I like the animations, and the site design is wonderful.
I have one nit in general about thinking of evolution as an optimization problem or hypothesis space—it assumes there’s some optimal solution to a given function, but that’s not necessarily the case.
There are optimal thermodynamic states at some scale for an organism, but at another level there are only stochastic processes which are in flux with other parts of a dynamic system. For genes in particular, there’s no optimization per se, but different types of pressures which drive population dynamics depending on the system state.
Maybe this post didn’t mean it this way, and I am reading too much into it. I think the context-dependent agent action described in the post is more in the realm of game theory than evolutionary genetics.
nis0s•6mo ago
I have one nit in general about thinking of evolution as an optimization problem or hypothesis space—it assumes there’s some optimal solution to a given function, but that’s not necessarily the case.
There are optimal thermodynamic states at some scale for an organism, but at another level there are only stochastic processes which are in flux with other parts of a dynamic system. For genes in particular, there’s no optimization per se, but different types of pressures which drive population dynamics depending on the system state.
Maybe this post didn’t mean it this way, and I am reading too much into it. I think the context-dependent agent action described in the post is more in the realm of game theory than evolutionary genetics.
By the way, this may also be of interest https://en.m.wikipedia.org/wiki/Evolutionary_game_theory