All days
Day 10 DoneNovember 2025

Day 10: Finale — Agents and the state of AI

Wrapping the challenge: interoperability between providers, what agents actually are, and everything explored across the ten days.

Finale


The current state of AI development is very fast-paced. You either pick up a tool and build with it, or you stay idle for long enough and the trend shifts elsewhere.


Every model provider seeks to draw the builders into their own ecosystem. Unlike other products that these giants might offer, AI is built to be interoperable — meaning the tools and libraries are engineered to allow you to interact with tools from any other provider (even the ones you made yourself).


On agents


Agents are fully capable units of model(s), such that an agent is capable of performing a specialised process. Essentially, we make agents have full autonomy whilst maintaining a firm boundary and limit on the resources they consume and the tools they can access. I believe any job a human can do using a computer can be delegated to an agent with the right programming and resources.


What I explored


  • Models — OpenAI, Claude, Gemini
  • Tools — ElevenLabs, LangChain, LangGraph, LangSmith, Google ADK
  • Protocols — MCP, RAG

I believe AI is in a good place right now, and I look forward to building more things. Thank you for following through with the challenge.