According to AI researchers and industry experts, the development of AI super-agents capable of performing complex, multi-step tasks autonomously could become a reality by 2025, primarily due to increasing computing power availability. These systems would combine various AI models and capabilities to handle sophisticated tasks like travel planning, software development, and business operations. Companies like OpenAI, DeepSeek, and Anthropic are actively working on such agents, with DeepSeek’s CEO suggesting that current hardware limitations are the main barrier to their development. The article highlights that the necessary computing infrastructure is rapidly evolving, with expectations that sufficient computational power will be available within 12-24 months. These super-agents would differ from current AI assistants by having improved reasoning capabilities, better memory management, and the ability to break down complex tasks into manageable steps. However, experts also warn about potential risks and challenges, including safety concerns and the need for proper governance frameworks. The development of these systems could significantly impact various industries and potentially lead to major workforce transformations. The article emphasizes that while the technology is promising, careful consideration must be given to its implementation and regulation to ensure responsible development and deployment.
Source: https://www.businessinsider.com/ai-super-agents-enough-computing-power-openai-deepseek-2025-3