A crystallographer who grows a crystal slowly — controlling temperature, managing saturation, allowing the lattice to assemble one molecular layer at a time — produces a structure of extraordinary internal order. The crystal is transparent because its regularity permits light to pass through without scattering; its faces are flat because each layer was deposited in alignment with those beneath it. Force the crystallization — supersaturate the solution, drop the temperature rapidly — and the crystal forms faster but with a degraded internal structure: voids, dislocations, boundaries where one region of order meets another at a misaligned angle. The crystallographer's dilemma is the relationship between speed of formation and structural depth, and the book uses this analogy to frame the question of what AI speed does to the depth of the formal sequences it fills. The analogy is imperfect — crystals do not learn — but the relationship between pace and order illuminates the structural question at the center of the AI age's effect on cultural production.
The analogy connects directly to Byung-Chul Han's critique of smoothness but provides the analytical specification his framework lacks. Han can describe the phenomenology of smoothness — what it feels like to encounter artifacts produced without friction — but he cannot specify what is lost structurally when the friction disappears. The crystallographer's dilemma provides the specification: what is lost is the process of entrance, the slow traversal of the sequence that builds the maker's structural understanding of its shape. The artifact may be identical in its formal properties to one produced through slow entrance, but the maker is different, and the difference matters for the capacity to produce the next artifact.
The analogy has a second dimension: not all friction is formative. Kubler was a scholar of pre-Columbian art, a field where barriers to entrance were primarily institutional, geographical, and economic rather than intellectual. The Mesoamerican potter did not face a deficit of productive struggle; she faced a landscape of constraints that prevented entrance into formal sequences altogether. The removal of access friction is not a loss but a liberation. The developer in Lagos who could not enter web application development because she lacked institutional infrastructure is not made shallower by the removal of that barrier; she is made possible. The distinction between formative friction (which builds structural understanding) and access friction (which prevents entrance) is the distinction that separates a rigorous critique of AI's effect on depth from a romanticized defense of difficulty for its own sake.
AI removes both kinds simultaneously, and that simultaneity is the source of the dilemma. The same tool that opens the sequence to the developer in Lagos also allows the developer in San Francisco to bypass the formative process that would have built her structural understanding. The same collapse of the imagination-to-artifact ratio that democratizes entrance also accelerates production to a pace that may degrade the depth of engagement. The crystal grows faster. Its internal order is compromised. The compromise cannot be separated from the liberation, because both are produced by the same mechanism.
The dilemma is, finally, a question about what kind of structure the AI age will produce. Kubler's formal sequences are cultural crystals — structures of linked solutions whose internal order determines their capacity to support further growth. A sequence with deep internal order can support a vast range of further development; a sequence with shallow internal order may appear complete without being deep, containing thousands of variations that are formally competent and structurally unconnected. The question is whether the ascending friction builders describe — the relocation of difficulty from the mechanical level to the level of judgment, vision, and structural understanding — produces a new kind of internal order that compensates for the order rapid crystallization degrades. The question is open. The answer depends on what makers do — whether they use the liberation from mechanical friction to engage more deeply with structural logic, or whether they allow production speed to substitute for the understanding that slow production once enforced.
The analogy is developed in chapter 6 of the book, drawing on crystallography as an analytical resource Kubler himself might have used given his preference for the physical sciences. The analogy extends Kubler's implicit concern with the relationship between pace of sequence development and structural depth, making explicit what his framework left largely implicit.
Pace and structural depth are related. The speed at which a formal sequence is filled affects the internal order of the resulting structure, in ways the analogy with crystal formation makes precise.
Friction is not uniform. Formative friction builds structural understanding through the process of entrance; access friction prevents entrance altogether. The first is valuable; the second is a barrier to be removed.
AI removes both kinds simultaneously. The tool that opens sequences to new entrants also accelerates production to paces that may compromise the depth of engagement — the compromise cannot be separated from the liberation.
Ascending friction is the open question. Whether the relocation of difficulty from mechanical execution to structural judgment produces a new kind of internal depth — or merely shallow density dressed as depth — depends on how makers use the liberation.
Structural order supports future development. Sequences with deep internal order support ranges of further variation that sequences with shallow order cannot; the long-term consequence of AI-speed filling is an empirical question whose answer is not yet determined.
The dilemma does not resolve by choosing speed or choosing patience. It resolves by understanding the relationship between them — and the practical resolution depends on empirical questions about which kinds of friction are formative, which are merely access barriers, and how makers actually engage with sequences when both have been removed. The answer will vary by domain and by individual practice; the book resists universal prescription and insists on the question as the load-bearing structure of the analysis.