Eighteen open problems · 5 of 19

Making a digital footprint unpredictable rather than private

Data removal services sell to people who can afford them, and they lose, because the record regenerates. The other move is to leave the footprint in place and make it useless: seed enough noise into search history, cookies, location and scroll patterns that a prediction model gets garbage. Not privacy, predictability. The buildable half today is the measurement, which is showing someone exactly how predictable they already are from public sources alone.

The open question. Is poisoning a profile better than trying to delete it?

What would kill it

Adversarial products aimed at companies with far more compute lose slowly, and the legal ground is not solid.

Conor's, prompted by watching someone run a full background check from a laptop with nothing proprietary.

Tell me why this is wrong

That is more useful than telling me why it is interesting. Reply to [email protected] or on LinkedIn.