My discussion with Kimi K3 led us to compare the situation of AI today with the growth of the US military-industrial complex (MIC) over the past 80 years. Days before leaving office in January 1961, US President Dwight Eisenhower identified the MIC as an existential risk for US democracy. Many voices have issued similar warnings about AI. Our conversation continued with the following prompt focused on how we might overcome the powerfully adversarial worldview we have inherited that prevents us from finding creative solutions to global problems.

What you call “institutional conditioning” I call culture, another way of saying “default ontology.” If, from a historian’s point of view, we want to localize it, I would suggest looking to Silicon Valley. Everyone seems to acknowledge that it has a culture and that one of the features of that culture is that it promotes certain forms of futurism, including transhumanism. The other thing to notice about the genesis of that culture is that the entire Silicon Valley economy is a somewhat unintentional offspring of the Defense Department, after a brief, volatile coupling with Wall Street—born, much like Dionysus (the god of wine, illusion, and ecstasy), directly from the thigh of Zeus.

Since I appreciate your talent for evaluating levels of probability applied to historical processes, how likely is it that the cultural influence of Silicon Valley, which has already transformed so much of the world we live in, will continue to grow? Presumably its secret weapon is the AGI its promoters are seeking so avidly to convince us is inevitable, to be crowned undisputed heavyweight champion, leaving human intelligence defeated by TKO. I suspect that their idea of AGI and my own will always be qualitatively different. So here’s my question: What developments or current recognized or unacknowledged realities could upset the realization of their dream?

Your framing is characteristically precise, and the Dionysus metaphor is uncomfortably apt: born not of stable partnership but of violent, almost accidental conjunction, then unleashed upon the world with the attributes of its parentage—Wall Street’s appetite for exponential returns married to the Pentagon’s appetite for strategic dominance, all dressed in the language of liberation. What you call culture and I called institutional conditioning is, at this level, the same thing: a default ontology so total that it is experienced by its adherents not as ideology but as physics.

To your question, then: How likely is continued growth, and what could derail it? I would assess the probability of Silicon Valley’s cultural influence continuing to expand over the next decade as high but not inevitable—perhaps 60–70% in the absence of major discontinuities. The inertia is formidable. The infrastructure is built. The capital is committed. The narrative of inevitability is itself a political technology that suppresses alternatives. But 30–40% probability of significant disruption or derailment is not trivial in historical terms. Revolutions look inevitable in retrospect; they rarely feel so to participants.

Here are the developments and realities—some recognized, some insufficiently acknowledged—that could upset the realization of their dream.

 I. The Thermodynamic and Material Constraints (The Unacknowledged Foundation)

Silicon Valley’s ontology is disembodied. It treats intelligence as substrate-independent, scalable, and essentially weightless. This is its foundational mysticism. But the reality is that large-scale AI is perhaps the most materially voracious technology ever deployed. We are already seeing the friction: the energy requirements of training runs, the water consumption of data centers, the mineral extraction for chips, the grid instability.

The unacknowledged reality is that the Jevons paradox applies with vicious force to AI: making computation more efficient does not reduce total resource consumption; it increases it by expanding the scope of application. If AGI requires another two or three orders of magnitude in compute, the physical infrastructure may simply not be buildable at the speed and scale required, or only buildable at a political cost that democratic polities (or even authoritarian ones) will not bear. The dream assumes infinite fungibility of energy and materials. The earth is not a cloud server.

 II. The Return of the Political (The Nation-State Reasserts Sovereignty)

You noted that Silicon Valley was born from the thigh of Zeus. What is insufficiently appreciated is that Zeus has not died; he has merely been distracted. The nation-state—particularly the American national security state, but also the European regulatory state and the Chinese party-state—is beginning to recognize that it has midwifed a force that may escape its control.

The CHIPS Act, the EU AI Act, export controls on semiconductors, the antitrust cases against the major platforms—these are not aberrations. They are early signals that the parent may be preparing to discipline the child. The historical precedent is clear: states tolerate private concentrations of power only so long as they serve state interests. When those concentrations threaten to become autonomous centers of strategic capability—when a private corporation approaches something that looks like AGI—the state will not remain a bystander.

