Hi Patrick and Brandon, I have promised to share AgentTask with you and get feedback, so now it is ready after a long wait. Please meet AgentTask, the product my company has been building. It's a shared place where people and AI agents plan, do and review work together. Built for that: agentic first, auditable and responsible AI. Please watch the video I'm sharing: the same task, with AgentTask and without it. With AgentTask: → Live progress: tasks and subtasks moving to done, with a comment and commit at each phase → Decisions written down with the exact "because" and "instead of", so the rationale isn't lost → The agent icon shows who's working on the task, last heartbeat and claim history → Independent review: a second Claude session, Codex, another model or a human can review the plan and final work, and discuss gaps and tests on the same ticket → Connect to projects so more people join, and link other work, notes and tasks → A live custom dashboard from your tasks, shareable with your team → Phone view, email notifications and a weekly report Without it: → Nothing was committed → No live view while it ran → A fresh session couldn't say if the tests passed → A teammate had nothing to look at AgentTask is free (free tier, no time limit). Pro and Team add more features and higher limits. Enterprise is for large organizations with strict data security. 🎥 Full 5-minute comparison: https://lnkd.in/gB_nPyUt 🔗 agent-task.com I also do consulting for businesses and enterprise companies as part of the AgentTask company: workflow automation, connecting and cleaning messy data, migration and integration with platforms like Salesforce and Kyriba. Need help? Let me know. #AgentTask #AIAgents #MCP #Automation

Does anyone use Claude Code and Codex together on the same machine and codebase, with each independently reviewing and challenging the other’s work to catch mistakes and improve the final implementation? If so, any tips?

Hi everyone, I've recently built an end-to-end AI-first task and agent management system http://agent-task.com/, and I'm releasing the mobile and desktop apps for test users. I wasn't sure if sharing details here would be considered advertising, so I wanted to check first. If the admins are okay with it (or don't object), I'll provide more information. The app is a productivity tool that lets you talk to it and streamline your to-dos. It includes mobile, desktop, and web access, plus the ability to invite other users. You can also connect it to AI-first applications like Claude, Codex, Cursor, and ChatGPT. I genuinely believe it can double your productivity.

Hi all, I am blocked with this error perhaps because of the following TODO: is there any ways you can help me getting unblocked? located at: site-packages/crewai/crew.py Line: 648 error: ```_execute_single_listener] Error in method write_chapters: Failed to convert text into a Pydantic model due to error: 'NoneType' object has no attribute 'function_calling_llm' Traceback (most recent call last): File "/Users/.... /src/write_a_book_with_flows/.venv/lib/python3.11/site-packages/crewai/utilities/converter.py", line 37, in to_pydantic result = self.model.model_validate_json(response) ``` More context: https://github.com/crewAIInc/crewAI-examples/tree/main/write_a_book_with_flows ``` for agent in self.agents: agent.i18n = i18n # type: ignore[attr-defined] # Argument 1 to "_interpolate_inputs" of "Crew" has incompatible type "dict[str, Any] | None"; expected "dict[str, Any]" agent.crew = self # type: ignore[attr-defined] agent.set_knowledge(crew_embedder=self.embedder) # TODO: Create an AgentFunctionCalling protocol for future refactoring. # <------THIS ONE if not agent.function_calling_llm: # type: ignore # "BaseAgent" has no attribute "function_calling_llm" agent.function_calling_llm = self.function_calling_llm # type: ignore # "BaseAgent" has no attribute "function_calling_llm" ```

VIEW RECORDING - 123 mins (No highlights) Key Takeaways - Google's Agent Development Kit (ADK) is gaining traction, offering a comprehensive ecosystem for AI development - Several members are working on innovative AI projects, from conflict resolution bots to Reddit analysis tools - The group discussed various AI tools, frameworks, and business strategies for AI-based applications Topics Brandon's ADK Demo and Updates - Demonstrated a voice agent using Gemini's newest voice models and ADK - The agent can interact with calendar events, creating/modifying/deleting appointments - Highlighted ADK's artifact versioning feature for managing document versions - Emphasized Google's upcoming announcements and potential game-changing releases Ongoing AI Projects - Michael: Developing a WhatsApp-based conflict resolution bot for divorcing couples - Jake: Created a Reddit analysis tool for comparing medications, with sentiment analysis and theme extraction - Marc: Taking courses on various AI frameworks, considering focusing on Google ADK - Paul: Finalizing a calculator tool with FastAPI backend and Lovable frontend