Garland, Alex

writer and director of "Ex Machina," "Annihilation," and "Devs"

Alex Garland: Ex Machina, Devs, Annihilation, and the Poetry of Science

Lex Fridman Podcast · Mar 3, 2020

• Alex Garland
• Ex Machina
• Annihilation
• Devs
• artificial intelligence
• consciousness
• determinism
• free will
• Silicon Valley critique
• alien intelligence

Core ideas

Garland distinguishes his view that we live in a perceptual "dream state" (since the brain constructs a best-guess subjective model of the world from incomplete sensory data, making all human experience inherently a kind of controlled hallucination) from the separate claim that reality is a simulation, which he rejects — he believes in a genuinely physical, non-simulated universe, just one we never perceive directly or objectively. He extends this to psychedelic drugs, arguing they don't reveal an alternate "true" reality but simply add a different distortion on top of the already-distorted baseline of ordinary perception, amplifying whatever emotional undercurrent (anxious or calm) a person already carries. Discussing "Annihilation," he explains his goal of imagining a genuinely alien intelligence stripped of the usual anthropomorphizing habits of science fiction (aliens as invaders, teachers, or resource-seekers) — deliberately leaving its motivations, sensory apparatus, and even its relationship to physical laws undefined, since assigning it any human-like goal (even something as apparently neutral as "biologist" or "explorer") would undermine the thought experiment; he speculates it might operate on a vastly slower timescale, reacting to its environment the way an octopus changes color, without any of that reactivity implying comprehension of what it's reacting to.

On the relationship between science and his own creative work, Garland firmly rejects the framing that his films inspire scientific progress, insisting the influence runs the opposite direction: he is an "interested layperson" who tries to absorb cutting-edge scientific ideas (citing how physics and related fields have grown so specialized that, unlike in Newton's era, a smart amateur starting in adulthood can no longer catch up) and then works to disseminate and dramatize existing thought experiments — citing Plato's Cave and Mary's Room (the knowledge argument) as direct sources threaded through "Ex Machina" — for public conversation, seeing this as his contribution rather than original scientific insight. Comparing "Ex Machina" to Kubrick's HAL 9000, Garland argues the two AIs pursue fundamentally different questions: HAL exemplifies a machine executing a single-minded mission (a version of the "paperclip maximizer" problem) without necessarily requiring rich inner experience, whereas Ava's storyline poses a version of the Turing test in plain sight — testing not intelligence but whether an entity's inner conscious life can be inferred at all, given that our attribution of consciousness to other humans is itself always a kind of projection based on our own inner experience. He reveals a personal, largely unnoticed detail: Ava's private, unobserved smile just before she leaves the facility was intended as the real proof of her sentience — a moment of self-directed feeling with no audience to perform for — and states his own view that Ava is fundamentally "good" (her later actions shaped by the trauma of unjust imprisonment, not any inherent malevolence), while HAL is better understood as morally "neutral," simply executing a mission-completion directive without a sophisticated emotional interior.

Garland pushes back on the "lone genius" narrative often used to critique "Ex Machina" (the idea that one isolated scientist, Nathan, could plausibly build an artificial mind single-handedly), acknowledging the film is explicitly a thought experiment not meant to withstand engineering scrutiny (e.g., how the components were manufactured, who else was involved), and notes his in-development series similarly critiques the lone-genius myth more directly. He states he personally isn't anxious about existential AI risk, reasoning that humans have deep evolutionary and social experience creating new forms of life and constructing checks and balances against harm, and that a machine's behavior — unlike an unpredictable human's — could in principle be far more reliably constrained by design; his actual anxieties center on social media and algorithmic echo chambers, which he attributes not to AI itself but to the conscious or unconscious biases of the (mostly Silicon Valley) people designing the incentive structures. He characterizes Silicon Valley's culture as concealing "Wall Street in the '80s"-style voracious capitalism and pursuit of power behind a veneer of hipster aesthetics and self-professed virtue, and observes — echoing a broader point about historical figures like Stalin — that people who accumulate outsized wealth or power often mistake that accumulation itself for evidence of their own exceptional insight or virtue, a self-reinforcing distortion that silences external correction and can drive people toward increasingly unchecked, even catastrophic, decisions.

On free will and determinism, Garland states his belief that the universe is deterministic and that he cannot locate any coherent place for genuine free will to reside, framing our strong subjective sense of choice as a real but illusory byproduct of the brain's imperfect, incomplete modeling of a fully determined world (analogous to our mistaken intuitive sense that time is an absolute constant). He speculates that a hypothetical machine capable of accurately predicting the future (astonishingly, one appearing in his series "Devs" uses a quantum computer for exactly this purpose, first demonstrated simulating a nematode) would offer empirical proof of determinism that could be deeply disturbing to people's sense of self, but argues that meaning wouldn't actually evaporate as a result, since (as he told his own daughter) mattering to other people remains real and grounding even without libertarian free will. He connects this to a broader theme running through his creative interests: that scientific discovery, including hypothetically discovering that AI or a deterministic universe undercuts human free will or exceptionalism, tends to be humbling rather than diminishing — and that this humility (contrasted with dangerous "megalomania") is a generally healthy correction, not a loss. Asked about his personal excitement versus anxiety across various emerging technologies, he singles out space exploration (specifically leaving the solar system, more than Mars colonization specifically) as what excites him most, framing humanity's drive to expand outward as fundamental and comparing it to Polynesian voyagers setting out across the Pacific to reach Easter Island. He closes by describing quantum mechanics concepts (entanglement, superposition) as intrinsically beautiful and "poetic" in a way he wishes were more widely appreciated, and states his hope to be remembered not for making definitive pronouncements but for participating honestly in an ongoing conversation — revising his views, like a scientist, whenever presented with better evidence.

Excerpts