Problem : Orchestrator agent delegates tasks to specialized sub-agents, a critic agent reviews the output -classic setup. But: every agent runs in isolation. None of them know what the others are working on, what assumptions they're making, or whether they're about to repeat the same mistake as the agent before them. The orchestrator ends up as a pure task router instead of an actual control layer. Solution: I built a shared state folder that all agents access through an MCP server. Each agent writes its current progress while working -not just at the end, but continuously: → The orchestrator sees in real time where each sub-agent stands, instead of only waiting for finished outputs → Sub-agents can read other agents' progress before starting - no re-deriving context that already exists → The critic agent has access to the full working history, not just the final result - reviews get sharper because it can see exactly where a decision was made The MCP server essentially turns the folder into shared memory between processes that would otherwise be completely stateless to each other. Result: Noticeably less redundant work between agents, and the orchestrator can actually orchestrate instead of just delegating and hoping for the best. This is my first pass at solving this. I'm sure there are more elegant ways to handle shared state between agents, and I'm genuinely curious how others are approaching it. Anyone else building similar shared-state mechanisms?

What AI tool are you currently learning or experimenting with?

Anthropic just had the busiest 48 hours of its year. Yesterday: Claude Sonnet 5 dropped. Near-Opus performance, way cheaper, now the default model for free/Pro users. Today: Fable 5 comes back online after being pulled by a government export order weeks ago. One company, two "new model" stories in the same week — for very different reasons. Which one are you more curious about? 👇

The Problem- 1.)A support Agent easily spends 2-4 hours answering the same refund policies & the shipping queries 2.) For many businesses, 50–80% of support emails are repetitive FAQs. 3.)Businesses take 4–24 hours to reply to customer emails while a customer wants it as soon as possible. 4.)When your business grows from 80 to 400 customer emails a day you have to hire more people to keep up with the task & provide the same accurate, answer every single time. Now, The solution: For a business getting 80+ support emails per day- 1.) 40-60 emails are repetitive FAQs and policy related queries which can be handled automatically which can save 3-5 hours of manual work every single day, which is 21 -35 hours a week. 2.)Because of the Automation response time can drop from 4-24 hours to under a minutes which is huge. 3.)Support quality will be more consistent, with the same policy-based answer every time and as the business grows you can avoid hiring a new person. Now let me walk you through a demo of this Agent 👇 in the video below. Now if you want it to use it or test it comment "Support" and i will send it to you. Harsh Singh

Excited to share my Day 1 build for the #AISChallenge. Today I created an AI automation newsletter that explains how businesses can reduce repetitive work using simple automation workflows. 💡 One thing I learned I realized that great AI content isn't just about writing prompts. It's about communicating ideas clearly. A clean layout, good visual hierarchy, and simple language make technical concepts much easier for people to understand. 🚀 One thing I'd improve On the next iteration, I'd make it more personalized by adding stronger branding, more real business examples, and a clearer story that connects with business owners. I'm taking this challenge because my goal isn't just to learn AI tools. My goal is to build practical AI automation solutions that help small businesses save time, reduce repetitive work, and focus on what matters most.