LLM Inevitabilism and the Ideological Capture of Strategic Reason

Three threads surfaced on Hacker News within the same week, and their simultaneity is worth pausing on. One dissected what its author called "LLM Inevitabilism" — the creeping assumption that large language models represent not just a promising technology but an inescapable historical trajectory. A second criticized what it termed "AI psychosis companies" — organizations restructuring themselves around AI adoption without credible evidence that the investment would pay off. The third, almost defiantly, declared that AI skeptics were simply deranged. Read together, these threads do not contradict one another. They are complementary diagnoses of a discourse that has stopped being a technical conversation and started being an ideological one.

Inevitabilism — the belief that a technology is not merely likely to transform society but fated to do so — is a recurring phenomenon in the history of technology. What makes the LLM variant particularly striking is how quickly it calcified, and how deeply it has penetrated institutional decision-making. Within three years of ChatGPT's release, expressing systematic doubt about LLM value propositions has become a kind of professional liability in certain circles. The "skeptics are crazy" declaration is not a logical argument; it is an excommunication notice. And the fact that it can be uttered without irony tells us something precise about where the discourse has arrived.

From Forecast to Creed: The Mechanics of Inevitabilism

The route from technical optimism to ideological orthodoxy follows a recognizable path. It begins with a genuine insight — in this case, that transformer-based language models demonstrate capabilities that earlier generations of AI could not match. That insight attracts investment, which attracts narrative, which attracts more investment. At a certain inflection point, the narrative detaches from its evidentiary base and becomes self-referential: the technology is transformative because people say it is transformative, because capital is flowing toward it, because careers are being built on it.

What distinguishes ideology from hypothesis is the treatment of counterevidence. A hypothesis about LLM impact generates predictions that can be tested and revised. An ideological commitment to LLM inevitabilism, by contrast, treats failures as implementation problems, timing problems, or the fault of skeptics who created a self-fulfilling prophecy by not believing hard enough. The explanatory scheme becomes unfalsifiable, and once it does, it stops being a tool for understanding and starts being a tool for social coordination — a way of distinguishing insiders from outsiders.

The philosopher Charles Taylor described something adjacent to this as "social imaginary capture": the moment when a particular vision of the future becomes so embedded in institutional practice and collective identity that questioning it feels not like intellectual challenge but like moral deviance. The "crazy skeptic" framing is a textbook instance. It performs the boundary work of an ideology under stress — asserting its perimeter precisely because it feels contested. When a belief system needs to pathologize dissent rather than answer it, that is a signal worth registering.

Strategic Capture and the Costs of Epistemic Closure

The practical stakes of this ideological drift are not abstract. When organizations internalize inevitabilism, several characteristic distortions follow in predictable sequence.

Attribution asymmetry becomes systematic. AI-adjacent successes confirm the thesis; AI-adjacent failures are attributed to execution, timing, or insufficient commitment. The technology itself is placed outside the explanatory frame. This is precisely how unfalsifiable belief systems maintain themselves in the face of mixed evidence — and it is deeply dysfunctional for strategic learning, because it eliminates the feedback loop that would otherwise allow organizations to update their bets.

Opportunity cost calculations break down next. The question "what else could this capital and attention produce?" becomes nearly unaskable, because inevitabilism transforms AI investment from a choice into an obligation. When adoption is framed as survival, the relevant comparison shifts from "AI versus alternative X" to "AI versus extinction." Under that framing, even marginal or negative-ROI deployments appear justified. The psychosis companies criticized on HN are not simply irrational; they are organizations that have correctly internalized the inevitabilist logic and followed it to its conclusion.

At the level of national strategy — where the stakes are largest — these distortions become potentially dangerous. States that accept the inevitabilist framing of an AI arms race tend to stop asking what kind of advantage they are actually competing for, and at what cost to other priorities. "We must not fall behind" is a reactive posture dressed up as strategy. It forecloses the more difficult questions: behind in what, relative to whom, and for what purpose? A country that cannot ask those questions is not pursuing a strategy; it is responding to someone else's.

The epistemically troubling feature of inevitabilism is that it is partly self-validating. Large capital flows toward a technology accelerate its development and adoption, which then appears to confirm the inevitabilist narrative. Skeptics seem to have been lagging a trend they helped slow. The feedback loop is real — but it tells us nothing about whether the technology is being deployed well, for good ends, at the right scale, or with adequate attention to second-order effects.

Skepticism as Strategic Infrastructure

To be precise about what is being argued here: none of this is a claim that LLMs are unimportant, that the technology is simply overhyped, or that skeptics are systematically right and enthusiasts are systematically wrong. The argument is structural. When the space for calibrated skepticism collapses — when asking "is this the right tool for this problem?" becomes a loyalty test rather than a legitimate strategic question — organizations and states lose the cognitive infrastructure they need to allocate resources well.

The actors who have historically navigated technological transitions most effectively are neither reflexive adopters nor reflexive resisters. They are the ones who maintained an optimistic hypothesis about a technology while continuously updating their picture of where it failed to deliver, where it created unexpected costs, and what conditions needed to hold for the optimistic scenario to materialize. That posture is not a psychological trait; it is an institutional design question. It requires that someone in the room be permitted to ask "but does this actually work, here, in this context, for this population?" without being read as obstructing progress.

The moment that question is treated as deviance — the moment the "crazy skeptic" framing takes hold and is not immediately challenged — is the moment strategy becomes theology. Theology is not necessarily wrong about everything, but it is a poor basis for resource allocation under uncertainty, which is precisely what strategic planning is.

The three HN threads that opened this column are not curiosities. They are data points about the current epistemic state of AI discourse. That the most visible response to criticism of AI adoption is to declare critics mentally ill rather than to engage their arguments suggests that the discourse has crossed a threshold. The technology may well be as transformative as its advocates claim. But if it is, that case should be makeable without pathologizing the people who ask for evidence. When it cannot be made that way, the problem is not the skeptics.