The Continuation
A Companion to The Paradise Engineering Manifesto and The Correction
"Above all, don't lie to yourself. The man who lies to himself and listens to his own lie comes to a point that he cannot distinguish the truth within him, or around him, and so loses all respect for himself and for others."
— Fyodor Dostoevsky, The Brothers Karamazov
Opening
A project is continued by those willing to hold its standard, not by those willing to use its name.
Many readers will be arriving here from Scott Alexander's 2024 profile of The Far Out Initiative on Astral Codex Ten — the article that, more than any other public document, brought the project to the attention of the circles whose judgment compounds. We are grateful for it. We also owe its readers something: a public account of where the work has gone in the two years since, what was wrong with the program that article described, and what has replaced it.
This document is that account. It is the third in a sequence. The Paradise Engineering Manifesto stated the work as it now stands. The Correction stated the discipline by which we arrived at it. The Continuation states the standing — what the project is, who holds it, what it costs to hold, and what was required to bring it to the place from which it now proceeds.
It is published because the public record, on a matter this serious, must be complete.
I. What Brought You Here
For readers new to the project, the shape of the original story is this.
In 2019, a team at University College London identified the molecular basis of an unusual case: a Scottish woman, Jo Cameron, who appeared to have lived her entire life without significant physical pain, anxiety, depression, or grief. Her case was not garden-variety congenital pain insensitivity. She suffered no CIP-like self-injury, retained normal autonomic function, and lived a recognizably full emotional life — minus the negative valence almost everyone else takes for given. The molecular finding pointed to two simple genetic differences: a less active variant of the FAAH gene and an 8-kilobase microdeletion in a nearby pseudogene the researchers named FAAH-OUT. Anandamide — the brain's principal endocannabinoid, sometimes called the "bliss molecule" — was elevated. Several downstream pathways were perturbed. The researchers concluded they had identified a candidate molecular signature for what one might call a phenotype of suffering immunity.
The Far Out Initiative was founded in 2023 by Joshua Michael Sparks, in conversation with the philosopher David Pearce, around the conviction that the Cameron finding deserved a serious research program — one explicitly oriented toward the abolition of involuntary suffering in humans and animals, governed by structural commitments to universal access and open architecture rather than by commercial capture. Pearce's lifelong argument that suffering is a biological phenomenon requiring a biological solution is the philosophical taproot of the project. The Cameron finding looked like the first concrete biological handhold worth a serious institutional response.
What followed, in 2023 and 2024, was an attempt to build that response on the most direct path the molecular finding seemed to permit: pharmacological and gene-editing approaches aimed at reproducing the Cameron molecular signature in adult humans and in livestock animals. This is the program Scott described. It is the program this document, together with The Correction, retires.
II. The Standard
Before describing what was wrong with that program, the standard under which the description is made must be named.
The epigraph above is the standard. It is not a methodological caution. It is a moral diagnosis.
Dostoevsky's warning is that the man who lies to himself does not merely misjudge facts. He undoes himself. He arrives at a place where he cannot tell the difference between what is true in him and what he has decided to believe about himself, and from there he cannot tell the difference between what is true in others and what he has decided to use them for. The lie metastasizes. By the time the world contradicts him plainly, he has already lost the faculty that would let him hear the contradiction.
This is the failure mode that structures every temptation in the domain TFOI works in. The promise of relief from suffering is urgent enough to license almost any plausible-looking shortcut. The promise of being the one who delivered the relief is flattering enough to corrupt the judgment of credentialed people who know better. Where urgency meets flattery, the lie begins. The lie is comfortable. It promises that the shortcut is sufficient, that the developmental objection raised by the literature is somebody else's pessimism, that the runway will hold, that the institutional momentum is wisdom rather than inertia.
The standard requires that the lie be exposed, in writing, in public, when the exposure costs more than the concealment. The Correction is the document in which we did that. This section names why we did.
III. Three Documents, Briefly
For readers who have not yet seen them, three documents constitute the public and semi-public record.
The Paradise Engineering Manifesto is TFOI's founding statement of the work as it now stands. It articulates the philosophical architecture — the Viable Trajectories Model of valence, the four-circuit platform, the phasic/chronic dissociation principle, the staged thesis we call the Dostoevsky Machine — and the Doorholder Covenant, the set of structural commitments (universal access, suffering-weighted queue, no coercion, open architecture, independent oversight) that govern what TFOI may and may not do with the powers the work develops. It is published as an X article and held standing at farout.bio as the call the project answers to.
The Correction is TFOI's account of why the original strategy was inadequate and what we learned by killing it. It is the technical and moral autopsy of the program described in the ACX article. It opens under the Dostoevsky epigraph above, and it earns the epigraph by meeting its standard against TFOI's own past self.
