Eh, I guess it's still better than the (thankfully very rare outside of deliberate glitches) error Nintendo consoles give you now. "The software closed because an error occurred". Thanks for that useful piece of advice.
I'm sure I'm sitting on the wrong side of countless user studies that say to the contrary, but when an exception or problem occurs, I really dislike the feeling of helplessness. Something isn't working the way it is supposed to, and often times I can only rely on myself to fix it; this is particularly relevant for companies with non-existent support services, like Google, where you're basically at the mercy of what other more adventurous and capable users are able to fix on their own. To me, tossing up a "funny" error message when there is no support channel is incredibly frustrating, because not only is there no one I can ask for assistance, I can't even do research myself because I have nothing to go on except for a caricature of the WebDevs or a Frowny-Face and the steps leading up to the problem.
Again, I'm sure I'm on the wrong side of countless studies here, but I can't figure out how this is better than at least some sort of error message.
This is completely off topic, but I was a UX study participant last year, it was an interesting experience. They told me to do as I usually do when surfing the web, and to verbalize my thought process.
I opened the website, and a huge nonmodal advertisement to win some $15 piece of garbage pops up. On pure instinct I click ctrl-w to close the tab, and state "what an obnoxious advertisement, I am boycotting this site".
It went downhill from there.
I also dislike the first person error messages. "We are sending ...", "Let me fix this" when you are clearly 'talking' to a computer.
The subsystem that understands the error doesn't have the context to tell the user what to do about it, but the subsystem with context doesn't understand the failure mode in enough detail to actually give the right advice.