9 ms·
What that really means is that they either: 1) aren't doing the sensor fusion we all think they are, out of inability to access the data. 2) the models (and t
by huffmsa 4y ago
What that really means is that they either:
1) aren't doing the sensor fusion we all think they are, out of inability to access the data.
2) the models (and therefore modelers) aren't good enough to use the data they have correctly.
- chillchilla 4y agoOr 3) they run loads of A/B experiments to optimize engagement or some target metric and they've reached a local minima and are unable to escape it without a lot of political will or Product Managers willing to stick their necks out.