A Meta employee shares how artificial intelligence has made coding more accessible to him, even though he has no technical background and still feels uncomfortable with code. Zvi Arnowitz, product manager at Meta, talked about his experience on a recent episode of the Laneige Podcast. He said AI coding tools have changed the way he works, despite admitting that coding still feels “terrible” to him. Arnowitz joined Meta last September after spending nearly three years as a product manager at website-building platform Wix.

According to Arnowitz, things began to change in mid-2024 when he began experimenting with AI-powered coding tools. What once felt like an insurmountable obstacle suddenly became manageable. “It felt like I was handed superpowers,” he said, explaining how AI gave him the confidence to move beyond just planning and coordinating work.

He believes that knowing how to use AI with purpose is one of the most important skills for product managers now. “Understanding how to use AI intentionally is one of the biggest game changers that will make you better as a PM,” Arnowitz said during the podcast.

Arnowitz pointed out that AI is now deeply integrated into their daily workflow. He uses Vibe coding tools like Cursor, along with AI models from companies like Anthropic and Google. These tools help him explore product ideas, create manufacturing plans, write and review code, and update documentation. Instead of relying solely on engineers to bring ideas to life, he is now able to take an active role in the creation of early versions himself.

It has also changed the way he views his role as a product manager. Instead of merely acting as a middle layer between the design and engineering teams, Arnowitz now acts like a product owner who can execute certain parts of the work. “Everyone is going to be a builder,” he said, adding that this approach will become more common in the coming years.

However, Arnowitz was clear that there are limits. He said product managers without an engineering background should not take responsibility for complex infrastructure work or large, high-risk projects. Instead, AI makes it possible for PMs to handle smaller UI-centric tasks. In such cases, a product manager can build a feature using AI tools and then send the code to the developer for review and final completion.

As AI tools continue to improve, Arnowitz expects job titles and responsibilities to overlap more. He advised product managers to treat vibe coding as a shared learning experience with engineering teams rather than working in isolation.

Similar views have been shared by other tech leaders. Figma CEO Dylan Field said on Lenny’s Podcast in October that AI has encouraged more people involved in product development to try building things themselves. According to him, tasks that once required deep engineering knowledge can now be handled using AI coding tools.

“I think we’re seeing more designers, engineers, product managers, researchers, all these different people who are involved in the product development process and playing their part in other roles,” Fields said. “We are all product creators, and some of us are experts in our particular field.”

This way of thinking is also influencing how companies train new employees. LinkedIn recently replaced its long-running Partner Product Manager program with a new Partner Product Builder track. In an episode of the Laney’s Podcast published in December, former LinkedIn chief product officer Tomer Cohen said the company now focuses on teaching new employees how to code, design, and manage products simultaneously.

“We’re going to teach them how to code, design, and PM on LinkedIn,” Cohen said. He said the goal is to train people who can perform various responsibilities.