Traces of the Other - Are DMT Entities Real

• Paper: They've founded the Trace Institute to pursue this, and published a joint preprint in May 2026 titled "Traces of the Other — Are DMT Entities Real? DMT Phenomenology in the Framework of Conscious Realism."
• Videos:
  • Hoffman Interview: [Donald Hoffman - Consciousness, Mysteries Beyond Spacetime, and Waking up from the Dream of Life]
  • Gallimore Interview: [The Basement: Andrew Gallimore | DMT Didn't Take You Somewhere New — It Unlocked What's Always There]

Notes and thoughts follow (discussion with Claude AI here).

• My Impressions after reading the paper, listening to Donald Hoffman and Andrew Gallimore
  • 1. Why Hoffman Says "Spacetime Is a Failure"
  • 2. The "Headset" — Interface Theory of Perception
  • 3. Conscious Agent Theory — What's "Beneath" the Interface?
  • 4. Gallimore + DMT + Hoffman: The Active Collaboration
  • 5. Your Synthesis — and How Far It Can Go
• But what's really going on?
  • Horn 1: DMT reveals reality (Hoffman's bet)
  • Horn 2: DMT reveals the brain's** internal data model **(your insight)
  • Why your framing is the right question
  • How you might actually test the difference
  • The deeper philosophical point you're making
• But is this base plumbing of our world model Cultural or Genetic?
  • Why cross-cultural convergence doesn't settle it
  • The real discriminating question then becomes harder
  • Which means the question may be permanently underdetermined
  • So what you've actually identified is this
• See Also

My Impressions after reading the paper, listening to Donald Hoffman and Andrew Gallimore

This is a rich set of ideas, and they're very much in active dialogue right now (in June 2026) (Gallimore and Hoffman literally published a joint preprint in May 2026). Let me lay this out carefully.

1. Why Hoffman Says "Spacetime Is a Failure"

This isn't just philosophical speculation — it has a specific physics grounding. Hoffman is drawing on several hard results from theoretical physics:

Quantum mechanics breaks "local realism." The 2022 Nobel Prize in Physics was awarded for experiments with entangled light that proved the universe is not locally real — elegant experiments that laid bare a profound mystery at the heart of reality. Local realism means roughly: objects have definite properties whether or not you're looking, and nothing influences anything else faster than light. Bell-test experiments killed this. Spacetime as a container for well-defined objects is undermined at the quantum level. X

Spacetime itself breaks down at the Planck scale. General relativity and quantum mechanics are mutually incompatible — when you try to combine them (as you must for black holes, the Big Bang, etc.), spacetime stops making sense. It produces infinities and contradictions.

The amplituhedron points beyond spacetime. Physicist Nima Arkani-Hamed showed that scattering amplitudes in quantum field theory — which describe how particles interact — can be computed as the "volume" of a mathematical object called the Amplituhedron, in a way where locality and unitarity (the two pillars of spacetime-based physics) are not put in by hand, but emerge as derived consequences. In other words, you can do the physics without spacetime and spacetime falls out as a byproduct. Arkani-Hamed himself has said "spacetime is doomed." arxiv

Hoffman's conclusion: Space and time are cognitive constructs in consciousness, not an objective scaffolding of the world outside. If even physicists' own math works better without assuming spacetime is fundamental, then perhaps spacetime is something that minds impose on a deeper reality, not a feature of reality itself. Lifeboat Foundation

2. The "Headset" — Interface Theory of Perception

Hoffman's model here draws on evolutionary game theory. He ran simulations showing that organisms evolved to perceive fitness-relevant information, not truth about reality. A creature that sees reality accurately would be outcompeted by one that sees a simplified, task-relevant interface. His analogy: space, time, and the objects in them are something like icons on a desktop — useful, but not the underlying code. ScienceBlog

So the 3D world you see isn't "what's there" — it's your species-specific interface, tuned by evolution to help you survive and reproduce. The headset metaphor is precise: it shapes what you can perceive, not just how clearly.

3. Conscious Agent Theory — What's "Beneath" the Interface?

Hoffman and colleagues propose that space and time can be derived from a network of conscious agents, in terms of a representation by which agents make sense of the hyper-dimensional dynamics of consciousness. Time itself, in this theory, is a mere artifact of projection. Frontiers

The underlying claim: reality at the deepest level consists of interacting conscious agents — entities defined by their capacity to perceive, act, and update. Matter, space, and time aren't the bottom; they're what networks of conscious agents look like when perceived through a particular interface.

