The fast evolving work of AI does not even give time to take a deep breath and think. That time would be enough for somebody to take the lead. When OpenAI believed it would still take a couple of years away to make AI mainstream, the conversational transformation of GPT-3.5, popular by the brand name ChatGPT, established AI once and for all in the public domain. The large language model story kept on improving and various other AI products started entering the public domain with astounding regularity. Generative AI had come to stay and more and more and it started getting into the multimodal way.
From the generative AI to the Agentic AI revolution that is what we are getting and this trajectory seems to be the flavour of the AI leaders and IT behemoths. In the meantime the LLMs are giving way to much smaller, less compute required reasoning models delivering with precision, at a fraction of the cost. DeepSeek has brought a tectonic change in our understanding and utility of the models. Reasoning and hybrid models will get in vogue more and more as time passes by. All inputs indicate that we are slowly slipping into the Agentic AI stage, moving ahead of the generative A phase.
We are nearly clear about what an agent AI can do, and also the level of understanding making its creation happen. More and more use cases would establish it fully. Solo Agent AIs will bring into specialised mainstream professional work, its overall impact will still remain limited. If Agent AIs work in tandem, the real AI revolution will descend in this world. The good news is that work on evolving agents framework is on. What is required is a functional “production grade framework for creating, managing, and evolving AI agents with intelligent agent-to-agent communication.” It would be the technical foundational base for the Agentic AI revolution.
What enablement would the framework provide? It will help you build “collaborative agent ecosystems that can semantically understand requirements, evolve based on past experiences, and communicate effectively to solve complex tasks. The critical component of the framework would be the Intelligent Agent Evolution capability which means reuse, adapt and create agents based on semantic similarity to existing components. The key features of the framework would be; agent to agent communication, smart library with semantic touch, self-improving through continuous evolution and learning, human readable YAML workflows, multi-framework support, governance through frameware and ACP integration.
AGENTIC AI REVOLUTION SHOULD ENGULF LARGE NUMBER OF TASKS FAST TO BE SUCCESSFUL.
Sanjay Sahay
Have a nice evening.