What should we use instead?
Political and historical analysis? Just not simple extrapolation of a single variable. That seems woefully unconvincing, when it is itself the outcome of the complicated contingent process that is human history.
That point is part of my objection to Stephen Pinker's Panglossian thesis (that he's been peddling for some time now, when not hanging out with Epstein). He does things like collect stats to show violence has 'declined' over some recent period of time and then imply that means it will necessarily continue to do so. Given that it would only take a few nuclear weapons going off for that 'trend' to suddenly disappear in a mushroom cloud, it just doesn't seem a convincing argument, to me.
(It also seems to depend on how one defines 'violence', seems as if he only counts certain forms and ignores "the violence inherent in the system" - which may be a Monty Python quote, but still seems a valid point, to me - were the deaths at Bhopal not 'violence'? Or all the deaths from Fentanyl overdoses? Or deaths from extreme weather events due to climate change?)
I'd be a lot more convinced if I had, rather than some selective statistical trends, some ideological/political faith in a way of organising society and the economy that didn't have obvious massive flaws, that seemed likely to undo it.
That's the dream that socialism once seemed to offer (and libertarians and anarcho-capitalist types, and even the neo-liberal movement seem to have had their analogous 'grand solution to everything' - just none of them seem convincing to me at this point. Neo Liberalism seems to have failed entirely at it's original purpose, of ensuring 1930's type "Strongmen" never come to power again).
[Edit] This might just be low-serotonin levels talking.