I find it funny that what people are calling Atlas is basically the polar opposite of what I'd understood Atlas as from the OP, e.g. what resonated with me about the idea of discussing GM effort.

For me, the extreme examples of in game GM effort all have to do with things where the system obligates you to perform tedious multistep computation or resolution in order to run things, or where running NPCs to a reasonable percentage of their supposed abilities requires a lot of problem solving or optimization or planning.

For example, grapple mechanics to me are 'high in-game GM effort'. Or detailed hitpoint tracking for hundreds of units in mass-battle systems that don't do abstraction well. Or motion planning for ships or low maneuverability fliers.

Making stuff up on the fly on the other hand is low to zero effort.

The reason I introduced that extra category is that, no, that doesn't mean free form is always best, because some things which require effort give you a return on that effort. Whereas other things like tracking HP individually for every soldier in an army are basically wasted effort - you can do it, but there's no real benefits commensurate with the time it costs you.