- The first 10% of project development is now a commodity, creating a 'demos not memos' culture.
- Software quality is suffering as organizations optimize for token spend and feature velocity rather than engineering substance.
- Developing product 'taste' is equivalent to training a machine learning model; it requires high-frequency iterations and constant feedback loops.
- Agency is not evenly distributed; the most successful contributors are those who view software as material to be manipulated and ignore traditional organizational constraints.
Back to Feed
Why cultivating agency matters more than cultivating skills in the AI era | Max Schoening (Notion)
Notion's head of product, Max Shing, discusses how AI is transforming product development by shifting the focus from specific roles to individual agency and taste. He argues that the future of successful software lies in creating malleable tools that empower users rather than just shipping features.
Key Takeaways
- True product innovation today hinges on individual agency and the mastery of the development medium rather than just shipping code to production.
- Successful products are defined by a tiny, exceptionally good core superpower, while feature bloat represents a failure of design strategy.
- AI-driven product work requires managers to move away from static documentation toward iterative, hands-on prototyping in code.
Talking Points
Analysis
This conversation is critical for technical teams navigating the shift from manual coding to AI-assisted workflows. It provides a necessary counter-narrative to the 'feature factory' model currently plaguing many startups.
- Why this matters: The debate centers on the difference between superficial speed (shipping features) and structural progress (building reliable, extensible systems).
- Who benefits: Product leaders and engineers struggling with 'vibe coding' or feeling pressured to constantly adopt new AI tooling without clear business objectives.
- Contrarian takeaway: Shing suggests that inclusivity and total democratization might be counter-productive to building high-quality, 'obviously good' software. He argues that products should cater to the 'top of the class' rather than attempting to serve every user, a stance that challenges the growth-at-all-costs mindset.
Time saved:
Back to Feed