In the examples where Hoffman claims there is always an optimal survival function that is better than a reality-based survival function, it is not at all clear

In the examples where Hoffman claims there is always an optimal survival function that is better than a reality-based survival function, it is not at all clear to me that the optimal function is not actually a more realistic function than the allegedly realistic one. At best, the author's case seems to be that a) all mental representations are inaccurate to some extent (they are all approximations, at least), and b) a representation that is good for survival need not be accurate in aspects that do not matter for survival. To go from there to say that there is no correspondence between mental representations and reality seems to be a huge and unjustified leap, and to go from there to there being no reality seems to be a second and even larger stretch.