## Survival of the Stable How does a lifeless world, with no plan and no designer, end up overflowing with intricate living things? The answer begins with something far humbler than life: some arrangements of matter simply **last longer** than others. Long before any biology, chemistry was already throwing together countless structures. Most fell apart almost at once; a few happened to hold together. Nothing *chose* them. That is the whole idea of **survival of the stable**: if a pattern persists while its neighbours dissolve, you are simply more likely to still find it around. Picture a windy beach covered in shapes scratched into the sand. Most are wiped out in seconds. A few, by sheer luck of how they sit, resist a little longer. Nobody is selecting them on purpose. **Persistence itself does the filtering.** <viz id="0"></viz> **Watch the field form and thin out.** New arrangements keep appearing, but check the **counts at the bottom-left**: fragile and moderate shapes barely stay *alive* even though many *form*, while stable ones pile up. **Drag the spawn-rate slider**, or **click any shape** to see how long it lived. This is not **survival of the fittest** yet, just survival of whatever physically hangs together. It matters because evolution never starts from nothing. It starts in a world where some structures already outlast the rest. The next step is bigger. A stable thing merely lasts, but a **replicator** can make more of its own pattern, turning persistence into multiplication. <ref slide="2">What a Replicator Is</ref> ## What a Replicator Is A **replicator** is a pattern that makes copies of itself. That sounds simple, but it changes everything. A durable object survives by continuing to exist as one thing. A replicating pattern survives in a different way: even if particular copies are destroyed, the **lineage** can continue. This is the big shift from mere stability to evolutionary potential. A rock can last. A replicator can spread. Once copying enters the picture, what matters is not only how long one instance survives, but whether the pattern keeps reappearing through descendants. <viz id="1"></viz> **Press play** and compare the two sides. **Raise the destruction rate** and then **increase the copy rate** on the replicator side. Watch how individual copies can die while the overall pattern still remains present through new copies. The important insight is that replication creates a new kind of endurance. The pattern is no longer tied to one physical object. It can survive as a chain of copies across time. That is why replicators matter so much in evolutionary thinking. Once copying exists, small advantages can accumulate across generations instead of being trapped in one short-lived object. But copying is never perfect, and those small mistakes turn out to matter enormously. <ref slide="3">When Copies Make Mistakes</ref> ## When Copies Make Mistakes Copying is never perfect. Every so often a copy comes out a little different from its parent — a small error, a **mutation**. Most copies are faithful, but these rare mistakes are where variety comes from. <viz id="1"></viz> **Press play** and watch copies stream out. **Raise the mutation rate** to make mistakes more common. Most mistakes are harmless or bad and fade away — but a **good** mistake copies a little faster, so it leaves more descendants and slowly spreads through the population. Some mistakes are bad: the variant copies poorly and dies out. Some are good: the variant copies a little better, so it leaves more copies than its neighbours. Nobody chooses this — the good mistakes simply accumulate because they out-copy the rest. That is the engine of change. Imperfect copying supplies the variety, and differences in copying success decide which variants become common. Next we look at exactly which traits make a copy more successful. <ref slide="4">Three Replicator Advantages</ref> ## Three Replicator Advantages Not all replicators are equally successful. Three properties matter especially much: **longevity** (lasting long enough to copy), **fecundity** (making many copies), and **copying fidelity** (producing copies that stay similar enough to keep replicating). These are the classic advantages that make one replicator become more common than another. You can think of them as three levers. If a pattern survives longer, it has more chances to copy. If it copies faster, it can spread sooner. If it copies accurately, its useful structure is preserved instead of dissolving into noise. Even modest improvements can compound over time. <viz id="2"></viz> **Give each replicator its three traits**—longevity, fecundity, and fidelity—then **press play**. Raise all three and a population climbs and holds high; weaken any single one and it settles lower or dies out. **Watch the population chart** to compare the two. A useful way to read the simulation is to ask: which trait increases the number of future copying opportunities? Longevity protects opportunities, fecundity creates them quickly, and fidelity preserves the pattern that makes them possible in the first place. This is where the lesson begins to feel more like algorithmic scaling. A tiny edge per round may look unimpressive on its own, but after many rounds it can dominate. Raise all three advantages together and a replicator surges ahead; weaken any single one and it falls behind. These small, compounding differences in copying success are what decide which patterns become common — the foundation of everything we call evolution.