> Agriculture isn't the only way to sustain moderately-sized populations, it's just the most efficient in the short term. Semi-passive cultivation is arguably more sustainable

There is a theory by Jared Diamond [1] of how pre-Agricultural Jomon people of Japan were able to invent pottery 18K years ago: the wild-food rich environment of Japan, being moist, temperate, and surrounded by ocean - allowed them to have very settled lives while other hunter gatherers had to constantly move.

Outside of rate examples like this, there is little evidence that non agricultural societies reached even moderate populations.

1. https://erenow.net/common/gunsgermssteel/23.php

Semi permanent sites did exist in this timeframe. But they were much smaller in terms of population and rarely (if ever) walled cities like those which existed around 4000 BCE.

Completely agree with you that agriculture is not required to form non-permanent settlements that didn't last tens of thousands of years.

A non-permanent settlement conceivably could see continuous (non-permanent) habitation for thousands of years. If, as a semi-nomadic group, you return to the same location year after year, generation after generation, and you do so alongside a large number of nearby settlements, with whom you maintain social relationships... Well, it might not have a post office or gate, but it's a dynamic fulfills many of the requirements of a "city".

No one said it had to be a single, year-round, uninterrupted settlement. A seasonal, interrupted, settlement would be no different to an asteroid.