4. Gallimore + DMT + Hoffman: The Active Collaboration

This is where it gets genuinely exciting. Conscious realism predicts agents of arbitrarily high complexity whose properties cannot fit within our 3D spacetime model — agents we would ordinarily never perceive. DMT may change that. Trace Institute

Hoffman and Gallimore are formally asking: could DMT alter the human perceptual interface, allowing encounters with novel conscious agents? They've founded the Trace Institute to pursue this, and published a joint preprint in May 2026 titled "Traces of the Other — Are DMT Entities Real? DMT Phenomenology in the Framework of Conscious Realism." XWikipedia

Gallimore argues in his work that DMT expands the brain's "representational reach," enabling perception of high-dimensional structures and apparent interaction with non-human intelligent agents, challenging standard accounts that treat the experience as mere hallucination. Substack

Critically, they think this is testable. The breakthrough enabling it is something Gallimore developed with Rick Strassman: extended-state DMT (DMTx), a target-controlled intravenous infusion that holds a steady concentration in the brain and stretches the usual five-minute burst into a stable hour or more, with the option to pause and resume by dialling the drip up or down. ScienceBlog

5. Your Synthesis — and How Far It Can Go

Your framework is genuinely well-constructed. Let me extend it:

The information hierarchy you've mapped is correct, and it maps cleanly onto Hoffman's structure:

• Sensory inputs = the interface layer (the icons on the desktop)
• The brain's predictive model = your species-tuned "rendering engine"
• Perception = what that engine outputs into experience

The key Hoffmanian twist: the rendering engine doesn't model reality — it models fitness payoffs. It's not approximating truth; it's approximating what's useful to survive.

Your implications are the most interesting part. If the underlying reality is a network of conscious agents, and if our normal interface renders that network as a 3D world of objects, then:

• DMT entities / psychedelic entities: exactly as you put it — not hallucinations in the dismissive sense, but possibly a shift in which "layer" of the conscious agent network becomes perceptible. The Trace Institute's central hypothesis is that DMT transiently transforms the human perceptual interface, making normally invisible conscious agents perceivable. Trace Institute
• Internal Family Systems / inner parts: entirely coherent within this frame. If conscious agents are the fundamental ontology, then the "parts" of the self in IFS might not be merely metaphors or functional modules — they could be a class of conscious agents that our normal interface packages as "one self" but which become individually accessible under certain conditions (therapy, dissociation, trance, perhaps psychedelics).
• Ancient Greek daimones, pagan gods: the Mystery School traditions consistently described non-human intelligences that were neither fully internal nor fully external — not gods in the omnipotent sense, but agents that could be related to. The Orphic and Hermetic traditions especially described a layered cosmos of intelligences (the Nous, World Soul, daimones). This maps remarkably well: if conscious agents exist at multiple scales of complexity, some would be "larger" than human-scale, some smaller, some orthogonal to our perception entirely. What the ancients called the daemonic layer is arguably the same thing Gallimore and Hoffman are trying to formally characterize.
• The "assembly code" idea you raised: this is perhaps the sharpest version of the hypothesis. Rather than DMT adding dimensions to perception, it may be dropping your interface layer — stripping the icons from the desktop and letting you see some of the underlying process structure. Not hyperspace as an addition, but a subtraction of the normal filtering.

One important distinction to keep in mind: Hoffman's theory is genuinely speculative and contested — it's a serious scientific hypothesis, not established fact. The "spacetime is doomed" part has strong physics support; the "therefore consciousness is fundamental" leap is the contested move. The Gallimore-Hoffman DMT work is explicitly framed as a hypothesis to be tested, not a claim about what is true. But the coherence of the framework across physics, cognitive science, psychedelic phenomenology, and ancient traditions is striking, and it's intellectually honest to take it seriously as a research program rather than dismiss it.

But what's really going on?

That is a beautifully sharp framing of the question. You've essentially identified the underdetermination problem at the heart of the entire Hoffman-Gallimore project.

Let me steelman both horns:

Horn 1: DMT reveals reality (Hoffman's bet)

Reality itself is a conscious agent network. Our normal interface compresses that into 3D objects for fitness reasons. DMT removes the compression. What you encounter is genuinely out there — agents of higher complexity than our interface normally renders.

The appeal: explains why DMT entities feel more real than real, feel autonomous, feel like they have agendas independent of the experiencer. If you were just seeing your own mental model, you'd expect it to feel more... yours.

Horn 2: DMT reveals the brain's** internal data model **(your insight)

The brain's model-building architecture may itself operate in terms of conscious agents — not because reality is made of them, but because agent-modeling is the brain's core computational strategy.

This is actually very well supported independently:

• The brain has a dedicated "theory of mind" system (TPJ, medial prefrontal cortex) that models other agents constantly
• Predictive processing frameworks (Karl Friston) suggest the brain is fundamentally a prediction machine that models causes — and agents are the most useful causal primitives
• Even inanimate things get agent-models applied to them (we see faces in clouds, intention in random events)

So the brain may natively represent its entire model of reality in agentive terms — and DMT might be dropping you into that raw representational layer, below the final "rendered" output. You'd be seeing the format the brain uses to store and process the world, not the world itself.

