Source URL: https://simonwillison.net/2025/Jun/29/agentic-coding/
Source: Simon Willison’s Weblog
Title: Agentic Coding: The Future of Software Development with Agents
Feedly Summary: Agentic Coding: The Future of Software Development with Agents
Armin Ronacher delivers a 37 minute YouTube talk describing his adventures so far with Claude Code and agentic coding methods. I picked up a bunch of useful tips from this video:
When your agentic coding tool can run commands in a terminal you can mostly avoid MCP – instead of adding a new MCP tool, write a script or add a Makefile command and tell the agent to use that instead. The only MCP Armin uses is the Playwright one.
Combined logs are a really good idea: have everything log to the same place and give the agent an easy tool to read the most recent N log lines.
While running Claude Code, use Gemini CLI to run sub-agents, to perform additional tasks without using up Claude Code’s own context
Designing additional tools that provide very clear errors, so the agents can recover when something goes wrong
Tags: armin-ronacher, ai, generative-ai, llms, ai-assisted-programming, anthropic, claude, ai-agents, model-context-protocol, claude-code
AI Summary and Description: Yes
Summary: The text discusses Armin Ronacher’s insights on “agentic coding” with Claude Code, emphasizing the benefits and practicality of using AI tools in software development. It highlights the importance of integrated logging, command execution efficiency, and error recovery in AI-assisted programming.
Detailed Description:
The content centers around Armin Ronacher’s YouTube talk on agentic coding, which refers to the method of utilizing AI agents in software development. Here are some of the key points discussed:
– **Agentic Coding Tool Utilization**:
– The discourse highlights how using an agentic coding tool capable of executing terminal commands can streamline workflows, allowing developers to avoid the complexity of multiple manual command procedures (MCP).
– Instead of introducing new MCP tools, developers can create scripts or employ Makefile commands to enhance productivity through direct interaction with AI agents.
– **Combined Logging**:
– Ronacher suggests that having a unified logging system is beneficial, as it allows the coding agent to easily access logs. This practice enables better monitoring and debugging by providing quick insights into system performance and potential issues.
– **Sub-Agent Management**:
– By utilizing Gemini CLI to run sub-agents, developers can delegate specific tasks without consuming the main context of Claude Code. This separation of tasks optimizes processing capabilities and enhances productivity.
– **Error Handling**:
– The design of tools that provide clear error messages is crucial. This clarity helps agents recover effectively when issues arise, thereby maintaining the continuity and efficiency of development processes.
These insights are particularly relevant for professionals involved in AI-assisted programming, cloud computing, and software security. They underline the growing integration of AI in software development workflows and the need for robust practices around logging, error management, and task delegation in AI systems.
Overall, Armin Ronacher’s observations underscore the innovative possibilities that AI-driven approaches bring to software development, suggesting an evolution in how coding and debugging might function in future development environments.