a Pita. But they just raised $100mm at a billion plus valuation and charge $30/mo, so... fair expectations
1
2
Replying to and
They claim to sample HR at 100hz. Assuming short int precision that seems to work out to be ~82MB/day. I imagine they have a galvanic skin response too etc. So ~200-300MB/day? Not trivial but not horrendously hard to expose at scale, esp if only ~1-5% do anything with it.
1
1
Replying to and
Because it’s the right thing to do. Only 1-5% probably read TOS, but having one is the right thing to do. It is a costly signal of how you view the personhood of the counterparty in a transaction. Piece of meat? Sucker? Someone who deserves to have certain questions answered?
1
6
Replying to and
It’s not a commercial/business model decision unless you decide to pursue an open-model competitive strategy. That’s why it ends up in the regulators’ inbox. Societies tend to have shared norms about personhood that ends up setting regulations/rights.