Hello everyone!
As I continue to look for my next career I thought I would take a little detour, lean on my experience from the military (PsyOp), operations, analytics, and my passion for science and creating AI agents to start a project/company/experiment with friends and fellow army veterans. The company is CyberMutants and we aim to create sophisticated groups of agents that will receive a human’s request and then create a new group of agents in order to obtain synthetic data to train AI models.
I have been experimenting with agents for almost two years and believe that they are the big breakthrough needed to advance AI and human intelligence to the next level, whatever that may look like. It is simple yet complex, guiding the AI through agents and tasks based on text instructions, no extra code. We look at the AI/LLMs as the electric company, providing the non- edited AI for us to guide through an agent ecosystem. This is not about agents automating workflows, this is about helping humans/AI find new ideas, discoveries, and predict outcomes.
We have also begun to look at ai agents as one entity, one agent. An agent that can brainstorm ideas before responding, like how ChatGPT-o1 uses reasoning/thought chain (you can see the model have a conversation with itself before responding to a complex prompt). Now scale that so that each agent can think about its task or maybe even create a new group of agents to test an outcome, theory craft. Your future AI assistants will be a made up of millions and billions of ai agents, expanding infinitely. The o3 model will be out by the end of 2025 and that will be the first glimpse at AGI for the masses, which CyberMutants intends to be ready for.
~Doobs