How a developer used Google Gemini to replace his Kanban board

At a recent tech conference in San Francisco, I encountered Richard Murby, the business development director at Devpost, a platform dedicated to helping developers engage in hackathons. Our conversation delved into his innovative use of AI, particularly Google Gemini, in enhancing his work processes. Richard shared that he employs AI primarily to boost his productivity, maintain organization, and spark creativity. By training a Google Gemini model on a collection of his most effective emails, he has created a collaborative tool that assists him in crafting important communications. This model retains his unique voice and ideas while enhancing the clarity and impact of his messages. Moreover, Richard has engineered a Google Gemini Command Line Interface (CLI) that serves as a modern alternative to his traditional Kanban board. This markdown-based task manager integrates seamlessly with his calendar and email, allowing it to prioritize tasks based on real-time context from his daily schedule. Impressively, the system can conduct deep research on specific tasks and even draft emails directly into his notes, facilitating a more fluid workflow. Richard emphasized the ongoing dialogue with the AI as a significant advantage, expressing excitement about the potential for further expanding its capabilities. He highlighted two key areas where AI has profoundly influenced his work: first, by enabling Devpost to distribute AI tools effectively to developers worldwide, and second, by significantly enhancing his writing quality through real-time feedback and learning from past communication successes. However, he acknowledged that the journey to adopting AI isn't without challenges. One major hurdle is finding the time to explore and experiment with AI applications. Richard encourages developers to think critically about which tasks can benefit from AI and to embrace the trial-and-error process that leads to discovery. He draws parallels with hackathons, which provide crucial opportunities for testing ideas, failing, and learning in a supportive environment. For those who may not be as technically inclined, Richard offers simple advice: focus less on understanding the intricacies of AI and more on exploring its capabilities. He suggests being descriptive in prompts, asking insightful questions, and providing context to the AI. While immediate results may not be apparent, the combination of experience and AI context will yield time-saving benefits and enhanced performance over time.