I'd like to believe that certain processes have certain benefits, but I don't think we should accept that without some kind of evidence.
As my article indicates, pairing puts a drain on my mental capacity, and I highly doubt it could yield a more creative solution, if anything it'd limit my freedom in the process. This may vary based on experience in position however, as I indicated.
As to code quality, I see a lot of places that indicate this, yet I'm not seeing any kind of definitive proof of this. There are many ways to improve code quality. I doubt having multiple eyes watching the original coding inherently makes it any better.
Why would you think that pairing increases the time a developer can focus on a problem? If this is anecdotal evidence then I can offer my own anecdotal evidence that I've known many developers that can focus for very long periods of time on their won. I've indeed known many that have no ability to focus unless on their own.
I don't want to dismiss your points, as I suspect in some cases they are likely true. There are certainly people that can work well together, and pair programming is probably a good option. My overall concern with the approach is that descriptions of it are not well balanced, and fail to provide the bounds within which it is effective.
We’re a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.