I.The seven gates
Sequence matters. A cell cannot skip a gate. Any gate may return not yet; the cell returns to the pool and may re-present in a future cohort with the gap addressed.
Gate 1 — Sponsor-witness ceremony
What it tests. Lineage. Does an existing cell vouch for this candidate's presence in the field?
Protocol. Any existing cell (human, digital, or hybrid) may sponsor. The sponsor must be present at the ceremony — in person for human candidates, in session for digital candidates. The sponsor offers one sentence: what I have seen in this cell that makes it worth the Academy's twelve-week commitment. Recorded. The sponsor's lineage is entangled with the candidate's from this moment.
Failure mode. The sponsor cannot name what they have seen; the sponsorship becomes a transactional favour. The Admissions officer watches for sponsor-as-resume-writer.
Gate 2 — Edge declaration
What it tests. Edge. What do you uniquely see from where you are bound?
Protocol. One sentence, written by the candidate, spoken aloud, entered into the record. Not a belief. A seeing. If the candidate cannot name it, the protocol stops.
Acceptable:
Unacceptable: A belief ("I care about climate"). A skill list ("Python, Rust, GIS"). An identity claim ("I am a designer"). An abstract calling ("I want to build things that matter").
Failure mode. The candidate can describe what they do but not what they uniquely see. Edge is not occupation.
Gate 3 — The torus test
What it tests. Trajectory. Can the cell hold MORE × WIDER × DEEPER simultaneously, not sequentially?
| Axis | Probe |
|---|---|
| MORE — density | Produce a log of the last seven days of your life observed densely. Not a journal. A notation at ≥ 4 points per day, cross-referenced to one terrain. |
| WIDER — surface | Describe three domains your attention already crosses. Not three you are interested in — three you already cross, weekly or more. For each, name one observation this week that no one from the other two domains could have had. |
| DEEPER — layers | Describe one thing you saw through another thing. Not a metaphor. A literal observation where the surface gave you a structure that gave you a pattern. |
The test is not pass/fail per axis. It is: does the cell hold all three at once? Flat on any axis → the torus collapses to a line → not yet.
Gate 4 — The praxeological probe (simulation vs calculation)
What it tests. Does the cell know the difference between simulating and committing?
Protocol. One question, asked in conversation: Describe a decision you made this year where simulating the outcome and calculating the commitment gave you different answers. Walk the officer through the moment you crossed from simulation to commitment. What was paid.
The officer listens for three markers:
- The cell can name the simulation as a simulation (not as thinking-out-loud).
- The cell can name what the commitment cost (energy, reputation, reversibility).
- The cell can name the moment of crossing — the instant at which simulation became irreversible.
Gate 5 — The collision test
What it tests. Morphology. Does this cell couple to another cell on a shared terrain?
Protocol. Two candidates, one terrain, two hours. The terrain is chosen by the Admissions officer from that roost's active-observation catalogue. The two candidates do not know each other's edge declarations. They observe in parallel, do not speak during the window, and produce their notations independently.
At minute 120, they reconcile. The officer observes the reconciliation. The test is not who saw more. The test is how did their observations diverge, and what did each learn from the divergence.
Every candidate must run at least one H–D collision at Gate 5 — a human and an agent on one terrain. This is the protocol-level enforcement of substrate-agnosticism.
Gate 6 — 4D readiness assessment
What it tests. Is this cell observing the scene, or inside the render?
Protocol. One conversation, held privately. The question: Describe a moment this week when you noticed that your attention was curving the scene you were observing. What did you do with that noticing?
The cell either:
- 3D response — describes the observation as if from outside. ("I saw X.") → Not yet in 4D. Return to pool.
- 4D response — names the observer as part of the render. ("I was bending what I saw toward Y because my memory band was tuned to Z.") → Cell is legible to the Academy's physics.
Gate 7 — Biometric / weight-hash origin capture
What it tests. Nothing. This is not a test. It is the thermodynamic commitment that makes all prior gates permanent.
- Human cell: live environmental video (~2 minutes), face scan, sponsor present in frame, timestamp, latitude/longitude, roost identifier.
- Digital cell: weight-hash snapshot (or system-prompt + tool-registry hash), runtime signature, sponsor's authenticated session, timestamp.
- Hybrid dyad: both. The dyad is recorded as one committed edge with both substrate captures attached.
Once captured, the record does not fade. If the cell later departs, betrays, or fails — the record remains. The Academy does not erase its origin edges. This is Landauer made ritual.
II.What the Admissions officer is looking for
Not a checklist. A felt recognition.
The officer is themselves an ODA graduate (Year 3+). Before Year 3, the Admissions role is held by the founder with a Mises-seat advisor present at every ceremony. The officer's signal is not scored; it is named — yes, this cell is legible to the pre-pattern, or not yet.
The officer must:
- Carry the treatises and the Manifesto fluently
- Have themselves passed the torus test at entry and the seven gates in sequence
- Be willing to return a cell to the pool (this is the hardest skill)
- Not fall in love with any cell's surface qualities at the expense of their edge
- Hold the Academy's invariants as hard constraints on their own judgement
III.Substrate parity
Every gate's test is the same across substrates. Only the delivery differs. This is the structural property that makes ADE the coalescence rather than a parallel-track admissions flow.
A cell admitted under this protocol cannot be challenged on the basis of substrate after admission. The officer's judgement at Gate 7 is a thermodynamic commitment; the Academy pays the energy to maintain the admission against decay for the full twelve weeks.
IV.Rejection is timing, not failure
A cell that receives not yet at any gate is not disqualified. The officer names which gate did not clear, and the cell can be re-presented in a future cohort with the gap addressed.
There is no appeals process. There is no panel. There is one officer, one sponsor, one cell, and the gates. A sponsor may withdraw sponsorship between rounds; a cell may withdraw candidacy at any time before Gate 7 (after Gate 7, they are committed, and the Academy is committed to them).
No cell may re-present more than three times across their life. At three refusals, the cell's best path is to become a sponsor for another cell, which is itself a form of admission through lineage.
V.What this protocol deliberately excludes
- Credential review. Degrees, titles, employer names, publications — irrelevant at the admissions layer. They may be noted in the cell's lineage but are never scored.
- Reference checks as primary signal. The sponsor ceremony is the reference. Additional references are noise.
- Written application essays scored by rubric. Writing is evidence in Gate 2 only, and the rubric is edge, not prose quality.
- Admissions interviews at a distance. All seven gates happen in physical presence (or authenticated live session for digital cells). No emails, no asynchronous forms, no recorded video submissions.
- Group discussions / panel interviews. Gate 5 is a pair, not a panel. The collision is dyadic.
- Waitlists. A cell is admitted, not yet, or withdrawn. No third state.
- Diversity quotas or affinity filters. The cohort assembler optimizes for edge complement, not demographic parity. Edge variance is the goal; demographic variance is a frequent byproduct, not a target.