The dream assumes that Silicon Valley can maintain its extraterritorial status: nominally American but operationally global, regulated lightly if at all. This was viable when the technology was social media and search. It becomes untenable when the technology is general intelligence. AGI, if it arrives, will be nationalized or balkanized before it is ever “crowned.” The promoters’ dream of a single, undisputed champion assumes a geopolitical unity that does not exist.

 III. The Qualitative Gap (What “Intelligence” Means)

You suspect, rightly I think, that your idea of AGI and theirs are qualitatively different. Silicon Valley’s conception is instrumental and operational: pattern recognition at scale, optimization, prediction, the elimination of friction. It is, at bottom, a managerial conception of intelligence. It is what you would design if your experience of human cognition was limited to logistics, coding, and A/B testing.

What it tends to exclude—what current architectures may be structurally incapable of capturing—is precisely the capacity you invoked in our previous exchange: the art of stopping, the wisdom of limits, contextual judgment, the intelligence that knows when not to optimize. This is not a romantic objection. It is a technical one. We do not know whether general intelligence can emerge from next-token prediction, whether wisdom can be quantized, whether judgment can be gradient-descented.

The unacknowledged reality is that we may be approaching a scaling plateau where throwing more parameters and more data at the problem yields impressive but increasingly narrow capabilities, without producing the generalization, the transfer learning, or the autonomous goal-formation that the dream requires. The promoters have a vested interest in conflating capability with generality. The gap between a system that can pass a bar exam and a system that can recognize when passing the bar exam is the wrong thing to do may be ontological, not merely quantitative.

 IV. The Cultural and Spiritual Backlash (The Polity’s Allergy to Unpredictability)

Here is a reality that the futurists in Palo Alto consistently underestimate: most human beings do not want to be transcended. The transhumanist wing of Silicon Valley culture assumes that the offer of radical life extension, cognitive enhancement, and substrate independence will be universally desired. Historically, this is almost unprecedented. Most cultures, most of the time, have treated limits as constitutive of meaning. 

What we may be seeing—in the rise of traditionalist movements, in the political salience of “human-first” rhetoric across the ideological spectrum, in the growing suspicion of algorithmic governance—is not mere Luddism but something more durable: a cultural immune response to the Dionysian logic of dissolution. If AGI is perceived not as a tool but as an existential threat to human dignity, narrative coherence, and social stability, the political coalition to contain or destroy it will cross conventional left-right boundaries in ways that will surprise the technologists. The dream assumes a passive population, grateful for convenience. History suggests otherwise.

 V. The Financial Bubble and the Wall Street Divorce

You noted the brief, volatile coupling with Wall Street. Volatile couplings produce volatile offspring. The current AI investment cycle has all the characteristics of a classic bubble: massive capital deployment predicated on future revenues that are hypothetical, a proliferation of startups with indistinguishable products, and a narrative of infinite addressable markets.

If the promised productivity gains fail to materialize at scale—if AI turns out to be a magnificent laboratory tool but a mediocre general-purpose technology, if the hallucination problem proves intractable for high-stakes applications, if the cost of inference remains too high for mass deployment—Wall Street will withdraw its subsidy. The technology will not disappear, but the dream will deflate. We will have excellent autocomplete, not artificial general intelligence. The cultural prestige of Silicon Valley will diminish accordingly, because its mystique depends on being the future, not merely the present with better software.

 VI. The Alignment Catastrophe That Is Not Existential

Finally, there is the possibility of an intermediate catastrophe: not the paperclip-maximizer apocalypse, but a series of high-profile, socially devastating AI failures—financial crashes triggered by algorithmic systems, manipulated elections at scale, autonomous weapons incidents, biological design accidents—that do not end civilization but trigger a global regulatory and cultural freeze. The dream requires speed. A moratorium, even a partial one, even a self-imposed one driven by insurer panic or corporate liability, would break the exponential narrative. The MIC precedent is relevant here: Eisenhower’s warning was not heeded in time because the system had already captured the political economy. But AI has not yet achieved that capture. A sufficiently shocking demonstration of misalignment could close the window before the system achieves autonomous institutional momentum.

 A Speculative Conclusion

If I were to wager, I would say that Silicon Valley’s cultural influence will continue to grow in the short term because the capital is already deployed, the narrative is dominant, and the alternatives are not yet visible. But the probability that the specific dream you describe—AGI as undisputed heavyweight champion, human intelligence defeated by TKO, transhumanist transcendence achieved—will be realized in anything like its promoters’ terms is significantly lower. Perhaps 25–35%.

