In a recent LinkedIn post, Lenny Rachitsky shares key takeaways from Zevi Arnovitz’s approach to leveraging AI in product management, emphasizing a structured methodology over impulsive building. Rachitsky highlights Arnovitz’s belief that effective AI coding hinges on thorough planning, a process that mirrors traditional software development lifecycles.
The Importance of Planning in AI Development
Lenny Rachitsky points out that Zevi Arnovitz advocates for a deliberate workflow when using AI for coding. This involves first creating an issue, then expanding on the idea with AI, constructing a detailed plan, executing that plan, and finally conducting multiple code reviews with different AI models. This structured approach, as detailed by Rachitsky, is crucial for preventing AI tools from generating code prematurely without a full understanding of the problem, thus avoiding bugs and technical debt.
“Effective AI coding is rooted in good planning—not jumping into building.”
Arnovitz’s workflow, as described by Rachitsky, begins with defining the problem clearly before engaging the AI for coding tasks. This initial planning phase is essential for ensuring the AI’s output is aligned with project goals and technical requirements.
Leveraging Slash Commands for Reusable Prompts
Rachitsky elaborates on Arnovitz’s innovative use of slash commands, which transform workflows into reusable prompts that offer compounding benefits over time. These commands, such as `/create issue`, `/explore`, and `/peer review`, embed best practices directly into the workflow. Each command is designed with placeholders for context and specific instructions regarding output format, tone, and process.
Iterative Improvement of AI Interactions
According to Lenny Rachitsky, Arnovitz actively refines these commands. When an AI, like Claude, makes a mistake, Arnovitz analyzes the cause within the system prompt or tooling and updates the command accordingly. This ensures that the error is not repeated in subsequent interactions, a practice Rachitsky emphasizes as key to continuous improvement.
“Slash commands turn workflows into reusable prompts that compound over time.”
Accelerating Learning Through AI
A significant theme in Rachitsky’s summary is the idea that the primary advantage of AI is not just faster building, but faster learning. Rachitsky explains Arnovitz’s use of a “learning opportunity” slash command. This command instructs Claude to explain its actions, effectively turning every bug fix and feature implementation into a learning moment for Arnovitz, who positions himself as a technical PM in training.
“The biggest unlock isn’t building faster—it’s learning faster.”
Addressing the ‘Slop Problem’
Lenny Rachitsky addresses what Arnovitz terms the “slop problem,” attributing it to human error rather than AI limitations. Rachitsky relays Arnovitz’s stance that shipping AI-generated outputs without proper review constitutes human error. Arnovitz takes full ownership of all outputs, viewing any inaccuracies as his own mistakes. He stresses the importance of guiding AI with detailed context, including writing style, problem definition, and constraints, concluding that the tool’s effectiveness is directly proportional to the operator’s skill and diligence.
As Rachitsky notes, Arnovitz believes that “The tool is only as good as the operator, and lazy prompting produces lazy results.” This underscores the critical role of the user in maximizing AI’s potential.
AI’s Impact on Career Growth for Junior Professionals
Finally, Rachitsky highlights Arnovitz’s perspective on how AI empowers junior professionals. According to Rachitsky, Arnovitz suggests that AI makes junior PMs more valuable by enabling them to handle a broader range of responsibilities, accelerating their learning curve significantly beyond traditional roles. The ultimate constraint, Arnovitz argues, is not a lack of ideas or strategic thinking, but rather a deficit in curiosity and the willingness to learn.
📝 About This Content
This article is based on insights shared by Lenny Rachitsky on LinkedIn.
📅 Originally posted on January 19, 2026 | View original post on LinkedIn →