The optionality of the data is its power.
- The optionality of the data is its power.
- More optionality, more power, more danger, more explosive potential.
- Denaturing removes that optionality.
- For example, filter out any data that doesn't meet some k-anonymity threshold on some data stream.
- If you made the right denaturing call and it's the right signal you need, you can get most of the value with less of the danger.
- But the problem is you can't change what subdata to get later, so if you guessed wrong, you can't change your mind later.