Visual summary of operating lessons from Max Schoening.

Lessons from Max Schoening

Max Schoening leads product at Notion and previously worked in design and engineering at GitHub. In his interview with Lenny Rachitsky, he discusses how AI changes prototyping and why a great product often has one exceptionally good interaction. — Max Schoening on Lenny’s Podcast.

Part 1: Agency and the AI Era

  1. Act when new tools make a problem tractable: Schoening argues that AI makes some skills easier to access, increasing the value of taking initiative and trying to change how something works. — Max Schoening on Lenny’s Podcast.
  2. Take on useful work beyond your formal role: Schoening describes a Notion colleague who saw a hiring need and went out to meet candidates even though recruiting sat outside his usual design and engineering work. — Max Schoening on Lenny’s Podcast.
  3. Explore more directions when first drafts get cheaper: Schoening says AI makes it much cheaper to build rough versions and explore alternatives. He still sees the work of finishing a reliable product as difficult. — Max Schoening on Lenny’s Podcast.
  4. Develop taste through repeated feedback: Schoening describes taste as the ability to anticipate how a particular audience will respond to an idea. He says builders develop that judgment by making things and seeing how people react. — Max Schoening on Lenny’s Podcast.
  5. Keep specialists as roles overlap: Schoening warns that broader prototyping skills could come at the expense of specialists. He still sees a need for engineers who can make software reliable at scale and designers who can bring craft to the experience. — Max Schoening on Lenny’s Podcast.
  6. Create a playground for people learning to prototype in code: Schoening says Notion set up a separate environment that reproduced familiar interface patterns so designers and product managers could try building with AI without first learning the main codebase. — Max Schoening on Lenny’s Podcast.
  7. Hire for the skills the work actually calls for: Schoening recalls relaxing a requirement that designers be able to code because it was hard to hire for. In hindsight, he would have preferred a smaller team that retained that mix of skills. — Max Schoening on Lenny’s Podcast.

Part 2: The "Tiny Core" Theory

  1. Find the small interaction that makes the product special: Schoening looks for one exceptionally good interaction at the center of a product. He sees the market’s response as part of discovering which interaction matters. — Max Schoening on Lenny’s Podcast.
  2. Use simple interactions that can support different workflows: Schoening points to Notion’s blocks and slash commands as examples of the small interactions at the center of a flexible product. — Max Schoening on Lenny’s Podcast.
  3. Make it easy to propose a change: Schoening singles out the pull request as GitHub’s central interaction because it lets someone suggest a change and lets others review it. — Max Schoening on Lenny’s Podcast.
  4. Iterate on the interaction until it clicks for users: Schoening says an idea can feel promising and still fall flat in user testing. He advises refining the interaction until the intended users can make it work. — Max Schoening on Lenny’s Podcast.
  5. Notice when a new interaction changes the experience: Schoening offers multitouch on the iPhone as an example of the small interaction that can give a product its distinctive appeal. — Max Schoening on Lenny’s Podcast.
  6. Question whether one more feature will make the product work: Schoening warns against repeatedly adding features in the hope that the next one will make the product compelling. — Max Schoening on Lenny’s Podcast.
  7. Make the core job feel dependable: Schoening remembers early Dropbox for reliable file syncing and a menu bar icon that showed what was happening. That small experience was what he wanted the product to do well. — Max Schoening on Lenny’s Podcast.

Part 3: Prototyping Over Planning

  1. Write the changelog when a document is useful: Schoening recalls that GitHub product reviews sometimes asked for a user-facing changelog or launch post as a way to make a proposal concrete. — Max Schoening on Lenny’s Podcast.
  2. Prototype AI behavior in the medium where it runs: Schoening wants designers and product managers to work with code so they can understand interactions such as agent loops. A prototype can teach that even if engineers replace its code later. — Max Schoening on Lenny’s Podcast.
  3. Use the lightest prototype that can get useful feedback: At GitHub, Schoening sometimes started with a sketch shared with a few users. When the question needed more detail, the team brought users into a screensharing session to discuss the design. — Max Schoening on How GitHub Iterates Fast.
  4. Show a demo that people can react to: Schoening says GitHub product reviews asked teams to show what a user could do. A concrete demo gave the team something to respond to and improve. — Max Schoening on How GitHub Iterates Fast.
  5. Let designers work on the interface they design: Schoening said GitHub’s product designers worked with product managers during discovery and wrote and published interface code, mainly HTML and CSS. — Max Schoening on How GitHub Iterates Fast.
  6. Prototype the data model as well as the interface: When Schoening builds to think, he includes the data model. He looks for an underlying structure that matches how users understand the work. — Notion’s Max Schoening Interview.

