Chinese AI company DeepSeek has presented an architecture designed to increase the autonomy of coding assistants capable of performing multiple tasks with less human intervention. According to a paper from the DeepSeek Harness team, a person can oversee only three to five active AI agent sessions at the same time. The proposed system would distribute tasks among agents and coordinate their work, leaving humans to review results and make decisions. The initiative comes amid intensifying competition between DeepSeek, Anthropic and OpenAI, following a crypto trading experiment in which DeepSeek generated a 9.1% unrealized return, the best performance by an AI model in the test.
Source: Read the original article


