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

Episode Overview

• The episode explores the profound shift in software creation caused by AI, where raw technical skill is no longer the primary bottleneck, making human "agency" and initiative the most valuable assets.
• It details the evolution of product development from static planning (like PRDs and rigid design mockups) toward rapid, code-based prototyping and interactive demos.
• The discussion demystifies the concept of "taste," framing it not as innate magic, but as a trainable mental model used to predict user reactions in a zero-friction building environment.
• It challenges the myth of the "SaaS apocalypse," explaining why traditional subscription software will survive AI disruption due to the hidden, ongoing costs of maintenance and security.

Key Concepts

• The Shift from Skill to Agency: As AI lowers the barrier to entry for coding, technical execution is commoditized. The true differentiator is "agency"—the internal drive to ignore traditional job boundaries, make decisions, and steward an idea from concept to reality.
• Malleable Software: The future of digital tools is moving away from rigid, corporate-dictated products toward software that users can adapt, mold, and customize to fit their specific personal workflows.
• The "First 10%" is Free: AI has fundamentally altered the starting line of product creation. Because generating the first 0.8 version of an idea takes almost no effort, the competitive advantage has shifted away from getting an idea off the ground toward iteration and distribution.
• The Myth of the SaaS Apocalypse: While individuals can easily build internal tools with AI, the traditional SaaS model won't disappear. Companies pay for SaaS primarily for ongoing maintenance, security, and edge-case management—burdens they do not want to handle themselves.
• Taste as a Predictive Virtual Machine: "Taste" in product design is the ability to run an accurate mental simulation of how a specific target audience will react to an idea. It is a practical, learned skill developed through high-volume iteration and observation.
• The Death of the "Dead Fish" Prototype: Static design tools lack the dynamic, unpredictable nature of AI models. Theorizing in mockups fails to capture nuance, making interactive, coded prototypes essential for modern product design.
• Quality vs. Quantity in Features: Feature count is a vanity metric akin to lines of code. True product excellence relies on a small set of highly polished, combinatorial features rather than a bloated surface area.

Quotes

• At 0:00:11 - "I don't think agency is very evenly distributed in the world." - Highlights that while tools are becoming universally accessible, the internal drive to effect change remains rare.
• At 0:00:23 - "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." - Emphasizes the realization of personal agency in building systems.
• At 0:00:32 - "The first 10% of every project are now free. It takes almost no effort to now build the first version of a startup." - Explains how AI has eliminated the initial friction of getting an idea off the ground.
• At 0:03:36 - "Stop drawing dead fish." - Urges designers to move away from static mockups and build interactive prototypes to truly understand how a product feels.
• At 0:18:20 - "Malleable software is the idea that software works closer to the interest of the people that use it than the interest of the corporation that makes it." - Defines the user-centric future of adaptable software.
• At 0:25:40 - "Software is like a garden you need to tend to it. And the thing you pay for in the as-a-service is the maintenance." - Clarifies why SaaS persists despite AI making initial coding easier.
• At 0:28:48 - "I think the first 10% of every project are now free. So there is no point for most things to, for example, write a PRD... if you can just do the janky version and say, 'Here's the demo of what I think we should build.'" - Advocates for shifting from theoretical planning to rapid prototyping.
• At 0:29:58 - "Demos not memos... Give me something to react to. If you're going to write a PRD, just write the changelog or the blog post that a user would read." - Emphasizes the value of tangible artifacts over abstract documents.
• At 0:34:02 - "At the end of the day, feature count is the same silly metric as lines of code or tokens consumed." - Criticizes measuring productivity by output volume rather than necessity.
• At 0:37:37 - "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..." - Defines "taste" functionally as a predictive mental simulation.
• At 0:42:06 - "Why do automatic code review tools not work that well? ... you publish the work of you plus Claude, and you get bragging rights." - Illustrates the psychology of AI as a collaborator versus an automated critic.
• At 0:57:29 - "Let's just only make obviously good stuff. The origin... what does that mean? And it's like, you know it when you see it." - Establishes a high bar for intuitive product quality before shipping.
• At 1:04:03 - "If you get into the loop of 'if I just add one more thing to the product it'll be finally great'... that never works. Like if I really look at the truly great products, they all have one tiny core that is so exceptionally good." - Warns against feature bloat as a false savior for struggling products.
• At 1:08:19 - "People sort of turn off the brain when they're reviewing their own products... from a 'I'm a user, is this a good experience?' and they're more like 'I'm an employee of this company and I made a thing'." - Explains the necessity of frameworks to maintain objectivity regarding your own work.

Takeaways

• Stop relying on static design tools and start prototyping interactively with code to understand how AI integrations truly feel and behave.
• Replace written Product Requirement Documents (PRDs) with rapid, functional demos ("demos not memos") to gather immediate, tangible reactions.
• Ignore traditional job titles (like "I'm just a PM") and cultivate agency by doing whatever is necessary to ship a working idea end-to-end.
• Train your product "taste" by relentlessly exploring new apps, building side projects, and testing ideas to refine your predictive mental model of user behavior.
• Avoid the "one more feature" trap; instead, ruthlessly refine the single most important core utility of your product until it provides undeniable value.
• Use frameworks like "Jobs to be Done" to evaluate your own work objectively as a user, preventing yourself from falling in love with the mere act of creating.
• Evaluate progress based on the quality and combinatorial power of your features, rather than vanity metrics like total feature count or tokens consumed.
• When evaluating AI tools vs. building your own software, explicitly account for the long-term burdens of maintenance, security, and edge-case management, which represent the true cost of a product.