The author shares a range of projects accomplished with coding agents over the past month, including automating invoice processes, creating hydrological simulations, and developing machine-learning models for predicting US corn yields. They highlight their collaboration with tools like Gemini 3 Pro, OpenAI Codex, and Claude Opus 4.5, arguing that these systems demonstrate capabilities resembling artificial general intelligence (AGI). The author emphasizes that while AGI represents a significant milestone, it also marks the beginning of a broader journey in utilizing AI to transform various fields. Coding agents are framed as powerful tools that can benefit not just coders but a wide array of professionals, from scientists to diplomats. The author identifies command-line interfaces as the most effective way for beginners to engage with coding agents, despite their complexity. They provide a straightforward example of using a command line to replace text in a file, illustrating how efficient this method can be compared to graphical interfaces. The article also touches on the importance of diverse users experimenting with these technologies, suggesting that those who adapt best will thrive in the evolving landscape of AI. The author positions themselves as a "humanities person" with coding skills, aiming to bridge the gap between technical and non-technical audiences. They plan to introduce coding agents in a way that assumes no prior knowledge, making the technology accessible to a broader audience. The article hints at future developments in AI and coding agents, suggesting that innovation will depend on how effectively individuals and organizations utilize these tools for real-world applications.