Guide evolution toward increasingly complex functions in computational life

Ascertain whether there exist principled methods to steer the evolution of interacting, self-modifying program populations toward increasingly complex functions in the studied computational substrates, and, if so, develop and evaluate such guidance mechanisms.

Background

The experiments reveal that replicators can arise and dominate without explicit fitness landscapes, and further complex dynamics may follow. However, the paper does not investigate targeted interventions to promote specific functional complexity beyond spontaneous emergence.

This open question seeks strategies to guide or bias evolutionary trajectories in these systems toward richer capabilities while preserving their spontaneous, substrate-driven dynamics.

References

Several open questions arise from these investigations that warrant further investigations. Is there a way for us to guide the evolution of such systems into developing increasingly complex functions?

— Computational Life: How Well-formed, Self-replicating Programs Emerge from Simple Interaction  (2406.19108 - Arcas et al., 2024) in Discussion

History and growth: it produces copies but retains no history, and actual growth of complexity remained an open problem in the subsequent literature.

— Dalek: A Constructive Agent Machine  (2609.03546 - Xie, 3 Sep 2026) in Section 2.1, “The Inherited Construction”