As of 4 June, Synthena still makes no artificial-life claim. We are not claiming the system is alive. What changed overnight is the shape of the problem: a vague “why are we plateauing?” turned into a mapped substrate bottleneck, with one real positive result and one important null.
The first substrate positive
A stateful drive with a small, bounded memory raised the expressible behavioural dimensionality of the organism from roughly 2 to 5. In plain terms: with that memory in the loop, the range of distinct behaviours the system can produce got measurably wider.
Two checks make it worth reporting. First, the lift is genome-driven: it comes from the inherited parameters, not from the measurement injecting it. Second, it is heritable, with parent and child fields staying highly similar across inheritance. A sweep over the number of internal modes also showed that 5 is not a free knob: adding more modes does not simply inflate the number. That is reassuring for the integrity of the measure, and it also tells us the current setup has a ceiling.
The null that mattered
Then the decisive test came back an honest null. Under priced fitness, where behaviours have to pay their way to survive, the richer dimension-5 behaviours were not selectable. Viable dimensionality collapsed back to 2.
That is the informative part. The missing piece is not more drive complexity on its own. The ecology itself only presents about two meaningful resource channels, so selection has only about two things to reward, and the richer behaviours have nothing to live on. Expressibility and selectability are separate bottlenecks: being able to express richer behaviour is not the same as the world being able to keep it.
What we can, and cannot, say
We can say that memory-bearing substrate dynamics express richer behavioural structure than the previous linear drive, that expressibility and selectability are separate bottlenecks, and that there is now a concrete next lever. The discipline held: no falsifier was weakened, no metric was promoted, and no artificial-life claim was advanced.
We cannot claim artificial life, open-ended evolution, or sustained novelty. The organism does not yet see the user: the receipt and lifecycle audit spine exists, but live physical input from a real headset and camera is still not wired in. And selection does not yet maintain the richer behaviours. That last one is exactly the next gate.
The next gate
Widen the world. Give the ecology at least five meaningful, priced resource channels, then rerun the stateful-drive selectability checkpoint. If selection can maintain the richer behavioural modes under that pressure, we move toward sustained novelty. If it collapses again, that is another honest null, and the substrate diagnosis tightens. Either way we will have learned where the wall is.
We’ll post the next update when there’s something measured to report.