In the book's lucid first part, "Science Fiction and Extro-Science Fiction," Meillassoux rigorously analyzes the difference between SF and XSF. In SF, he claims, "[I]t is a matter of imagining a fictional future of science that modifies, and often expands, its possibilities of knowledge and mastery of the real" (4- 5).
If Popper misunderstands the problem, Immanuel Kant, for Meillassoux, appropriately adopts an XSF imaginary in the form of the transcendental deduction to respond to it in the Critique of Pure Reason (1781); nonetheless, Kant fails adequately to develop it by confining it to the correlationist circle (i.e., that objects-in-themselves produce in us representations with which we are always already correlated).
Meillassoux persuasively shows how Popper, in privileging epistemology (i.e., experiments and theories), evokes a SF imaginary to deal with an XSF problem of ontology (i.e., the stability of physical laws).
In the compelling "Transcendental Deduction and the Three Types of XSF Worlds," Meillassoux introduces Kant's response to Hume's problem: transcendental deduction.