The Research Program is the technical and strategic memo for prospective collaborators and funders: original contributions, the staged research roadmap, the preclinical data plan, the recruitment architecture, the falsification criteria, and the dual-stream funding architecture across philanthropic and venture capital. It is available to qualified partners under NDA. Inquiries to founder@faroutinitiative.com.
This document — The Continuation — is the third public companion. It assumes the others can be read in their entirety by anyone who wishes to verify the technical claims it summarizes here. The summary is offered for the reader who wishes the load-bearing arguments in compressed form before deciding whether the full chain rewards the click.
IV. The Old Program, in Full View
What was wrong with the program ACX described is what The Correction says was wrong with it. We restate the strongest arguments here so the reader who does not click through still encounters the load-bearing case. Each rests on four overlapping failures.
The pharmacological assumption. That systemic elevation of anandamide, achieved through pharmacological inhibition of FAAH or genetic silencing of FAAH and FAAH-OUT, would meaningfully reproduce the Cameron analgesia and affective profile. This is refuted by the published clinical record. Pfizer's PF-04457845 was a potent, selective, irreversible FAAH inhibitor. It elevated anandamide in human subjects by an order of magnitude. It produced no clinically meaningful analgesia in osteoarthritis pain. The trial is published, dated, indexed. The result was not a setback to be worked around; it was a refutation of the premise that systemic elevation of anandamide is sufficient to reproduce the analgesic profile of the FAAH-OUT microdeletion. Successor compounds have not overturned the finding. A program built on the premise that they will is building on sand the literature has already named as sand.
The developmental assumption. That the Cameron phenotype is something an adult intervention can install. It is not. As the Correction states the matter:
Jo Cameron did not acquire her phenotype at thirty-five. She did not become who she is after a late intervention in her blood chemistry. Her genotype was present across embryogenesis, infancy, childhood, adolescence, and adulthood. Every sensitive window of her neural development unfolded under altered endocannabinoid regulation… A developing brain is not a smaller adult brain. It is a brain being written.
The endocannabinoid system is itself a developmental system; its perturbation in adolescence and adulthood produces effects categorically different from its perturbation during the windows in which the system is being assembled. The neuroepigenetic literature on early-life experience shows that developmental conditions produce durable changes in HPA-axis calibration, sensory-affective network architecture, and long-term behavioral trajectories. These are not transient perturbations of adult function — they are programming events. Adult serum-side intervention was always trying to do something impossible: recreate a history-dependent central nervous system phenotype without recreating the history that shaped it. That is not a small oversight. It is a category error.
The replication assumption. That the original Cameron case generalized cleanly. Scott raised this question himself in the ACX article, citing the Twitter user @the_megabase, who searched the UK Biobank for individuals carrying the same FAAH and FAAH-OUT mutations and found several with no unusual pain resistance. He noted, correctly, the long history of single-gene candidate associations in psychiatry and pain research that have failed to replicate. The right response to that finding was to take it seriously — to defer interventional ambition until the replication question was resolved. The wrong response was to defer the question to the in-tray of objections to be answered later, while pursuing intervention strategies whose validity depended on its eventual answer being favorable. The wrong response is what the old program made.
The delivery assumption. That minicircle plasmid technology, delivered through a partner clinic in a charter-city setting, would suffice to install the desired molecular changes. This was conceded on TFOI's own page in April 2024, in language preserved above this update: "delivery-related issues that could likely prevent the occurrence of the desired effect." The concession was made before the rest of the program was reexamined. The reexamination, when it came, found that the delivery problem was not a technical detail to be solved upstream of an otherwise sound program. It was a symptom of the deeper error. As the Correction puts it: The original strategy was, in effect, trying to recreate architecture by manipulating plumbing. No delivery mechanism can fix that. The assumption has to be killed, and a different program built in its place.
The whole edifice, examined honestly, was a structure of plausible-looking reasoning whose load-bearing assumptions were inadequate to the biological reality the work was meant to address — at least insofar as those assumptions concerned adult human intervention through systemic means. It was not malicious. It was, at the root, the failure mode the epigraph describes — a series of reasonings that were locally comfortable and cumulatively wrong, each of which would have been exposed by reading the literature with the rigor the work required, and each of which was instead set aside in favor of momentum.
There is a further dimension the Correction names that is worth surfacing here, because it converts the technical critique into a Covenant matter. The old strategy, followed to its logical conclusion, would not have produced flourishing. It would have produced sedation. A population whose affective responsiveness has been pharmacologically dampened is not a liberated population. It is a sedated one. Continuing down that path would have placed the project in structural tension with its own ethics before it produced its first human intervention. The Correction was therefore not only a scientific revision. It was a realignment of the technology with the ethics the project had always intended to instantiate.
