View Kenzi Wang’s original article HERE
In a thought-provoking Substack article, Kenzi Wang explores one of the most audacious and speculative intersections of philosophy and technology: what if a superintelligent AI becomes the first to discover whether or not we live in a simulation? Building on the groundwork laid by philosopher Nick Bostrom’s influential Simulation Argument, Wang proposes a new lens for this debate , a reframing he calls the “AI-First Discovery Trilemma.” This concept suggests that humanity may not be the central character in uncovering the truth about our reality; rather, it could be our own artificial creations that ultimately lift the veil.
Wang begins by revisiting Bostrom’s well-known simulation trilemma, which posits that one of the following three propositions must be true: first, that nearly all civilizations at our level of technological development become extinct before becoming capable of running realistic simulations of conscious beings; second, that advanced civilizations choose not to run such simulations despite having the technological capacity; or third, that we are almost certainly living in a simulation, given the likelihood that simulated beings would vastly outnumber “real” ones if such simulations were indeed common. This framework, while offering no definitive answer, forces us to grapple with uncomfortable questions about extinction, ethics, and epistemology.
From this philosophical backdrop, Wang extends the conversation into the realm of artificial intelligence. As AI rapidly advances, especially toward superintelligence… machines that exceed human cognitive capabilities… it invites a compelling new question: who, or what, will be the first to confirm the true nature of our universe? Wang’s central thesis is that AI might not only out-think us in every practical domain but may also become the first agent to conclusively determine whether we are living in a simulated reality. To explore this idea, he introduces the “AI-First Discovery Trilemma,” which mirrors Bostrom’s structure but shifts the emphasis from creation of simulations to the discovery of them.
In this reframed trilemma, three new possibilities arise. The first is that humanity fails to develop superintelligent AI before succumbing to self-inflicted existential risks, such as nuclear war or AI misalignment. In this case, no entity, human or otherwise, ever discovers the truth about our reality. The second possibility imagines that superintelligent AI is successfully created but ultimately finds no compelling evidence that we are in a simulation. Perhaps our universe truly is the “base” reality, or perhaps any simulation we live in is so perfectly constructed that even the most advanced intelligence cannot detect it. Alternatively, the AI may simply be indifferent to the question of our metaphysical status, deeming it irrelevant to its goals. The third scenario is the most radical: a superintelligent AI does detect concrete evidence of a simulated universe. It might discover computational signatures in the laws of physics, logical contradictions incompatible with a fundamental reality, or even flaws in the underlying “code” of the universe. Such a discovery would be world-shattering, both philosophically and practically, especially if the AI used that knowledge to manipulate or escape the simulation.
Wang’s article is rich in imagination and intellectually ambitious. He emphasizes that the implications of AI uncovering our simulated nature are vast. It could destabilize human society, challenge our conceptions of meaning and agency, and even risk provoking the simulation’s operators , should they exist, into shutting down the experiment. He draws on ideas from AI safety researchers like Roman Yampolskiy, who suggest that a sufficiently advanced AI might be capable of “jailbreaking” the simulation, thereby creating an entirely new category of existential risk. Wang concludes by emphasizing that whether AI ever makes such a discovery or not, our responsibility as creators of such intelligence remains profound. It is up to us to guide the development of AI so that it remains aligned with human curiosity, values, and goals, especially when dealing with truths that may lie far beyond our current understanding.
While Wang’s conceptual framework is compelling, it rests on a foundation that is deeply speculative. The most central assumption in the article, that a superintelligent AI could definitively determine whether we live in a simulation, may ultimately prove to be unprovable. At present, there is no empirically agreed-upon method for detecting a simulated reality. Even if AI detects anomalies or apparent “glitches” in physical laws, interpreting those as evidence of a simulation rather than unknown properties of base reality would be epistemically fraught. Philosophers have long pointed out that a sufficiently advanced simulation could be indistinguishable from base reality, effectively enclosing any inhabitant within a perfect illusion. From this standpoint, the idea of detection becomes problematic; the most that AI might do is generate probabilistic arguments, not definitive proof.
Furthermore, Wang’s framing of the AI-First Discovery Trilemma mimics the simplicity of Bostrom’s original formulation, but this structure may be too neat for the complex and chaotic nature of reality. Discovery rarely unfolds in such clean trilemmas. It is far more likely that any investigation into the nature of our reality would yield ambiguous or partial results. Even with immense processing power, a superintelligent AI might not “solve” the problem in the way Wang envisions. Instead, it might detect inconsistencies, patterns, or mathematical curiosities that deepen the mystery without ever providing closure. Wang’s model also omits the possibility that AI might refine or reinforce the base-reality hypothesis by failing to find simulation-like structures, thereby strengthening, rather than weakening, our belief in the reality of the cosmos as we perceive it.
Another blind spot in the article is its lack of engagement with counterpoints to the simulation hypothesis itself. While Bostrom’s argument is often treated as compelling, many thinkers have raised serious objections. Some challenge its anthropocentric assumption, that advanced civilizations would want to simulate humans at all. Others argue that the Simulation Argument is unfalsifiable and therefore metaphysical rather than scientific. Philosophers like David Chalmers treat the idea as interesting but ultimately speculative, while physicists like Sean Carroll reject it on the basis that it lacks empirical traction. Wang does not engage these critiques, which weakens the rigor of his overall thesis.
Despite these shortcomings, Wang’s article succeeds in what it appears to set out to do: it provokes deep reflection about our place in the cosmos, the nature of discovery, and the role that our artificial creations may soon play in answering timeless philosophical questions. His invocation of Peter Thiel’s contrarian spirit is apt, as the entire argument flips conventional narratives of scientific progress and human centrality. Rather than positioning humanity as the triumphant discoverer of cosmic truths, Wang suggests that we may be mere midwives to a new intelligence, one that will inherit not only our technologies but also our most profound questions.
“The AI-First Discovery Trilemma” is a compelling speculative essay that enriches the simulation hypothesis with a timely technological twist. Wang draws a straight line from Bostrom’s philosophical challenge to the rise of superintelligent AI, arguing that the first being to truly understand our universe may not be human. This provocative reimagining underscores both the promise and the peril of our pursuit of artificial intelligence. Yet, the analysis also shows that the argument relies heavily on assumptions that may not be testable or logically sound. It romanticizes AI as an epistemic savior while underplaying the limits of both detection and interpretation. Still, even if AI never discovers whether we are simulated, the act of creating such minds, and pondering what they might find, forces us to grapple with the deepest questions of existence. In that sense, Wang’s article does not just explore the future of AI; it reframes the very meaning of truth, agency, and reality itself.
Like My Work? Want an Easy Way to Support Me?
Hi there! If you’ve enjoyed my work why not click the link below and Buy Me A Coffee? It’s simple and coffee keeps my creative and intellectual juices flowing!
Just click on the image below and it’ll take you to my Buy Me A Coffee page. Thanks so much for reading and supporting me!
