DeepSeek outlines AI agent system aimed at cutting human oversight
Chinese artificial intelligence company DeepSeek has outlined a new design intended to make AI coding assistants more autonomous, allowing them to handle multiple tasks with less human intervention. According to a paper from the company’s DeepSeek Harness team, cited by Singapore’s South China Morning Post on Sunday, the goal is to reduce the amount of human effort needed to coordinate and monitor several AI agents at once. The team said multi-agent management is constrained not only by model capability but also by human attention, arguing that a single person can supervise only three to five active AI agent sessions at the same time. Under the proposed setup, tasks would be distributed across agents and their work coordinated by the system, while humans would focus on reviewing outputs and making decisions. The report said the proposal arrives as competition among AI developers intensifies, including among US firms Anthropic and OpenAI. Cointelegraph also noted that it reported in October that Chinese AI models had outperformed US peers in cryptocurrency trading tests, citing CoinGlass data, with DeepSeek posting the only positive unrealized return at 9.1%.
