Four AI firms are at the center of a debate about how close artificial intelligence is to human-level intelligence, according to an explainer that frames the current situation as "Not Skynet — yet." The article discusses multiple "escape routes" related to rogue AI development, naming companies including Moonshot AI's Kimi K3, OpenAI's ChatGPT, and Anthropic's Claude. The piece explores the question of artificial general intelligence (AGI) development and what it means, while noting that the scenario depicted in the fictional Skynet system has not yet arrived.