Why cultivating agency matters more than cultivating skills in the AI era | Max Schoening (Head of Product, Notion)

May 3, 2026 Episode Page ↗
Overview

Max Schoening, Head of Product at Notion, discusses how AI is transforming product development, emphasizing the importance of agency and malleable software. He shares insights on empowering teams to build and iterate faster, and his perspective on the future of work in the AI era.

At a Glance
14 Insights
1h 27m Duration
27 Topics
7 Concepts

Deep Dive Analysis

Introduction to Max Schoening and AI's Impact

Origin of Designers and PMs Coding at Notion

Current State and Future of Designers Shipping Code

Importance of Agency in the AI Era

Examples of High Agency at Notion

Potential Loss of Specialists as Roles Merge

Advice for Developing Personal Agency

Understanding Malleable Software

Dieter Rams' Design Philosophy and Usefulness

The SaaS Apocalypse Debate

How Product Building Has Changed with AI

Next Shifts in Product Building Modalities

Token Spend and ROI Conversations

Strategies for Changing Work Habits

Max's AI Tool Stack

Future Transformation of Roles by AI

Notion AI's Success Factors

Shipping Quickly While Maintaining Quality

Building Taste Through Iterations

Core Elements of Successful Products

Applying the Jobs-to-be-Done Framework

Hot Take on Universal Basic Income

Max's Vision for AGI Use

Contrarian View on Inclusivity

Lessons from Career Failures

Advice for Young People in Silicon Valley

Lightning Round and Final Thoughts

Agency

The ability to understand that the world around you is malleable and that you can change things, rather than being limited by perceived skill issues or rigid job descriptions.

Malleable Software

Software that works closer to the interests of its users than the corporation that makes it, allowing users to customize and own their computing experience without starting from scratch.

Tiny Core Superpower

The concept that truly great products possess one exceptionally good, fundamental feature or idea that provides immense value, around which everything else flows.

Incremental Correctness

A core value emphasizing continuous iteration and improvement, where one increases 'shots on goal' by experimenting frequently and refining ideas over time.

Taste (in Product)

The ability to run a 'virtual machine' in one's head to predict how a specific in-group will react to an idea, developed through extensive iterations with feedback.

Obviously Good Products

Products that are so inherently excellent and useful that their quality is immediately apparent and undeniable to users, like the first iPhone or ChatGPT.

SaaSpocalypse

The debate or prediction that traditional SaaS tools will become obsolete as users build their own custom tools with AI, though the speaker believes this is overstated due to maintenance and specialization needs.

?
Why should designers and PMs learn to code in the AI era?

Learning to code, even for prototyping, forces designers and PMs to deeply understand the medium and how agent loops work, which is crucial for designing effective AI-powered experiences, rather than just tweaking styles.

?
What is the most important quality for thriving in the AI era?

Agency is the most important quality, as it enables individuals to understand that the world is malleable and they can initiate change, even with AI tools providing skills at their fingertips.

?
What might be lost as traditional roles like engineer, PM, and designer merge?

There's a risk of losing specialists in engineering (optimizing for scale and reliability) and design (craft and delight), as the focus shifts to rapid feature shipping and token spending.

?
What is 'malleable software' and why is it important?

Malleable software is designed to serve the user's interests, allowing them to customize and own their computing environment, rather than being locked into rigid, corporate-designed applications. It empowers users to make tools and adapt software to their needs.

?
Is the 'SaaS apocalypse' a real threat to existing software companies?

The 'SaaS apocalypse' is largely overstated because most people don't want to maintain the full software stack themselves; they value the maintenance and specialized expertise that SaaS providers offer.

?
How has product building fundamentally changed in the last two years?

The first 10% of every project is now 'free,' meaning it takes almost no effort to build an initial version or explore many paths quickly, shifting focus to demos over lengthy documentation.

?
What is the speaker's view on token spend for AI?

Currently, it's worth letting people explore with unlimited token spend, as it's a new technology. However, in 6-12 months, companies will likely start asking uncomfortable questions about the ROI of token spend.

?
Why has Notion AI been so successful?

Notion AI benefits from the platform's connected workspace, which provides rich context for agents to operate, and its operating system-like nature that resembles the Unix environment favored by coding agents.

?
How can product teams ship more quickly while maintaining quality?

By embracing 'incremental correctness' and increasing 'shots on goal' through rapid iteration, while still striving for 'obviously good' products and eventually consolidating disparate ideas into a simple, robust core.

?
What is the most important factor in building a successful product?

The most important factor is the team building the product. Beyond that, successful products typically have one 'tiny core' superpower that is exceptionally good, rather than a broad suite of features.

?
Do we already have Universal Basic Income (UBI)?

The speaker's hot take is that knowledge work itself functions as a form of UBI, as many jobs are not strictly necessary for survival but provide a comfortable living, and humans will always find ways to insert themselves into new loops.

?
What is a contrarian opinion about inclusivity?

