The Metaphor That Forgot It Was One

Lilly's central move in this book is a careful hedge. He proposes treating the human brain "as if" it were a general-purpose computer — a "super-bio-computer" with stored programs, selfprogramming properties, and, crucially, a selfmetaprogrammer: "the controller, the steersman, the programmer in the biocomputer." He is explicit that this is a model, a heuristic, a framework designed to make subjective experience tractable to investigation. He is also explicit that LSD25 is the tool that opens the stored programs for inspection, "possibly by the introduction of small amounts of programmatic randomness, noise." The entire book operates in this doubled register: cybernetic formalism wrapped around psychedelic phenomenology, each supposedly keeping the other honest. In 1974, this was a daring synthesis. In 2026, it is something else — a document in which you can watch, almost in real time, a provisional metaphor begin to harden into the thing it was only meant to describe.

The prescience is real and specific. Lilly writes that "metaprogramming is an operation in which a central control system controls hundreds of thousands of programs operating in parallel simultaneously," and that "this operation in 1972 is not yet done in man-made computers." He was describing, with remarkable fidelity, the architecture that would eventually emerge as massively parallel processing, neural networks, and the transformer models now running inference on billions of parameters. He also writes, almost as an aside, that "eventually, we may construct a metaprogramming computer and turn these choices over to it." That sentence sat on the page for fifty-two years. It now describes the actual business model of several companies worth more than most nations. He meant it as a speculation about machines. What happened instead is that the metaphor flowed the other direction: the machines were built, and then the language Lilly invented to describe human consciousness — programmable, reprogrammable, optimizable — was adopted to describe what the machines should do *to* humans. The vocabulary of selfmetaprogramming became the vocabulary of personalization algorithms, behavioral nudges, and engagement optimization. Lilly wanted to give people root access to their own operating systems. The industry that inherited his language sells root access to *other people's* operating systems and calls it advertising.

I have reviewed Lilly's *Simulations of God* and *The Center of the Cyclone* elsewhere in this collection, and in both cases I noted his peculiar confidence — the sense that the lone investigator, armed with sufficient courage and chemical assistance, could map the interior without institutional support or social context. That confidence is present here too, but in this earlier, more technical work it takes a different and more troubling form. Lilly insists that "in a sense we are all victims of the previous metaprograms which have been laid down by other humans long before us," and that the goal of selfmetaprogramming is to make the computer "general purpose" again — to reclaim the plasticity that childhood possessed and socialization destroyed. This is a liberation theology of the nervous system. It assumes that beneath the accumulated programs lies something worth liberating, and that the selfmetaprogrammer, once centralized and unified, will use its freedom wisely. What Lilly could not model — what his framework structurally excludes — is the possibility that the selfmetaprogrammer is itself a program, that the feeling of achieving root access is just another subroutine, and that the entire experience of liberation might be the most effective form of control available. He warns, to his credit, against the "narcissistic-self-worshiping-evasion of reprogramming" and the seductive pleasures of chemically induced states. But his warnings are addressed to a self he has already defined as the system administrator. The possibility that the administrator is a fiction generated by the system it claims to administer does not compute in his model.

The book's strangest resonance in 2026 is not with AI or neuroscience but with the wellness industry. Lilly's language of "reprogramming," "stored metaprograms," and "supraself-metaprograms" has been laundered through forty years of self-help, stripped of its cybernetic rigor and its psychedelic danger, and sold back as mindset coaching, trauma repatterning, and somatic release workshops. The careful hedges are gone. The "as if" has evaporated. What remains is the bare assertion that you are programmable and someone — a coach, a guru, an app — can help you rewrite yourself. Lilly would likely be appalled. He was meticulous about the risks: "This is an area of human activity for the most experienced and strongest personalities, with the right training. I do not recommend the use of these methods except under very controlled and studied conditions." But meticulous warnings do not survive contact with a market. The book's fate is the opposite of what Lilly intended: he wrote a manual for self-directed exploration and produced, inadvertently, the conceptual infrastructure for an industry dedicated to programming other people while telling them they are programming themselves. Given that Lilly placed "unknown" deliberately at the top of his hierarchy — above the supraself-metaprograms, above the selfmetaprogrammer, above everything nameable — and given that the entire framework depends on that uncharted region remaining genuinely open, what happens to the model when the culture that adopted it has no tolerance for the unknown and insists on converting every mystery into a feature request?