Why Notion warns that AI token spend hides actual product ROI

This article summarizes an episode of Lenny’s Podcast’s video series  featuring Max Schoening, head of product at Notion.

Made by Tech in Asia with the help of Gemini

Max Schoening head of product at Notion, sees a massive shift in tech industry. As AI drastically lowers the need for deep technical skills, raw coding ability is no longer the main roadblock to success.

This shift changes how companies hire, speeds up product launches, and highlights the gap between simply generating code and actually building lasting software.

When action matters more than skills

AI has permanently removed the standard excuses for delayed projects. With instant code generation available, the focus shifts entirely to workers who take initiative and own the final results.

Schoening argues that lacking skills is no longer a valid excuses. “Before, it was very easy to always say, ‘I will never be able to do this because [it is a] skill issue,’” he notes. He adds that even with tools available, what matters most is agency, which is unevenly distributed.

He says that workers must act like founders instead of hiding behind job titles. Schoening shares that employees who take initiative and shape their environment will succeed, while others will face struggles.

Building independence at work

Finding these highly motivated workers is difficult, but teaching that mindset across a large company is an even bigger management challenge. Creating true independence requires building a work culture that actively encourages daily experimentation.

Schoening says that constant building breaks down a worker’s fear of failure. He notes that regularly experimenting and creating new tools builds a powerful momentum that helps employees realize they actually have the power to alter their company.

He pushes employees to take ownership of their projects. “I tell this to myself as well: ‘Do you drive Notion like it’s stolen?’” Schoening says, adding that employees can contribute with the agency even if they are not founders

The falling cost of starting new projects

Empowering workers to experiment lowers the cost of starting new projects. AI enables fast, low-cost prototypes, creating a flood of early tests. The advantage now belongs to teams that can polish those cheap tests into finished software.

“The first 10% of every project is now free,” Schoening says. “There’s no point in writing a PRD if you can just do the janky version and show, ‘Here’s the demo of what I think we should build.’”

While starting a project is cheap and fast, crossing the finish line remains brutal. “If you’re generous and say the first 90% is now done,” Schoening argues, “the last 10% is still actually 90%. That’s always the hardest.”

The upcoming costs of AI

Pushing those projects to the finish line requires a lot of testing, and running endless AI tests creates a financial drain. Counting the raw amount of computer processing gives leaders a false sense of safety. Eventually, strict budget reviews will clash with this fast building mindset.

Changing software as a user right

The hidden value of keeping systems running

Building lasting products around a single focus