Key Moments
AI era skills: Why cultivating agency matters more than job titles | Max Schoening (Notion)
Lenny's PodcastKey Moments
AI makes the first 10% of any project free, enabling rapid prototyping, but true success hinges on 'agency' rather than job titles. The real challenge is not skill, but the mindset to leverage tools and change the world around you.
Key Insights
The first 10% of any project is now effectively free due to AI, drastically reducing the effort needed to build initial versions of products or startups.
Agency, the belief that one can change things and influence their environment, is the most crucial differentiator for success in the AI era, even more than specific skills.
Productivity in software development has increased, but software quality has not necessarily followed suit, highlighting a need for better engineering and craftsmanship amidst rapid iteration.
Malleable software, which allows users to customize and adapt it to their needs, is crucial for increasing user ownership and counteracting the rigidities of current software design.
Developing 'taste' in product development involves building a virtual machine in one's head to predict user reception, requiring iterative 'reps' akin to training an AI model.
Great products are defined by a single, exceptionally good 'tiny core' superpower, rather than an accumulation of many features, a principle seen in products like the iPhone's multi-touch or GitHub's pull request.
AI liberates early-stage product development, but agency is the true superpower.
The advent of AI has fundamentally changed how products are built. Max Schoening highlights that the initial 10% of any project – the ideation, basic setup, and early prototyping – is now essentially free. This drastically lowers the barrier to entry for creating the first version of a startup or a new feature, making it almost effortless to explore initial concepts. However, he posits that possessing the necessary skills is no longer the primary differentiator. Instead, 'agency' – the belief and drive to effect change and shape one's environment – is paramount. This internal locus of control, the conviction that one is not bound by predefined roles or limitations, is what truly separates those who thrive in the current landscape from those who may fall behind. This agency, he suggests, is not evenly distributed and is the key to navigating the increasingly malleable world of software.
Rethinking roles: Designers and PMs coding to master the material.
Schoening advocates for designers and product managers engaging with code, not necessarily to directly ship production code, but to deeply understand the medium they are designing with. He uses the analogy of a "playground" — a separate, LLM-friendly environment optimized for rapid iteration — to help teams overcome the initial fear of coding. This hands-on experience with code allows them to grasp concepts like 'agent loops' and design more effectively for AI-driven interactions. While some designers and PMs do contribute to production code, the primary benefit, he argues, is gaining a mastery of the material. This contrasts with the traditional approach where designers might only manipulate static mockups in tools like Figma, missing the nuanced understanding of how AI interacts with functional code. The goal is not just to 'vibe code' faster, but to design in the actual medium that the final product will inhabit, leading to a more profound understanding of the system's capabilities and limitations.
Malleable software: Empowering users and reclaiming ownership.
The concept of 'malleable software' is central to Schoening's vision for the future. He defines it as software that serves the interests of its users more than the corporation that builds it, pushing back against the rigidity of traditional app structures where customization is often impossible. He likens current software environments to being unable to rearrange one's living room, a level of restriction we would never accept in physical spaces. The ideal, he suggests, is software that feels like an operating system for one's life, offering ownership and adaptability. While he notes that tools like Notion are moving in this direction, he believes they can and should become even more malleable. This vision is supported by the idea that software should be like a garden, requiring continuous tending and adaptation, rather than a static, immutably designed product. The 'as a service' aspect remains crucial, as most users do not want to maintain the entire technology stack, but they do desire more control over their digital environment. This leads to a future where general-purpose tools, possibly delivered as a service, will offer significant malleability, challenging the 'SAS apocalypse' narrative by evolving rather than disappearing.
Taste: The uniquely human skill in the age of AI.
In a world where AI can generate content and code at an unprecedented scale, 'taste' emerges as a critical human differentiator. Schoening describes taste as the ability to run a 'virtual machine in your head' that predicts how a specific group of users ('in-group') will react to an idea or product. Developing taste is not an innate trait but a skill honed through extensive practice and feedback – akin to training an AI model. It requires an iterative process of inputting ideas, observing reactions, and refining one's judgment. He draws parallels to Japanese craftspeople who perfect their skills over lifetimes. For designers, he suggests, this involves having side projects where they are responsible for the entire product lifecycle and constantly exploring new tools. Exposure to tasteful examples, whether through design or even well-crafted physical objects, helps individuals recognize what is 'obviously good.' This distinguishes them from mere executors, enabling them to define quality and direction in a landscape increasingly populated by AI-generated outputs.