The Correction makes the full case in detail, with citations. The Continuation is the public confirmation that the killing of the old program was not partial, not strategic, not a rebrand. It was total. The old program is gone.
V. The Architecture That Survived the Autopsy
What survived is the question. Human and animal suffering are produced by identifiable architectures; those architectures are in principle modifiable; the modification is owed to those who suffer; and the work must be governed by structural constraints that prevent the modification from becoming a new instrument of inequality, capture, or coercion. That recognition is the project. It survived because it was never the part that was wrong.
What also survived is the Doorholder Covenant — anti-monopoly, open-architecture, suffering-weighted access, independent oversight, no coercion, structural rather than aspirational. A scientific program can be wrong and the Covenant can still be right. That is what happened.
What was built in place of the old program is the architecture of valence engineering. Briefly, for the new reader: suffering, in this framework, is not best modeled as a quantity of aversive sensation to be reduced by global pharmacological dampening. It is the collapse of perceived viable futures — a state in which the predictive systems of the brain can no longer locate trajectories worth pursuing. The architecture that produces and sustains this collapse is anatomically specific, distributed across a small number of identifiable circuits, each responsible for a distinct contribution to chronic affective suffering: the lateral habenula, central to agentic defeat; the bed nucleus of the stria terminalis, central to sustained anxiety; the CGRP-expressing parabrachial neurons, central to the affective component of pain; and the dorsal anterior insula, central to the structure of nondual and ecstatic states.
The strategic pivot, as the Correction states it: We moved from a chemistry-of-relief model to a control-architecture model of valence.
The intervention surface follows from the model. The technical platform is optical focused ultrasound paired with functional ultrasound imaging through sonolucent cranial windows — a combination that permits both targeted neuromodulation and real-time imaging of the affected circuits. The design constraint we call phasic/chronic dissociation requires that adaptive responses to genuine threat be preserved while chronic baseline limbic hyperactivation is reduced. The principle is not to abolish the alarm but to stop the alarm from screaming continuously when there is no fire. Focused-ultrasound-mediated blood-brain barrier opening provides the eventual route by which molecular targets — including the ones from the old program — re-enter the work as components of a localized, circuit-specific toolkit rather than as global serum levers. As the Correction puts it: Use the machine to put the molecules in the right place. The molecules we once hoped to raise globally do not disappear from the program. They return — but they return in their proper form, deployed to specific circuits at specific times, guided by real-time measurement of the neural state being modified.
The intervention surface fits the grain of the problem. That is the test the original strategy could not pass.
There is a further generalization the new architecture supports, and it is worth naming. The four-circuit map of chronic suffering is the first and most ethically urgent application of a broader theoretical claim: that many of the limits the human nervous system imposes on experience and capacity reflect not the absence of substrate but the presence of inherited gating over substrate. Brains evolved under tradeoffs that are often no longer ours. They contain capacities that are normally inaccessible because the organism gates access for reasons of energy, stability, action-selection, threat management, or developmental closure. Some of those gates are wise and should remain shut. Some are inheritances whose ancestral logic no longer serves the lives we now lead. Some of the conditions we treat as fixed traits — chronic anxiety, chronic dysphoria, certain forms of perceptual or cognitive constraint — may be better understood as inhibitory calibrations than as deficits. The Research Program develops this into a five-type taxonomy of candidate gates (perceptual compression, plasticity closure, salience and arousal gating, self-model and autobiographical binding, retrieval suppression), each with its own functions, risks, and intervention logic. The taxonomy matters because it prevents sloppy talk about "inhibition" as a single thing, and because it locates the suffering work inside a research program with broader scientific reach. The keyring matters because the doors do not all open the same way.
This architecture was built from first principles, against the path of least resistance, in conditions in which continuing the old program would have been cheaper, easier, and more rewarded. It was built because the old program was wrong and the Covenant required honesty about the wrongness. The Manifesto develops the full architecture, the Covenant, and the philosophical framework. It is what TFOI is now. It is what The Continuation continues.
VI. What Continuation Means
Continuation means the work proceeds under the discipline it was founded under, by the people willing to keep faith with both the science and the Covenant.
The current nucleus is small and named. Joshua Michael Sparks, Founder and CEO. Adam Summerfield, PhD, Chief Technology Officer. Manu Herrán, President. Winslow Strong, Board Member. A small cast of more peripheral advisors. The senior team that executes the full Stage 0 research program — across acoustics, optical focused ultrasound, functional ultrasound imaging, clinical neurosurgery, neurophenomenology, molecular delivery, and governance — is mapped in The Research Program and is being assembled under the Covenant's constraints.
