Well the essay has a lot to discuss, part of which is already (or will be) addressed up and down thread, so towards your TL;DR:

Yes of course, I'm not suggesting to disrupt circadian rhythms. And yes, lookup tables for solar days will always be required, but I would argue this is an inherent complexity to how we measure time in relation to our behavioural patterns and environment. However doing that by using variously large timezones that do not quite match solar days at their edges anyway, with a lot of them changing their offsets by an hour for half the year, and some of them using half-hour offsets throughout the year, that is complexity added for administrative reasons which are partly obsolete and largely irrelevant to the question off what would benefit humanity as a whole the most.

If everybody were to use one single timezone you would memorise your relative offset to noon/midnight pretty fast. Like it's one number to remember, e.g. where you are 4:40am is noon, 4:40pm is midnight, your offset is -7:20. Having those times be (roughly) 12 (for half the year) is just tradition and something we have every child learn. We could teach them about solar offsets just as well. It's not even really more complex, arguably much less so since you remove the need to confuse them with the chaos that global timezones have grown to be historically.