Having a friendly face to bend the rules creates a breeding ground for iniquity. We can’t have both - and the current direction seems to be to prefer reducing inequalities, rather than having people help break the rules for others that they like.
Under testing, everyone seems to be pretty bad at recognizing liars [2]. It’s especially hard to get good because it’s often an open loop, since good liars are never caught at all (classic toupee fallacy — you think you’re good at catching liars because you don’t know about the ones you missed), you never really get the chance to calibrate.
[1]: maybe with specific definitions of “average” and a very lopsided distribution, but I’m doubting that the people who think they have exceptional judges of character actually believe that they are above the mean but near the mode; usually, if someone is “exceptional,” I assume they’re in a high percentile
[2]: https://www.theatlantic.com/science/archive/2021/03/how-to-s... — https://news.ycombinator.com/item?id=26607386