nexcorp’s Post

PREDICTION: The future is not just about agents, its about networks/systems of agents where agents can discover and collaborate with other agents. Here's where we are now: We have made great progress this year in creating single purpose agents like customer service, company research, etc. This has been made possible by better and better reasoning capabilities in the LLMs, better structured output. Better function/tool calling. More advanced RAG (Retrieval Augmented Generation) -- and in some cases, even early implementations of multi-modality (audio/video support). That's all *awesome*. I'm here for all of it. But next year, we're going to go much further. 2025 is going to be the year of multi-agent *networks*. Agents will be able to "declare" their capabilities, their inputs and outputs and highlight their "experience". In the same way that current LLMs support tool calling (whereby a set of tools is made available to the LLM, which it can then use to formulate a more useful response), in the future *agents* will be provided access to a network/system of other agents. It can then invoke these other agents in order to handle some number of tasks that the agent can't do itself. There are already powerful dev frameworks for building these kinds of multi-agent systems (CrewAI comes to mind), and there are low-code platforms for building this -- like Agent.ai, that I'm working on. These are early days, and a lot has to happen to fulfill this network of agents vision. But it feels like we're moving that way very, very quickly. Excited to be part of this movement!