Do people see the same things on DMT?

The claim, stated carefully

Ask people who have inhaled N,N-DMT what they saw and a striking number of them reach for the same vocabulary. Lattices and grids. Fractal corridors. Tessellated fabric that seems to be under construction. And, far more often than a purely random visual system would predict, other beings: figures that seem aware, seem to be waiting, and seem to be showing the person something.

That much is not folklore. It is in the published record. A survey published in the Journal of Psychopharmacology in 2020 by Davis and colleagues collected accounts of entity encounters occasioned by inhaled DMT and found that these reports are common, structured, and personally significant to the people who have them. Rick Strassman's clinical work at the University of New Mexico, later described in DMT: The Spirit Molecule, ran into the same thing decades earlier in a controlled hospital setting, with volunteers reporting contact they did not expect and could not explain.

So the honest starting sentence is not "people see the same things." It is narrower and more interesting: people describe the same things, at a rate that is worth measuring properly.

The gap between those two sentences is the entire reason this project exists.

Why "the same thing" is harder to establish than it sounds

Every one of the following can manufacture apparent agreement without anything unusual happening. None of them are exotic. All of them are ordinary features of how humans perceive and talk.

Priming and expectancy. Almost nobody arrives at DMT unprimed anymore. The imagery is on YouTube, in art, in podcasts, in the word "machine elves" itself. If you have heard the description before you take the compound, your report is no longer independent evidence for it.

Shared hardware. The visual cortex has a geometry. Bressloff and colleagues showed that the recurring form constants of hallucination, tunnels, spirals, lattices, cobwebs, fall out of the mathematics of cortical architecture. Two strangers can produce the same grid because they are running the same tissue, not because they visited the same place.

Language collapse. A person has roughly a dozen usable words for an indescribable visual event. Everyone reaches for "fractal", "geometric", "entity". Vocabulary convergence is not perceptual convergence.

Pareidolia and pattern completion. The face-detection system does not switch off in altered states. It gets louder.

Retrospective editing. Memory of a two minute experience is reconstructed, and it is reconstructed using the language the person encountered afterward.

Any serious attempt to answer this question has to survive all five. Most published discussion of DMT convergence does not even try.

What the evidence actually supports today

Here is the current state, held to what can be defended:

• Strong: entity encounter reports are common, cross-cultural, and phenomenologically consistent enough that researchers treat them as a category rather than as noise.
• Strong: the DMT state has a measurable neural signature. Timmermann and colleagues, in Scientific Reports in 2019, characterised the experience with multivariate EEG and found changes distinct from ordinary waking activity. This tells us the state is real and physiologically specific. It does not tell us what is being seen.
• Weak to absent: any controlled demonstration that independent, unprimed observers converge on specific visual content beyond what shared neuroanatomy and shared culture already predict.

That last line is the one that matters, and it is the one usually skipped. The convergence claim is currently unfalsified, not confirmed. Those are different things.

The strongest case against

We link the skeptical arguments deliberately, and we link the good ones rather than the easy ones.

Andrew Gallimore, who has spent more time modelling the DMT state than almost anyone, has publicly argued against the strongest version of the laser code claim. Zeus Tipado has made a similar argument in public: that lasers and DMT are not going to reveal the mysteries of the universe, and that treating them as if they will damages the credibility of the field. There is also a 2026 preprint titled "Detecting Code of Reality with No Lasers and No Dimethyltryptamine", which reports the effect turning up in the absence of the two ingredients supposedly responsible for it.

If you only read the enthusiastic material, you have not read the topic. A project that claims to be an honest instrument has an obligation to hand you the counter-argument at the same moment it hands you the claim.

How this gets settled

Not by more testimony. Testimony is already abundant and it has already stopped being informative, because it is contaminated by exposure.

It gets settled by blinding. The design question is simple to state and hard to execute: can a person who has not been shown the reference set produce a description that matches it above chance, scored by someone who does not know which condition they are scoring?

That is what the symbol registry is being built to support. It is early and it is small, and the counts on the registry pages are the real counts, not aspirational ones. A convergence result that arrives fast is a convergence result that was contaminated. We would rather publish a slow honest null than a fast interesting artefact.

What would change our mind, in either direction:

• Toward convergence: unprimed participants, blinded scoring, and a match rate that holds after controlling for prior exposure and for the form constants that cortical geometry already predicts.
• Away from convergence: the same protocol returning chance-level matching, or the effect surviving removal of the compound entirely, which would relocate it from pharmacology to expectation.

Both outcomes are publishable here. That is the point of the instrument.

Before you go looking

This page is not an instruction. If you are reading it because you are considering the experience yourself, two things belong in the same breath as the phenomenology.

First, this material can be destabilising in ways that outlast the experience. Argyri and Evans have written specifically about ontologically challenging psychedelic experiences and how the resulting distress is managed. That literature exists because the distress is real and reasonably common, not because it is rare.

Second, N,N-DMT is a controlled substance in most jurisdictions, and nothing here is a recommendation to obtain or use it. Where we document protocols, we document what is being investigated and under what conditions, not how to reproduce it at home.