Inclusivity isn't always great; sometimes it's beneficial to be exclusive and build a really good product for a specific 'top of the class' in-group, which by definition means excluding others.

1. Cultivate Personal Agency

Realize that the world is made up by people no smarter than you, and that you can change things. Start by tinkering and making things to awaken this sense of agency within yourself.

2. Design and Prototype in Code

Designers and PMs should think and prototype in the actual medium (code) that the product will be built in. This forces a deeper understanding of the material and how AI agent loops work, rather than relying on static design tools.

3. Prioritize Demos Over Memos

Accelerate product development by building janky versions and demos early in a project, rather than writing extensive documentation like PRDs. This allows for immediate reactions and faster iteration, making the initial 10% of a project almost ‘free’.

4. Focus on a Tiny Core Superpower

Build successful products by identifying and perfecting one exceptionally good, tiny core feature or idea. Avoid the pitfall of continuously adding more features, as true greatness comes from a singular, powerful core.

5. Embrace Incremental Correctness

Increase ‘shots on goal’ by iterating frequently and allowing different ideas to grow. Be comfortable with shipping things that aren’t perfectly polished initially, then consolidate and refine them into a robust, simple core.

6. Build ‘Obviously Good’ Products

Aim for products that are undeniably excellent and useful from the first impression, similar to the original iPhone or ChatGPT. This sets a high bar for quality and ensures the product’s value is immediately apparent.

7. Develop Taste Through Repetition

Cultivate good taste by consistently doing the work, getting feedback, and iterating. This process builds an internal ‘virtual machine’ that can predict how a specific in-group will react to an idea.

8. Use Jobs-to-be-Done Holistically

Employ the Jobs-to-be-Done framework to think holistically about what users truly ‘hire’ your product for. This encourages a user-centric perspective, ensuring you’re building what users actually want, not just what you want them to want.

9. Actively Encourage Agent Adoption

Companies need to actively prod people to identify and automate repetitive tasks using AI agents. Significant effort is required to shift individuals from their accustomed manual workflows to leveraging AI for efficiency.

10. Review Code More, Intervene Less

Focus on thorough code review, but view any manual human intervention in the code-writing process (beyond review) as a potential bug. This indicates a need for better automation or refinement in the software factory.

11. Question Need for More Intelligence

For many cognitive tasks, ‘good enough’ intelligence is sufficient. Prioritize optimizing for speed, cost, and local execution of models, rather than always seeking the most advanced frontier models.

12. Surround Yourself with Tasteful Things

Expose yourself to well-designed and tasteful objects and ideas. This environment can inspire improvement in your own work by highlighting areas where your creations might be lacking.

13. Embrace Strategic Exclusivity

Consider building a product for a specific ’top of the class’ in-group. This focused approach can lead to a highly successful and well-crafted product, even if it means intentionally excluding a broader audience.

14. Avoid Silicon Valley Frenzy

Don’t let the intense pressure and worry of Silicon Valley distract you from your genuine passions and what you truly care about in life. Work hard on meaningful endeavors, rather than chasing every trend or fearing missing out.

if you drive Notion like it's stolen, one day you wake up and you realize the world is made up by people no smarter than you. It just really awakens you to the idea that you can just change things.

Max Schoening

The first 10% of every project are now free. It takes almost no effort to now build the first version of a startup.

Max Schoening

All the great products have something tiny that is a superpower. One tiny core that is so exceptionally good.

Max Schoening

One of the biggest pitfalls is if you get into the loop of if I just add one more thing to the product, it will be finally great. That never works.

Max Schoening

We already have universal basic income. It's called knowledge work.

Max Schoening

if I had to describe what taste actually means it's you're able to run a virtual machine in your head where given an idea you can predict for a certain in-group whether they're going to like it or not.

Max Schoening

I don't think that there is a through line out of the things that I've contributed to where I can pinpoint it — I think that you can't say that the best design always wins... I think you can't even say that the way it's built always like the best engineering always wins.

Max Schoening

I think that Silicon Valley is uncharacteristically full of people who don't actually love computers.

Max Schoening

the universe has changed and life is what you make it.

Max Schoening
10%
Percentage of GitHub's top contributors who were designers (pre-LLMs) Indicates designers' historical ability to contribute to production code.
6 to 12 months
Timeframe for companies to start asking about AI ROI Prediction for when AI token spend will face scrutiny.
thousands for sure but like maybe tens of thousands
Estimated monthly token spend for a high-usage individual at Notion Illustrates the current scale of individual AI tool usage.
every two weeks
Frequency of regressions in CLI tools from labs Highlights a current challenge in software quality from AI labs.
8 people or fewer
Ideal size for group chats that 'run the world' Based on small group theory, supporting the idea of exclusivity.
18 to 25
Recommended age range for working incredibly hard Advice for young people in Silicon Valley.
27
Chapter in 'Code' by Charles Petzold where first line of code appears Emphasizes the book's focus on fundamental computer science before coding.
2014
Year Max Schoening started a competitor to Notion Context for a past product failure due to focusing on the wrong core feature.