Too much of a good thing?
Start seed 120426, let 120 ticks pass, then trigger Bloom. Observe grazers, hunters and subsequent resource pressure. Reset and replay with Drought instead. Note the delay between intervention and population response.
LOCAL SIMULATION / NO LOGIN / NO TRACKING
ARTIFICIAL
LIFE ZOO.
Give a creature energy. Give it an appetite. Give its world a few rules. Then observe how much can happen without a plan.
A SMALL WORLD
WITH CONSEQUENCES
VOL. 01 · 2026
WITH CONSEQUENCES
VOL. 01 · 2026
01 / LIVE HABITAT
The living petri dish
Preparing habitat…
HABITAT / 52 × 32 CELLS
SEED —
Population history
LAST 240 TICKSGrazers —Hunters —Scavengers —
Lineage / ancestry
—Lineages appear after simulation begins.
Parent IDs identify direct simulated descendants; this view shows recent births, not a claim of biological phylogeny.
Observation log
RECENT EVENTS / TEXT EQUIVALENT- Enable JavaScript to run the model. The guide and editorial material remain readable.
02 / THE MECHANISM
Complex behaviour.
Unromantic rules.
Nothing here is thinking.
Agents have positions, energy, type, a heritable speed trait and a unique ID. Every tick, they sense nearby cells, move toward a relevant target, consume resources, lose energy, sometimes reproduce and eventually die. An agent has no model of itself, no subjective experience, no intention and no claim to consciousness.
Emergence is an observation.
Population waves, clustered feeding and collapses can arise when local movement and resource renewal interact. Patterns that look coordinated do not require a coordinator. The model is deliberately toy-scale: it is not an empirical estimate of any real ecosystem.
Order matters.
The seed fixes initial positions, food distribution and the pseudorandom stream. Per tick: regrow food → shuffle a snapshot of agents → sense and move → eat or hunt → pay metabolism → reproduce → remove deaths → record totals. Capping agents prevents runaway browser load.
ENERGY ACCOUNTING / SIMPLIFIEDEnext = Enow + intake − metabolism × morphology
Reproduction splits a parent's energy with a child. Mutations change one speed trait within bounds; the same seed and action sequence reproduce the same trajectory on this code version.
03 / DO THE SCIENCE, NOT THE MYTH
Three field experiments.
One knob, new future.
Save a JSON state; capture comparison A. Change only metabolism, run another 80 ticks and capture B. Load the saved state and apply a different metabolism value. Compare the resulting histories instead of attributing every change to one species.
What makes a newcomer survive?
Introduce violet scavengers into an established habitat. Watch their births, energy and extinction risk, then repeat the same intervention from an identical saved state. This is a metaphor for ecological competition, not a prediction about invasive species.
04 / THE LIMITS OF THIS UNIVERSE
The simulation
is not the territory.
No genetics beyond one numeric trait. No disease, geography, seasons, mating, multi-resource competition, actual evolution by natural selection at realistic timescales, or empirical calibration. Variation is seeded pseudorandomness. A repeated seed is a reproducibility tool, not a guarantee of real-world prediction.
Artistically, this is a miniature bestiary. Scientifically, it is a transparent agent-based toy model. That distinction matters.