Andrew Ng explained that he often uses AI as a “brainstorming companion” while driving, using voice-mode chats to have extended back-and-forth conversations rather than quick one-shot prompts. He switches among different AI tools especially ones like Claude Code and OpenAI Codex—to leverage each model’s different strengths for tasks such as coding. He finds that long, context-rich interactions yield better results than minimal prompts, though sometimes a brief “lazy prompt” can surprisingly be more efficient if the model infers the user’s needs. When he reaches his destination, he has the AI summarize the session and sends that summary to his team. His approach reflects a broader shift among leaders using multiple AI tools integrated into daily workflows for coding, ideation, and productivity.
