So I had to look up the Boltzmann constant and... That's a lot of math.

I think you have a point on the decreasing human temperature. It looks like the decrease is at 0.05°F every decade, which actually is quite a bit. If it was something like 0.005°F, I'd say that that's a problem for the people of the year 2500 to solve.

That said, the reason it's been decreasing seems to be due to medical advances and not some change in the Earth's gravity or climate change. I would be surprised to see humans in the year 2500 having an average body temperature of 72.9°F, or closing in on 0°F in the year 3,984. I imagine there will be fluctuations, but there's got to be a lower limit to what is physically possible.

I'd still defend the Celsius number, since even though there are changes due to air pressure, it's changing over space and not time. In the year 2500, water at sea level will still freeze at 0°C.

I think my big thing is I'm less concerned about a logically consistent scale, and more towards a scale that's geared to the emotional side of temperature.

Thinking outloud moment

If we are going for the emotional side of temperature specifically, we would also need to factor in wind, humidity, sunlight, what season it is, etc. and that's a lot of variables, and even then that's how you get the wind-chill factor. But even that is almost completely subjective. I feel like that scale would go from "IT'S GOTTA BE NEGATIVE A MILLION FUCKIN' DEGREES" to "I FEEL LIKE IM ON THE SURFACE OF THE SUN, so like a bazillion degrees" and then we go to the traffic report.

Either way, it's not a perfect scale, but I'd still take that over the other two.