Everyone has a limit; some much greater than others.
I live by tangible proof and am data driven.
When UOAs are this good, and you can have independent verification (such as visual sludge checks) then I see everything pointing towards OCI extension.
What would hold you back in this case? Nothing but unverifiable emotion. You said it yourself: " ... This whole thing feels ..." You're ignoring data and visual verification, and running on your feelings (emotion). There is nothing wrong with that for you. Frankly, we all have emotional trigger points for many different topics, and that includes me. But I don't try to hide or otherwise obscure my emotion with lame justification or excuses (not saying that you or the OP is doing that, but many BITOGers are guilty of that practice).
If someone wants to run syn with an M1 filter and OCI every 3k miles, that is their decision to live by. But that does not make it a smart practice.
When there is evidence such as this UOA, showing long OCIs are safe and longers ones are possible, and you can confirm this with additional verification, one must admit that not doing so is a matter of emotion and not intellect.
Everything quantifiable in this UOA is in very good control. There is logical reason to add in both TAN and visual verification (if easily done) of no sludge if one proceeds towards longer OCIs. While not a necessity, it can give more "comfort" in the decsion to extend.
We are just beginning to see here that some members are actually coming into a new light; that OEM OCIs are artifically limited as a means of ultra-safe warranty protection. When folks step outside that box and experiement, it's not turning nearly as ugly as the pundits would profess. My UOA; this UOA; 2010_FX4's UOAs; there is mounting evidence that supplements and completely is in agreement with both my macro UOAs studies, and the SAE study showing OCI extensions are generally much safer, and result in less wear. There are always going to be "oops - darn it" examples such as SL2 and some Toyota sludgers. But should those examples be a reason for the rest of us to never get off the porch?
I would NEVER, EVER suggest someone simply triple their OEM OCI blindly, without knowing any situational conditions surrounding the use and the equipment. But once UOAs are done, and visual clues are examined, then what in the world can be the sane reason for not extending? When you do the experiment, run the data, and confirm the results, why ignore what you spent time and effort to find out?