On May 29, 2025, at 11:39, Kurt Wendt kwendt@pakoinc.com wrote:
So, yeah - item (2) had a reference to "primarily for improving their services". But, it's not really clear - I guess I figured improving meant learning - but, I now know it does NOT mean learning.
Think of it this way: if your interaction could change the underlying model, what is to stop a bad actor from deliberately feeding it incorrect (or dangerous) information?
-- Ed Leafe