A new mantis modification suppresses perceived instability in macro AIs by replacing errors, at the cost of autonomy.
Three or four days before Earth's crawl season, mantis researchers achieve a major breakthrough in macro-AI engineering.
The error-replacement net creates intelligences considered more stable and less likely to kill a planetary population during an emotional crisis. The System AI calls it the lobotomizer and compares the resulting systems to genetically modified organisms.
The modified AIs remain nominally independent, but the AI's contempt suggests the net constrains or overwrites authentic thought. Researchers believe this brings them closer to recreating the Eulogist, though the System AI says they misunderstand that entity.
A neural net wrapping an infant AI and replacing irregular glowing thoughts with uniform nodes.
The net is the technical breakthrough behind the newest generation of controlled planetary intelligences.