The 'tiny core' superpower and the pitfall of feature creep.
Schoening emphasizes that truly great products are not built by simply accumulating features, but by possessing a single, exceptionally good 'tiny core' that acts as a superpower. This core element is what draws users in and makes them return, even if other aspects of the product are less refined. He cites examples like the multi-touch interface on the first iPhone, GitHub's pull request system, or Notion's block-based editing and slash commands as embodiments of this principle. The pitfall to avoid is the belief that adding 'just one more thing' will make a product great; this incremental approach often dilutes the core value. Instead, the focus should be on perfecting this central superpower. This can arise from a combination of luck and market agreement, but it requires relentless iteration on that core idea until it truly resonates with users. The lesson is to identify and perfect this singular strength, rather than chasing a diffuse set of features.
Agency as the key to navigating the future and personal fulfillment.
Schoening returns to the concept of agency when advising individuals on how to develop it within themselves. He reiterates the powerful realization, often attributed to Steve Jobs, that the world is made up of people no smarter than oneself. This awakening is crucial for fostering a sense of agency. He stresses that agency is cultivated through 'making' – tinkering, building, and creating things. This act of production, whether it's a physical object or a digital creation, naturally leads to learning and improvement. The alternative is to feel like a passive component in a large machine, waiting for instructions. For those seeking to develop agency, the advice is simple: start by making things. This process naturally awakens the understanding that change is possible and that one has the power to effect it. This internal drive is ultimately more valuable than having specific skills or adhering strictly to job titles, allowing individuals to shape their careers and their impact on the world.
The 'SAS apocalypse' is exaggerated; tools will become more general.
While some predict the demise of Software as a Service (SaaS) with the rise of AI-powered custom tool building, Schoening believes the 'SAS apocalypse' is largely exaggerated. He argues that the value of SaaS lies not just in the functionality but in the 'as a service' aspect: the maintenance, ongoing development, and specialized expertise that end-users are typically unwilling or unable to manage themselves. He likens it to preferring to buy a steak from Costco rather than hunting and butchering it oneself. Instead of disappearing, SaaS tools will likely become more general-purpose – think of advanced word processors or spreadsheets – but will still be offered as a service. This means users can leverage their AI capabilities to customize and adapt these general tools without taking on the full burden of infrastructure and maintenance. The ongoing value proposition for SaaS providers will be in their ability to maintain these complex systems and continuously improve them with specialized teams, ensuring that users can focus on their core tasks rather than managing their toolchains.
Universal Basic Income is here: It's called knowledge work.
Schoening offers a provocative 'hot take' on Universal Basic Income (UBI): he suggests that knowledge work, in essence, already functions as a form of UBI. He frames this with a degree of humor and a touch of seriousness, noting that humans have generally built complex hierarchies and jobs that often provide sustenance and a baseline of contentment beyond basic survival needs. While acknowledging that not everyone has this privilege or security, he believes those who are discussing this topic are likely already beneficiaries of this 'knowledge work UBI.' The implication is that the nature of work and our perceived needs are complex societal constructs. As AI automates more tasks, the conversation might shift from 'if we will need to work' to how humans will continue to find purpose and insert themselves into evolving systems, leveraging their inventiveness to create new roles and values regardless of automated capabilities. He suggests that many in the tech industry are lucky to be in this fortunate position and should appreciate it.
Mentioned in This Episode
●Products
●Software & Apps
●Companies
●Organizations
●Books
●People Referenced
Common Questions
Agency refers to the understanding that the world is malleable and can be changed by you. Even with AI providing skills, agency is what truly matters, enabling individuals to shape their environment and work effectively.
Topics
Mentioned in this video
Mentioned as a company that is powered by Work OS.
Mentioned as a company associated with Vanta.
Mentioned as a company powered by Work OS and as one of the major AI labs.
Mentioned as a company powered by Work OS and as one of the major AI labs.
Mentioned as a company that is powered by Work OS.
Design tool used for prototyping chat interfaces, but its role is shifting as designers move towards prototyping in code.
Max Schoening's previous employer where he led design and product, and notes that designers were significant contributors to the GitHub codebase.
Mentioned as an example of a SaaS tool that some might consider building their own alternative for, but the speaker argues against this due to maintenance and specialization.
