Likely this should apply the municipality/commune, but without the year it concerns, it’s not that helpful. I assume it’s acceptable to just remove it.
Some municipalities publish population data not only for the whole municipality, but for the individual places as well. Generally removing these without further investigation is not helpful.
Keeping until more precise data can be added is probably a better idea than removing.
Also, personally in think that the population key is not a good fit for OSM, and data consumers should load the data via WikiData when they need it. I’m nonetheless not removing that key…
I think isn’t a matter of precision, it’s just that the data lacks the year and is applied to the wrong OSM feature.
If we want better data, we could indeed try to add that to Swiss municipalities. French municipalities have that (from 2015, I think). So it would be removing and adding.
In case of the municipality Veyrier there are (as far as I can see) two villages that belong to that municipality: Veyrier and Vessy. It could be that municipalities with multiple villages/hamlets/boroughs/… publish population information for each of these, let’s say e.g. each 5000 for Veyrier (the village) and Vessy. That’s 10000 in total for Veyrier (the municipality).
So in that case, 10799 was probably wrong, as it was tagged on the village and not on the municipality. However, the currently tagged Wikidata item on the village node describes the municipality, not the village. That’s wrong as well.
That said, I agree with @habi that Wikidata is a better source to use for data consumers (provided that the correct Wikidata item is tagged).