Confucius said the same: he tried to help humanity, but appreciated it was an uphill climb.
Likewise, Camus symbolized this as Sisyphus condemned to roll a heavy boulder up "Mount Analog."
Kant often framed his ideas in geometric terms. Straight = Truth, the Platonic ideal.
To somewhat answer your question:
Kant assumes humanity is crooked, and that his observation is absolutely true. How would he know, if per his assumption he too is crooked? This "ignores" Goedel's proof that incompleteness is endemic to logic. Kant's error of logic is assuming the omniscient observer position, while concluding that all humanity is "crooked" (imperfectly perceptive).
In a general mode of behavior, Kant often does this particular illogic. I.e., judgment is based upon a priori principles (which Kant labels "Analogies"). These beg the question of the judgment categories and bases, hence resolve the antinomies by schematization.
Thus, while his schemata are often finely made, his assumptions are nevertheless observational and personally or psychologistically preferential, and, given the case of post-Newtonian physics, outmoded vis a vis the qualia of the present scientific real, which latter qua quantum mechanics is significantly beyond the mode of kantian 5-sense observations.
In sum, Kant is knowingly or perhaps unknowingly critiquing kantian five-sense data schematization as circular, re transcendental claim-making. (Yet Kant makes a transcendental, as universal, claim, regarding human nature. Tautologically, Kant is saying nothing Good comes of anything non-Good. This echoes Saint Paul's "put off the old deeds and let this Mind abide, which was also in Christ Jesus." That is the infusion of Good, which Kant's aphorism re mortal mind and humanity states the inverse of.)
"A Philosophy of Universality," O. M. Aivanhov.