Mentioned for facing criticism for attempting to create a leaderboard for AI usage, which the speaker understands as a way to encourage adoption at scale.
A company that automates compliance and risk management for businesses, helping them earn trust.
Mentioned as a company associated with Vanta.
Mentioned as a 'fanboy' example of companies designing software from an 'ivory tower,' limiting user customization.
Mentioned as a company associated with Vanta.
Mentioned as a company associated with Vanta.
Venture capital firm that was going to provide a term sheet for a Notion competitor, but rescinded due to Notion's pivot.
Mentioned as a company that is powered by Work OS.
Mentioned as a company that is powered by Work OS.
Max Schoening's current company, a workspace software that aims to be malleable and act as an operating system. It has recently integrated AI features.
Mentioned as a company associated with Vanta and as an example of running models locally.
A modern developer platform for B2B SaaS companies, providing APIs for enterprise features like SSO and SCIM.
Mentioned as an example of the flip side to app-based software, offering malleability but requiring more effort to manage.
An Anthropic model variant mentioned in the context of fast inference costs.
An Anthropic model variant mentioned in the context of fast inference costs.
Mentioned as the first Notion assistant being launched before it, and as an example of an 'obviously good' product.
Mentioned in the context of the desktop app having multiple tabs for co-work, and as an example of software quality issues.
Mentioned as a tool that now has features like markdown folding, which Max Schoening's team worked on in 2014 for a competitor to Notion.
Mentioned as an example of a critical SaaS tool that companies like Anthropic rely on, and that nobody wants to rebuild due to its complexity and scale.
A terminal emulator recommended for its quality, contrasting with generally 'terrible terminals' most people use.
A phone application that Max Schoening is currently exploring and finds to be well-done, with a focus on coding on the phone.
Mentioned as part of Max Schoening's history at GitHub.
Mentioned for his talk 'Stop Drawing Dead Fish,' which influences the thinking about designing chat interfaces dynamically rather than statically.
Mentioned as an example of someone at Notion with high agency, blurring engineering and design, and being a key recruiter.
An example of someone at Notion demonstrating high agency by shifting from strategy docs to prototyping in Figma and then to building prototypes directly.
Mentioned in the context of Dieter Rams criticizing designer chairs which are beautiful but not necessarily functional, similar to some museum pieces.
Mentioned for his criticism of designer chairs, highlighting the importance of usefulness and malleability in design.
A journalist who tweeted about understanding and using Notion thanks to Notion AI.
Mentioned as 'Andre' in the phrase 'software we uh Andre right like software is eating the world.'
Author of 'Code: The Hidden Language of Computer Hardware and Software,' recommended for understanding how computers work.
Author of 'Tools for Conviviality,' which contrasts tools that enhance human autonomy with those that are destructive to it.
His quote 'Life is what you make it' is a life motto Max Schoening returns to, emphasizing agency and living in the moment.
Mentioned for his idea 'How Buildings Learn,' which connects to the concept of malleable software adapting over time.
Mentioned for the idea that you can change things around you because they are made by people no smarter than yourself.
Mentioned in relation to the idea that society is not capped by intelligence.
Mentioned for his ideas on manual intervention in code within a software factory.
A book, popularized on Stack Overflow, about creating legible systems that neglect the reality of how they function on the ground, relevant for executives.
A TV show that, when reframed with AI replacing God, becomes a heavy but impactful watch.
A book that contrasts technology that enhances human autonomy with tools that diminish it.
A book and its film adaptation that Max Schoening enjoyed, sparking excitement about the future of humanity and space exploration.
A book recommended for understanding the fundamental workings of computers, even for programmers.
An open-source split keyboard that Max Schoening likes for its ability to reclaim agency in one's computing life, allowing for custom builds.
A physical tool that Max Schoening appreciates for its quality and usefulness, even if it's a considered a higher-end pocket knife.
Cited as an example of a product that was 'obviously good' upon release.
More from Lenny's Podcast
View all 28 summaries
71 minSnapchat CEO: Why distribution has become the most important moat | Evan Spiegel
86 minHow Anthropic’s product team moves faster than anyone else | Cat Wu (Head of Product, Claude Code)
96 minWhy half of product managers are in trouble | Nikhyl Singhal (Meta, Google)
83 minHard truths about building in the AI era | Keith Rabois (Khosla Ventures)
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free