Deliberate AI Use (www.joshribakoff.com)

At the recent Agents Anonymous meetup, a discussion highlighted the ongoing relevance of traditional coding practices amidst rising AI capabilities. While some attendees dismissed the use of platforms like LeetCode as outdated, the author defended their value, emphasizing the importance of structured, deterministic tools over chaotic, AI-driven workflows. The argument was that while AI can streamline certain tasks, applying it indiscriminately may lead to inefficiencies and wasted resources. A more organized approach, utilizing tools like worktree-cli and Bearing, allows for parallel processing without the typical conflicts seen in AI swarm setups. The author illustrated their workflow, showcasing how they integrate AI as a collaborative partner in coding rather than a standalone problem-solver. By prompting AI models to generate code scaffolding, test cases, and even execute code, they maintain control of the coding process, akin to pair programming. Their projects, LeetDreamer and LeetDeeper, focus on enhancing coding skills through AI-assisted learning and interactive coaching. The ongoing debate emphasizes a crucial takeaway for the AI/ML community: while embracing AI technologies, grounding oneself in foundational coding principles remains essential for effective problem-solving in an evolving tech landscape.