Agency Trumps Skill: In the AI era, the ability to self-direct and execute projects ('agency') is more valuable than specific technical skills, which are being democratized by AI. (Claim 35)
The 'Tiny Core' Product Strategy: Truly great products are built around a single, exceptionally well-executed core feature, rather than a broad set of mediocre ones. (Claim 54)
AI Commoditization to 'Good Enough': The focus of AI product development will shift from raw intelligence to factors like speed, cost, and local execution once models become 'good enough' for most tasks, challenging the assumption that customers always want the frontier model. (Claims 31, 21)
The Blurring of Technical and Non-Technical Roles: AI tools are enabling non-engineers (designers, PMs) to contribute directly to code, while the principles of software engineering will permeate all business domains as software creation costs approach zero. (Claims 10, 22, 39)
Strategic Positioning Over User Experience: A product with a superior user experience (e.g., Heroku) can lose in the market to a competitor (e.g., Kubernetes) that offers a better strategic value proposition to the organization, such as empowering existing teams and avoiding vendor lock-in. (Claims 27, 46)
2014
Schoening's startup had a term sheet from True Ventures withdrawn after Notion, a future competitor, created a portfolio conflict for the VC firm. (Claim 49)
Pre-LLM Era (at GitHub)
Worked as a design leader and engineer at GitHub under Nat Friedman. He notes that during this time, the culture emphasized 'demos, not memos' and that designers accounted for roughly 10% of top contributors to the GitHub codebase. (Claims 1, 4, 6, 20, 60)
Career History
Held roles as a product manager at Google and ran the design team at Heroku, where he developed insights into product strategy and user experience, particularly through the lens of Heroku's competition with Kubernetes. (Claims 2, 3, 5, 27, 46)
Current (at Notion)
As a leader at Notion, he is overseeing a shift where designers and PMs use AI to code, prototyping for AI has moved from Figma to custom playgrounds, and the company maintains an 'unlimited' internal AI token spend policy to encourage exploration. (Claims 14, 19, 32, 41)
▶The 'Tiny Core' Product PhilosophyMay 2026
Schoening posits that all truly successful products are built around a single, exceptionally well-executed feature or capability. He identifies this 'tiny core' for companies like GitHub (the pull request), Heroku (git push deployment), Notion (block editor), and Figma (real-time collaboration).
This philosophy suggests that investors should prioritize companies that demonstrate obsessive focus and excellence on a single, powerful user value proposition over those with a wide but shallow feature set.
▶Agency Over Skills in the AI Era
A central thesis of Schoening's is that as AI democratizes technical skills, the most valuable human quality becomes 'agency'—the drive and ability to take initiative and see a project through. He believes this trait is more critical for success than traditional skills, which can now be augmented or automated by AI.
For talent acquisition and management, this implies a shift in focus from assessing specific technical proficiencies to identifying and cultivating proactive, self-directed individuals who can effectively leverage AI as a tool.
▶The Blurring of Roles and Rise of the GeneralistMay 2026
Schoening observes and encourages a trend where designers and product managers are increasingly using AI tools to contribute directly to production code, breaking down traditional silos. He predicts that the principles of software engineering will permeate all business domains as the cost of creation plummets, leading to more generalized, technically-enabled roles.
This trend could disrupt the market for specialized SaaS tools and create opportunities for platforms that empower generalists to build and manage complex workflows without deep specialization.
▶The Future of Software: Malleable, General-Purpose ServicesMay 2026
Schoening predicts a market shift away from the rigid, form-based SaaS of the 2010s and back toward more general-purpose, malleable tools reminiscent of 1990s applications like FileMaker Pro. However, he maintains that these tools will continue to be delivered 'as a service,' as users will not want the burden of maintaining the underlying infrastructure.
This signals a potential decline for niche, single-purpose SaaS applications and a major opportunity for highly customizable, AI-native platforms that provide powerful, general-purpose primitives for users to build their own solutions.