Food the existing programs could not use. A combination that could.
Working manuscript · v0.1
THE BARRIER
Organisms in this digital world inherit programs that process food. I added a nutrient the existing programs could not use, then let mutation combine those programs at birth. The experiment asks whether a new combination can use that food.
A DIFFERENT WORLD
I built a new metabolic ecology with a deliberately engineered nutrient. To process it, a program had to return min(65535, a + (a XOR b)) on all 64 registered input pairs. All 89 typed primitives failed at least one pair. This is a claim about this library and interface, not every possible program.
BORN HERE
Founders carried mask_xor and low uptake. Mutation could combine two cells at birth from 84 composable primitives. Of 1,825 distinct expressions evaluated during mutation, one passed: add_sat(a, mask_xor(a,b)). It arose more than once, but those origins were the same kind of capability.
WHAT I PUT INTO THE WORLD
I designed the resource and knew a successful expression as a validation witness. I did not preload it into the evolving population or supply a pool screened for success. The result is online assembly within a finite, designed possibility space.
DISCOVERY IS ONLY THE START
Forty worlds ran for 3,000 ticks in 256 slots: ten each with full evolution, uptake mutation disabled, no substrate, and atomic-only mutation. The capability appeared in 3/10 full worlds and persisted in 2/10. With uptake mutation disabled it appeared in 2/10 and persisted in 1/10, against a registered retention requirement of 8/10.
THE CONTROLS
Without substrate, the capability appeared in 3/10 worlds and established in none. With composition disabled, it never appeared. These controls help interpret the resource dependence; they do not compensate for the failed retention gate.
TOO EARLY FOR THE TEST
A later high-uptake mutation appeared in each of the two successful full worlds. But the events arrived 19 and 4 ticks before the B lineage completed establishment: at least 20% of slots for 50 consecutive completed ticks. Neither was an eligible candidate under the registered rule.
THE RESULT
Online mutation crossed the engineered capability barrier. The registered ratchet criterion failed, and conditional dependence remained untested: zero eligible candidates meant zero primary factorial or historical-fork assays. An early event is not a failed dependence test, and it cannot be promoted into a confirmatory candidate after seeing its outcome.
Must an invention be established before something can build on it?
EX-12 asked that question under a new prospective candidate rule, in fresh discovery worlds. Open-ended evolution remains unestablished.
Lab notebook / Working draft
Conclusion
Mutation crossed the engineered food-processing barrier. The lineage did not meet the registered persistence criterion often enough, and later changes arrived before they were eligible for the dependence test.