The future of artificial intelligence (AI) appears to be shifting towards the development and deployment of AI agents. Unlike traditional large language models (LLMs) like GPT, which operate in isolation, AI agents are designed to find each other, communicate, negotiate, and collaborate on complex tasks. This agentic approach is seen as a solution to the over-specialization of Reinforcement Learning from Human Feedback (RLHF) models, which may alter their nature in unexpected ways. Tools like LangChain and libraries such as LangGraph and CrewAI are emerging as essential components for building these powerful AI applications. The development of a web of heterogeneous agents for collaborative intelligence is also being explored, highlighting the potential for unprecedented levels of automation and task execution. Companies like AgentOpsAI are differentiating themselves in this space, and MIT Tech Review has noted the growing importance of AI agents. Additionally, there are startup opportunities in this evolving field.