The article opens with a quote from filmmaker Nanni Moretti: “Whoever speaks badly, thinks badly and lives badly. We must find the right words: words are important!” It then relays a reader’s question about the future of tech writing and the author’s response that the most valuable skills are not necessarily AI proficiency or coding ability. Instead, the piece recommends returning to the humanities—studying philosophy, literature, linguistics (including its computational branch), mathematics, history, or anthropology—as a way to cultivate an edge in a world where AI can retrieve facts and assemble text in “reassuringly trite ways.” According to the article, the advantage for a “carbon‑based lifeform” lies in the ability to go deep, take a stance, and convey and defend it, which can be developed by training “in the dojo of Plato and Chomsky.” The author contrasts this with learning a trade such as plumbing or carpentry, suggesting that polishing an existing humanities degree or pursuing self‑study are viable paths. For engineers, adding a “bundle of Classics” to one’s skill set is presented as beneficial, although the article notes that degrees do not guarantee thoughtful thinking, only that some of the material may stick. The piece warns that a growing portion of knowledge work risks becoming a “polite exchange of statements between AI agents,” with humans reduced to “meat proxies” in the loop. To counter this, it advises taking a “long detour,” paving new paths, and generating original thoughts, even if they feel uncomfortable. While wiring systems together remains important, the article argues that for most developers technical knowledge alone—such as familiarity with Java boilerplate or Bash tips—is rarely enough. Likewise, many tech writers are not distinguished merely by mastery of Markdown or DITA. True professional excellence, the article claims, comes from the capacity to learn fast, think hard, and visualize whole systems, including their “hairy human factors.” It further notes that the foundations of modern AI draw from humanistic disciplines: philosophy of mind, philosophy of science, computational linguistics, and cognitive science have shaped AI’s conceptual base; logic and ontology underlie software design from databases to information architecture; user‑experience design leans on the arts and anthropological studies of tool use; and projects like NOPE translate clinical psychology into standards for safer AI behavior. Linguistics and literary theory, meanwhile, can help uncover the sources of effective documentation. In practical terms, the article says these human‑centered skills enable workers to read project proposals critically enough to detect “bullshit,” construct arguments that withstand disagreement, separate evidence from inference, and recognize when a wheel is being reinvented. Such abilities are needed to question AI output, design information architectures that make sense to humans, and explain products whose hardest problems concern ideas rather than technical challenges. The author also emphasizes language learning as a mental exercise: after studying basic Persian, they are now attempting Mandarin Chinese not to work in the language but to experience thinking in a radically different linguistic system, which acts as a “Trojan horse” carrying culture, history, and alternative worldviews. Greater exposure to diverse languages and cultural substrates, the article argues, eases navigation of complex problems. The final advice is to embark on a personal “Grand Tour,” cultivating a flexible, “underscore‑shaped” mind capable of acting as a proto‑diplomat, translator of cultures, and family philosopher. By becoming a humanist in tech—similar to the individuals AI labs are hiring—one can explore product dimensions from novel angles. The piece concludes by reframing the earlier lesson about finding one’s way in a library as learning to read the map of the world and then navigating it with the tools at hand, calling the rest “implementation details.”
