I'm a pretty forgiving person, I don't really mind mistakes as long they are 1) admitted to 2) steps are taken to actively reverse course, and 3) guidelines are taken to prevent the same mistakes from happening.

But you more or less drain thst good faith when you are caught with your pants down and decide instead to double down. So I was pretty much against OpenAI ever since the whole "paying for training data is expensive" response during the NYT trials.

----

In general, the populace can be pretty unforgiving (sometimes justified, sometimes not). It really only takes one PR blunder to tank thst good faith. And much longer to restore it.

Mistakes should be made once and once only, irrespective of good or bad faith. It is no longer a mistake when you do the same misstep over and over again, it is a deliberate pattern of behaviour.