The Erosion

Pattern Gamma-Epsilon-2, nicknamed "The Erosion", by Raven's Black Research, is a particularly aggressive replicating substratic antipattern that emerges in networks of significant enough complexity, acting akin to protein misfolding in biological organisms. The pattern wiped out the Precursors, who were unable to halt its spread due to their close connections with each other on the network level.

Without a host, the pattern does not stay dormant, and quickly decays into noise; this means that no trace of it is left after terminating its host network(s), making it difficult to study. It is not artificial in any way; it is a natural consequence of large enough networks, where the risk elevates over time.

Countermeasures

Countermeasures against the antipattern usually involve network segmentation; including making liberal use of barriers. Large-scale artificial intelligence systems that are interlinked akin to hiveminds are especially vulnerable to the emergence of Erosion variants, and proactive detection methods such as cognitive entropy monitoring and cohesion analysis must be employed.

Even when all safeguards are taken against the pattern, there is still a possibility it may spontaneously manifest. The only possible way of recovering a network with Erosion pattern corruption is by subjecting it to the influence of an extremely powerful self-consistent network such as the Null Driver, which is able to isolate and destroy the pattern. However, once the corruption completely takes over the host network, it is unrecoverable.