Right, but in order to convert lat/lon differences into a straight line/great circle distance, one must first agree upon the shape of the Earth (you refer to that yourself as "knowing how far apart lat/lon points are"). Using it the way you propose would reduce your argument to "the Earth is round, therefore these are the distances, therefore the Earth is round"
This is an excellent formulation of what I call the
circularity objection. You hear it many times from Flatearthers: the model RE is using to prove some point itself assumes an RE model, so the proof is circular, or question begging.
Now it’s true that if I worked out a distance between two lat/lon coordinates by using spherical assumptions, then used that distance to ‘prove’ RE, that would be highly question-begging.
But what if I work out the distance in another way? The original way that (short) distances were accurately worked out was by a chains, literally a chain of links spread out on the earth’s surface like a big tape measure.
First question in this debate: do we all accept that a measurement essentially like a big tape measure, perhaps suitably adjusted (I will come on to that later) is indifferent to the shape of the earth, and distances based on this are not circular?
If not, what other measure of distance would you use? Car odometers are actually the same form of measurement – a bit like rolling up the tape measure for every turn of the wheel and reusing it.