As a developer, i suffered burnout all the time. I would get too emotionally vested into the code and work late to finish it. With Pivotal Labs, we coded in pairs, and it was considered rude to work on code outside of your pair. This plus other techniques like putting strong emphasis on emotional safety created an environment where burnout just disappeared.
Having to communicate the entire time your thought process, not really being able to just take the wheel and investigate things quietly for a few minutes, and if you're letting them drive, getting them to go to the right spots in the code and modify the right things (especially if english is their second language) not easily being able to get up and walk away from the computer if you need to ruminate on something for a moment...yeah, just fries my brain.
Still need to do it at my work, since I'm managing junior developers right now that just don't see some of these things without pair programming, but it's exhausting.
Part of that is the coaching junior developers that need extra handholding, but I've found it almost as tiring even amongst peers of equal skill, if it happens for an extended period of time.
Maybe its like furniture movers... it takes time to build up that endurance and mastery to get to the point where it is possible to do for 3 - 4 hours at a time. Pivotal was also strict with taking 15 minute breaks... one in the AM around 10:30am and the other in the afternoon around 3pm.... but each pair can decide on their own.
When I was a team lead in pairing teams, I did allow (and encouraged solo time), but the code written during solo would be consider spike code... code to learn and try stuff out. Ultimately to make it into prod code, it had to go through a pair.
Who wouldn't? I get tired after 1.5 hours of programming solo.
>not really being able to just take the wheel and investigate things quietly for a few minutes
It's not disallowed. I do this all the time when pairing.
I have a suspicion that people who don't like pairing have more of a visceral rather than intellectual dislike.
I wonder how many mathematicians would put up with something like that. Full-time pair proving.