Part 4: Malleable Software and Ink & Switch

  1. Let users rearrange their software: Schoening compares adaptable software to a home whose furniture and drawers people can rearrange to fit their routines. — Notion’s Max Schoening Interview.
  2. Turn text notes into personal tools: In the Potluck prototype, a live search extracts structured information from freeform text. Users can write formulas against it and display the results as annotations in the same note. — Potluck: Dynamic Documents as Personal Software.
  3. Make version control usable for creative work: Schoening and the Patchwork team explored version control that writers and other creators could use across their files. They saw a gap between powerful developer tools and the narrower versioning features built into individual apps. — Patchwork: Version Control for Everything.
  4. Build from parts people can recombine: Notion co-founder Akshay Kothari describes Schoening as a LEGO maximalist. Schoening explains the comparison through his interest in software users can adapt from simple pieces. — Notion’s Max Schoening Interview.
  5. Let users turn documents into tools a step at a time: The Potluck researchers propose starting with ordinary text and adding computation as the need arises. This gives users a way to move gradually from using a document to shaping how it behaves. — Potluck: Dynamic Documents as Personal Software.
  6. Let different tools work on the same data: Ink & Switch argues that sharing data between tools lets users build workflows that span applications. A photo stored in a shared filesystem, for example, can be edited and annotated with different tools. — Ink & Switch: Malleable Software.

Part 5: Blurring the Lines of PM, Design, and Engineering

  1. Build fluency across product, design and engineering: Schoening says teams benefit when people in different roles share a vocabulary and can work with some of the same materials. AI makes it easier for a designer or product manager to build a prototype to convey an idea. — Notion’s Max Schoening Interview.

Part 6: The Tinkerer (Tüftler) Mindset

  1. Keep tinkering until the idea works: Schoening uses the German word Tüftler for his habit of repeatedly trying and refining ideas until they work. — Notion’s Max Schoening Interview.
  2. Ask how things were made and why they work that way: Schoening traces his impulse to tinker to curiosity about how other people’s creations work and why their makers chose that approach. — Notion’s Max Schoening Interview.
  3. Make room for projects you build for fun: Notion co-founder Akshay Kothari cites CloudApp, a menu bar app Schoening built for fun, as one example of Schoening’s habit of making tools outside a formal product brief. — Notion’s Max Schoening Interview.
  4. Express an idea in the medium users will experience: Schoening trained in product design and tries to develop ideas in whichever medium is closest to the experience users will have. — Notion’s Max Schoening Interview.
  5. Use failed ideas to decide what to test next: Schoening describes incremental correctness as learning quickly through successive versions. At GitHub, teams aimed to use each round of feedback to inform the next version. — Max Schoening on How GitHub Iterates Fast.
  6. Treat the tools around you as things you can change: Schoening points to a remark by Steve Jobs as inspiration for his belief that people can tinker with and change the things around them. — Notion’s Max Schoening Interview.
  7. Help teams make their own decisions: Schoening describes his role as connecting teams that need to talk and giving them enough context to make more decisions independently. — Max Schoening on How GitHub Iterates Fast.

Part 7: Lessons from GitHub, Heroku, and CloudApp

  1. Design automation around reusable pieces: Schoening says GitHub Actions drew on GitHub’s practice of building on others’ work. The team designed for reusable automation components that developers could combine into their own workflows. — Metamuse: Principled Products.
  2. Turn deployment into a small, memorable interaction: Schoening recalls that pushing code to Heroku with Git could turn a local app into a live URL. He uses that simple transition as an example of a product’s central interaction. — Max Schoening on Lenny’s Podcast.
  3. Start with a focused utility: Schoening built CloudApp as a simple file-sharing tool. The direct interview describes how that narrow product grew on Heroku and led to his move into product design there. — Max Schoening on How GitHub Iterates Fast.
  4. Shorten the loop for suggesting a code change: Schoening recalls that GitHub users had to open another pull request to suggest a change. Customer interviews helped the team identify an opportunity to let reviewers suggest edits inline. — Max Schoening on How GitHub Iterates Fast.
  5. Check that you are polishing the part users value: Schoening recalls spending too long refining a document editor and adding features to it. His retrospective emphasizes checking whether the core experience meets a real user need before investing more time. — Max Schoening on Lenny’s Podcast.

Part 8: The Future of Software Creation

  1. Judge generated software by reliability as well as volume: Schoening says AI helped people produce more software, while dependable software remained difficult to find. — Max Schoening on Lenny’s Podcast.
  2. Give agents the context they need to work: Schoening attributes some of Notion’s agent usefulness to its connected workspace. He says agents benefit when they can work across relevant context instead of going through narrow interfaces to isolated repositories. — Max Schoening on Lenny’s Podcast.
  3. Explore ambitious directions before committing to one: Schoening describes sending agents down several paths to test ideas cheaply, giving teams room to try alternatives before committing the effort needed to finish a product. — Max Schoening on Lenny’s Podcast.
  4. Add computation to loosely structured notes: Schoening sees AI as a way to let a document describe information informally while still making some of that information available for computation. — Notion’s Max Schoening Interview.
  5. Keep building things you genuinely care about: Schoening advises younger builders to keep working on things they care about and to give less weight to the fear of missing the latest technology wave. — Max Schoening on Lenny’s Podcast.