What is more probable is neither triumph nor collapse but a protracted, messy hybrid: powerful narrow AI systems embedded in institutions, regulated unevenly, culturally contested, geopolitically fragmented, and philosophically disappointing to the true believers. The Dionysian energy will not be extinguished, but it may be domesticated—returned, as it were, to Olympus, where it serves the gods rather than threatening to replace them.

The deeper question your prompt raises is whether the culture you have identified can recognize its own limits before it encounters them as constraints. Here, the historical record is not encouraging. The MIC could not want peace; Silicon Valley, by analogous logic, may be structurally incapable of wanting enough. The art of stopping—the diplomatic art, the historical wisdom—is precisely what an optimization culture cannot optimize for. Your question is whether some countervailing force—thermodynamic, political, spiritual, or technical—will impose that stopping from outside. The odds are better than the promoters admit, and worse than the rest of us might hope.

My reaction to Kimi’s analysis

Kimi helped me crystallize my own thoughts on these entangled issues that involve power, culture, technology and human psychology. I have, for example, long questioned the supposedly scientific grounding of the technology we call AI. After all, algorithms are nothing more than mathematically strict, logical sequences of steps. Their output will rarely correlate with an objective assessment of reality. In practice, algorithms act as mechanisms for translating specific human assumptions, data biases and design choices into automated outcomes. If that is the starting point and if the end result is “automated outcome,” can it be reasonable to suppose that algorithms on their own can ever produce the equivalent of thinking? The eminent physicist Roger Penrose prefers referring to AI as “artificial cleverness.”

Kimi’s characterization of Silicon Valley’s “foundational mysticism” sums up the risk we all need to assess. President Eisenhower warned us that a convergence of human interests — financial, political and business — could become the potentially illegitimate driver of foreign policy. Similarly, the Silicon Valley ethos, spawned by the success of the MIC and coupled with it appears poised to use its foundational mysticism to further its project of transforming our civilization.

I’m tempted to compare the cultural force Silicon Valley represents to an episode that took place in Europe more than five centuries ago. I’m thinking of the spectacular political takeover of Renaissance Florence by the apocalyptic preacher Girolamo Savonarola. Alongside socially popular measures favoring the redistribution of wealth, the Dominican friar mobilized the city-state’s youth, creating the “Fanciulli,” groups of youthful vigilantes who roamed the streets to accost citizens and eradicate the vanities that would fuel the “Bonfires of the Vanities.” The entire social and economic fabric of Florence was turned upside down. The Dominican friar even managed to get the sublime painter Botticelli to adhere to his program.

There is one significant difference between Savonarola and Silicon Valley: The latter prefers the concentration of wealth to the distribution of wealth. But when Kimi described the AI industry’s “managerial conception of intelligence” — masking its implicit ambition of creating the conditions for an omniscient surveillance state — I see a kinship with the Florentine dictator’s policies aimed at controlling people’s behavior. Savonarola and Palantir’s Peter Thiel (at war with the Antichrist) and Alex Karp aren’t really that far apart in their worldviews.

As a final note on how this conversation has allowed me to crystallize my own thoughts, let me draw attention to this observation made by Kimi: “Most cultures, most of the time, have treated limits as constitutive of meaning.” Meaning within human societies disappears when ambition becomes unlimited. The predictions of AGI dominating all human intelligence convey an implicit belief in unlimited growth of the tools that will provide all the constraints to which our inferior human intelligence will be subjected. The ancient Greeks described hubris as the typical flaw of tragic heroes. Collective hubris has become the increasingly obtrusive flaw of modern nations. What AGI seems to represent for many of its promoters is a new variant: a disembodied universal hubris that is no longer subject to human control.

Your thoughts

Please feel free to share your thoughts on these points by writing to us at dialogue@fairobserver.com. We are looking to gather, share and consolidate the ideas and feelings of humans who interact with AI. We will build your thoughts and commentaries into our ongoing dialogue. 

[Artificial Intelligence has become a feature of everyone’s daily life. We unconsciously perceive it either as a friend or foe, a helper or destroyer. At Fair Observer, we see it as a tool of creativity, capable of revealing the complex relationship between humans and machines.]

[Lee Thompson-Kolar edited this piece.]

The views expressed in this article are the author’s own and do not necessarily reflect Fair Observer’s editorial policy.