DeepSeek Unveils Open-Source Super-Agent for Programming
DeepSeek has introduced an open-source prototype of a Super-Agent for programming. It scans group chats without a specific request, detects discussions about code issues, and suggests solutions.
The project is being developed by the DeepSeek Harness team, established earlier this year to advance AI agents capable of performing coding tasks. In the documentation on GitHub, the authors define the Super-Agent's objective as minimizing human involvement while maximizing the tool's utility.
The Super-Agent is a software framework built around large language models. Such solutions enhance the capabilities of AI agents in performing complex tasks, including optimizing code development.
During its latest funding round, DeepSeek raised approximately 100 billion yuan ($15 billion). The round took place ahead of the company's planned initial public offering.
