 But when you're communicating with other people, it's not just what you say, but how you say it, of course, that conveys meaning. We had people communicating either with their voice, talking, or typing to the other person. And typing is, when we type to each other, a lot is missed, okay? So we had them communicating sincere or sarcastic statements. You knew as a sender that they were sincere or sarcastic. We had our senders predict how many of the recipients would be able to guess correctly out of these 20, how many are sincere and how many are sarcastic. We also had the recipients then tell us of these 20, which are sincere and which are sarcastic. Now, in the abstract, it's not going to be a surprise that in this experiment, people could actually detect sarcasm better when you could hear what they had to say than when you couldn't, right? So when you can hear what they had to say, they get about 75% of them right. They still miss quite a few, but better than chance. But when they were just typing to each other over email, they were barely better than chance. They got 56% right, okay? So that's kind of not a surprise in the abstract that we can interpret things better when we hear what somebody has to say, then we just read it. But that obvious thing wasn't so obvious to the folks who are actually communicating. They predicted that their recipients would get around 80% accurate, regardless of whether they were talking to the person or typing to the person. They didn't seem to appreciate that typing, AJ, such a hard worker, isn't that nonsense? They didn't seem to appreciate that their sarcasm wasn't coming through their fingers. So they didn't think that how the medium through which they were communicating mattered for how well they were actually communicating to another person, right? So this thing that's obvious in the abstract is not so obvious in practice. They were being egocentric. It also wasn't obvious to the recipients of this. The recipients who were 75% right when they heard what the person had to say, but were only at chance when they were reading, they actually thought they got 90% of these on average right in both the email and the voice condition. So they were the most overconfident and they were totally clueless about how the context in which they were communicating mattered for their ability. So here we see a problem of egocentrism, right? So, John, I'm teasing you about something. I think it's obvious I'm teasing you. I sent out this funny email, right? Then you, I think it's obvious, it's clear. You go off the handle because it's not obvious to you and you're pissed off. That's how friction can arise, for instance, from egocentrism, just one of many sorts of examples like that. Yeah, and we've seen the rise of the shorthand, the slash S, in written communication in these formats to denote sarcasm and the emojis and trying to add an extra color to what we're typing. What's fascinating to me is how poor we are at predicting what that future behavior or reaction will be in that moment. So there was a violent attack of someone in a car that was posted on Reddit. And if you actually look at the comments, the comments were all, man, I would have ran that person over. I would have got out of the car and gotten violent with them. I would have reacted in such a bold way based on what was happening to them. And the person replied to these people that in the moment, I froze. And we have this idea and this prediction in our mind that if someone is racist in front of us or someone is violent in front of us or someone gets in an accident in front of us, we in our altruistic state will rise to the occasion and be the most heroic version of ourselves. But that prediction is actually quite faulty in reality.