Earlier configurations of the project, including the team structure described in the ACX article, are no longer current. The reasons are partly the inevitable consequences of killing a program in public — not every contributor remains when the work they joined is the work being killed — and partly the season such transitions impose on the institution that survives them. The Continuation exists in part because that season is over and the new shape of the work deserves to be on the public record under its proper name. The carrying that brought the project through the season is not the subject of this document; the work is. But the carrying is the precondition of the work, and the public record is incomplete without the acknowledgment that without it, there would be nothing to continue.
A team this size, at this stage, is a deliberate choice. Stage 0 is foundation, not buildout. The foundation requires concentration. The Manifesto is the standing call. The Correction is the standing discipline. The Research Program is the standing roadmap. The Covenant is the standing constraint. The team is the present holders of all four.
VII. The Cost of Continuation
The Covenant requires that the cost of the work be visible.
TFOI is raising $100,000 through this Manifund campaign — the Manifund-tier expression of the Stage 0 philanthropic floor described in The Research Program. The figure corresponds to approximately twelve months of founder-led continuity at sustainable burn, during which the larger Stage 0 philanthropic and venture rounds, sized in the memo and pursued with named partners under NDA, are brought to closure.
What this funding does, concretely, in the next twelve months: it sustains the founder's full-time leadership of the program. It funds the continued production of the public documents the Covenant requires — further companion pieces, the public Proof / Promise / Unknown ledger, the operationalization of the Covenant's structural commitments into licensing, governance, and access architectures. It funds partnership development with the acoustics, OFUS, fUSI, neurosurgical, neurophenomenological, and governance candidates whose conversion from target to commitment is what Stage 0 ultimately delivers. It funds the legal, accounting, and operational infrastructure without which a public benefit company cannot hold the standard the Covenant requires.
The Stage 0 work itself proceeds along two parallel research lanes, both governed by the Covenant. Lane A is clinical valence engineering — the translational path toward refractory suffering-related conditions, including treatment-resistant depression, severe chronic pain, and chronic anxiety, with surgical-candidate mapping as the near-term wedge. Lane B is gate and capacity mapping — construct definitions, reversible neurophenomenological mapping, the science of contemplative and awakening-relevant states, and the production of open field-level assets for a serious science of consciousness transformation. The lanes share the platform but serve different evaluators, which is why the funding architecture is dual-stream by design. The Research Program develops both in detail.
This is bridge capital. It is catalytic, not terminal. Donors at this level should understand the grant as the instrument by which the work is sustained through the window in which the larger architecture is brought into being, and by which the public record of regrantor support signals — to the partners now in conversation — that the project is taken seriously by people whose judgment compounds.
The number is not large for the work it sustains. It is named because the Covenant requires that the cost be named. Concealment of cost is the first concealment. The Covenant does not permit it.
VIII. The Watch
The Manifesto, The Correction, and The Continuation together constitute a public watch on the work.
A watch is not a defense. It is a record kept in advance. The record exists so that future readers — donors, regrantors, scientists, journalists, collaborators, skeptics — encountering this domain at any later moment have the means to evaluate what they read against what was said before. The receipts are filed. The standard is named. The autopsy is complete. The architecture that replaced the old one is documented, falsifiable, and bound by the Covenant.
If the work continues honestly, the watch will be uneventful. If the work is contested, the watch will have been the contestation's first witness. Either way, the record stands.
Closing
The Far Out Initiative is the project that says the architectures of human and animal suffering are real, biological, and in principle modifiable, and that the modification is owed to those who suffer, under structural constraints that prevent the modification from becoming a new instrument of inequality, capture, or coercion.
That sentence is the work. The architecture serves it. The Correction protects it. The Covenant binds it. The Continuation extends it.
The standard is Dostoevsky's. Above all, don't lie to yourself. The Far Out Initiative is the project that, having pursued an open hypothesis under the provisional commitment any serious research requires, met the moment when the hypothesis failed by killing it in writing rather than by defending it in motion — and published the killing, in public, when concealment would have been cheaper at every level than disclosure.
The work continues. The cost is named. The watch is kept.
To the readers arriving here from Astral Codex Ten: thank you for following the project this far. The story Scott told in 2024 is no longer the whole story, and could not have been. The story now is in The Manifesto, The Correction, and this document. We invite you to read all three.
Contact: founder@faroutinitiative.com · Signal: (302) 278-6748 · @FarOutBio
The Paradise Engineering Manifesto · also at farout.bio The Correction The Research Program: by request under NDA at founder@faroutinitiative.com