Before the replicator / direct swing 0423 Aug 2026
Gate-conditioned future repertoire

The gate opens futures.

The hidden developmental state we learned to control determines how many different organisms the same later causes can still produce.

39organisms
64future interventions each
9,984new trajectories
4gate states compared
01 / The direct test

Put the system through the same 64 futures. Count no categories. Set no threshold.

From each passage-4 state—open, closed, untouched, and blind-full—we exhaustively applied every ordered pair of eight later actions. Thirty-two steps later, we measured the continuous geometry of all 64 resulting fields.

The question was brutally simple: did opening the gate leave more genuinely different futures reachable?

02 / More directions remain alive

The open state had about five more effective future dimensions than the closed state.

Effective rank of the 64-future map

open44.1
untouched41.9
closed39.5
blind full37.5
+4.62

open minus closed

95% interval [1.24, 7.94]. The open map averaged 44.13 effective dimensions; the closed map averaged 39.51.

+4.62effective dimensions over closed
+6.62effective dimensions over blind full
11.7%relative expansion over closed
03 / The futures spread apart

The result is visible in raw outcome distance, not only a spectral summary.

Mean pairwise Hellinger distance

open0.155
untouched0.151
closed0.147
blind full0.145

Open futures were farther apart from one another than closed futures: +0.00786, interval [0.00234, 0.01366].

That direction was positive in 27 of 39 organisms and 14 of 16 families. Blind repeated forcing narrowed the repertoire even more.

04 / Later causes regain leverage

Opening the gate made the system more responsive—especially to what happened second.

+0.0071

first-action sensitivity

interval [0.0012, 0.0134]
+0.0134

second-action sensitivity

interval [0.0090, 0.0177]

The opened state did not merely contain more variation. It preserved the ability of later causes to matter.

Even compared with untouched development, the open state had stronger second-action sensitivity: +0.0043, interval [0.0003, 0.0083].

05 / The map itself was rewritten

The same action no longer meant the same future.

0.369

same-action open/closed divergence

Give open and closed states the exact same action pair; their final fields are separated by mean Hellinger distance 0.369.

0.074

future-map geometry correlation

The pattern of which action pairs lead to similar or dissimilar outcomes was almost unrelated across open and closed states. Interval [0.044, 0.104].

The gate changes the system’s interpretation of causes.

06 / Across the cohort

Fourteen of sixteen families expanded their raw future breadth.

g01-0.007
g02+0.009
g03+0.002
g04+0.016
g05+0.008
g06+0.006
g07+0.015
g08+0.008
g09+0.013
g10+0.039
g11+0.007
g12+0.000
g13+0.001
g14+0.023
g15-0.035
g16+0.011

Orange is broader when open; cyan is narrower. Family sizes range from one to four organisms, so the population summary above remains the main quantitative view.

07 / What we found

Information plasticity became developmental possibility.

We first found a causal-developmental gate. Then we learned to hold it open. Now the complete action map shows what the gate controls: the dimensionality, breadth, and causal organization of the organism’s reachable futures.

This is a concrete bridge from the Levin-style idea to Flow Lenia. A hidden informational configuration precedes morphology, can be causally manipulated, persists after release, and changes the space of forms that later interactions can produce.

08 / The next bold swings

Stop observing the repertoire. Start navigating it.