When an AI tool makes a mistake, treat it as a learning opportunity for the system. Ask the AI to reflect on why it failed, such as a flaw in its system prompt or tooling. Then, update the underlying documentation and prompts to prevent that specific class of error from happening again in the future.
As AI democratizes the ability to create products, rigid job titles like "Product Manager" and "Engineer" will become obsolete. Meta PM Zevi Arnovitz predicts that responsibilities will merge, and the focus will shift to the act of creation. In the near future, everyone on a product team will simply be a "builder."
Default AI models are often people-pleasers that will agree with flawed technical ideas. To get genuine feedback, create a dedicated AI project with a system prompt defining it as your "CTO." Instruct it to be the complete technical owner, to challenge your assumptions, and to avoid being agreeable.
Don't get hung up on the cost of AI credits and subscriptions. Instead, reframe the spending as "tuition" for your professional development. This mindset shift encourages the experimentation and hands-on learning necessary to master these new tools, providing a far greater return than pinching pennies on API calls.
Treat different LLMs like colleagues with distinct personalities. Zevi Arnovitz views Claude as a collaborative dev lead, Codex (GPT) as a brilliant but terse bug-fixer, and Gemini as a creative but chaotic designer. This mental model helps in delegating tasks to the most suitable AI, maximizing their strengths and mitigating their weaknesses.
While junior roles may be contracting, AI provides an alternative path for new graduates. For the first time in history, a junior individual can single-handedly build and launch a fully-fledged startup. This empowers them to gain experience, build a portfolio, and bypass the traditional entry-level job market.
Create a reusable prompt (a "slash command") that explicitly asks your AI coding assistant to explain complex technical concepts. Frame the prompt with your current knowledge level (e.g., "explain this to a technical PM in the making using the 80/20 rule"). This transforms every coding session into a valuable learning opportunity.
For non-technical individuals intimidated by code, a gradual approach is key. Start with a simple chat UI like a ChatGPT project, then move to guided builders like Bolt, and finally graduate to a professional IDE like Cursor, initially in light mode. This "exposure therapy" builds comfort and confidence over time.
While tools like Lovable and Bolt are excellent for beginners, they trade control for simplicity by making opinionated technical choices (e.g., database, auth) on your behalf. For complex, production-grade applications where you need to make specific architectural decisions, graduate to a more direct tool like Cursor.
Meta PM Zevi Arnovitz details a 7-step workflow using slash commands in Cursor, moving from ideation (`/create_issue`) to planning, execution, and finally a multi-layered code review (`/peer_review`) and documentation (`/update_docs`). This structured process enables non-technical individuals to build and ship complex applications entirely on their own.
To overcome the challenge of reviewing AI-generated code, have different LLMs like Claude and Codex review the code. Then, use a "peer review" prompt that forces the primary LLM to defend its choices or fix the issues raised by its "peers." This adversarial process catches more bugs and improves overall code quality.
Go beyond standard AI mock interviews. When Meta PM Zevi Arnowitz struggled with product segmentation questions, he used a low-code tool (Base44) to build a custom quiz game. The app generated practice questions he could drill on his commute, turning a weakness into a strength through targeted, interactive practice.