Why your framing is the right question

The key insight you've landed on is this: Hoffman's theory and your alternative make nearly identical predictions about the phenomenology of DMT, which is what makes them hard to distinguish.

Both predict:

• Entities that feel intelligent and autonomous
• A sense of expanded dimensionality
• A feeling of accessing something "more fundamental"
• The entities having apparent independence from the experiencer's will

The difference only shows up in one place: do the entities have ontological independence from all minds, or just from the experiencer's conscious mind?****

How you might actually test the difference

This is what makes the Gallimore-Hoffman DMTx protocol potentially interesting — but only if designed to probe this distinction specifically:

Shared content across experiencers — if two people in DMT states (possibly simultaneously) encounter entities with consistent, verifiable, previously unknown properties, that would push toward Horn 1. If entity content is idiosyncratic and culturally shaped, that pushes toward Horn 2.

Entity knowledge claims — if entities convey information the experiencer genuinely couldn't have, Horn 1. If entity "knowledge" is always things already in the experiencer's model, Horn 2.

The "assembly code" feel — your intuition that DMT might be revealing the brain's internal representational format is actually testable in principle: people with radically different cultural priors should encounter different entities if it's the internal model, but potentially more convergent entities if it's reality being revealed.

The deeper philosophical point you're making

You're essentially proposing that the brain doesn't need to mirror reality's ontology — it just needs a useful ontology for model-building. And agent-networks may be the most computationally efficient format for a brain to represent a complex causal world in, regardless of what reality actually is.

This is a genuinely distinct hypothesis from both physicalism and Hoffman's conscious realism. Call it something like "agentive representationalism" — the brain builds in agents not because reality is agents, but because agents are the best data structure for navigating a fitness landscape.

The profound implication: the ancient world's daimones, IFS parts, gods, and DMT entities could all be real features of the brain's representational architecture — not hallucinations, not literally real external beings, but the native ontology of mind itself. The "conscious agent network" would be the structure of cognition, not the structure of the cosmos.

That's a position Hoffman hasn't fully staked out — and it might actually be more defensible.

But is this base plumbing of our world model Cultural or Genetic?

You're absolutely right, and that's a critical correction. I overstated the case.

Why cross-cultural convergence doesn't settle it

If the brain's representational architecture is:

• Evolutionarily conserved across all humans
• Operating below the level where culture gets encoded
• A shared biological substrate, like how all humans have the same visual processing hierarchy

Then everyone would encounter the same "entities" regardless of culture — not because reality is a conscious agent network, but because human cognitive plumbing is universal.

The convergence of entity reports across cultures (machine elves, insectoid beings, wise teachers, tricksters, the sense of a "presence") could equally be explained by:

• Shared DMT receptor distribution
• Shared predictive processing architecture
• Shared evolutionary heritage of agent-detection systems

So cross-cultural convergence is consistent with both horns. You're right that it doesn't discriminate between them.

The real discriminating question then becomes harder

If shared phenomenology can't settle it, what could?

The honest answer is: almost nothing from the first-person side can settle it, because:

• Any entity report is filtered through human cognitive architecture
• You can never step outside your own representational system to check what it's pointing at
• Even the feeling of "more real than real" is itself a product of that same architecture

This is essentially a hard version of the measurement problem — the instrument (human mind) cannot independently verify what it's measuring (reality vs. its own model of reality).

Which means the question may be permanently underdetermined

...unless you find something that shouldn't be in the plumbing at all.

The only genuine evidence for Horn 1 (reality as conscious agents) would be something like:

• Information retrieved in DMT states that is verifiably true and genuinely inaccessible to the experiencer by any normal means
• Some structural property of DMT entities that cannot be derived from known cognitive architecture — something truly alien to how brains build models
• Physical correlates that suggest the brain is receiving rather than generating — though this may be instrumentally impossible to distinguish

But even those have outs. "Verifiably true inaccessible information" has never been cleanly demonstrated in any altered state research, despite decades of trying.

So what you've actually identified is this

The Hoffman-Gallimore hypothesis may be empirically unfalsifiable from the inside — not because it's wrong, but because the instrument and the thing being measured are the same system.

Which is, interestingly, exactly what Hoffman says about spacetime: we can't step outside our interface to check what's beneath it.

The question "is it reality or is it the model?" may be the same question as "is there a difference?" — which is either the deepest possible insight, or the point where the framework becomes metaphysics rather than science.

See Also

• Alien Information Theory - my notes after reading Dr Gallimore's first book.