World's Fair 2026
Every company should have a Brain — Garry Tan, Y Combinator
Overview
The core thesis is that companies should be built as AI-native entities from inception, leveraging AI not as a mere tool but as a workforce. This approach allows a small team to achieve the output of a much larger traditional organization, fundamentally changing the economics of company building. The key lies in how work is structured and managed, akin to building an organization with AI agents as employees and a curated knowledge base as the company's collective memory.
Who should watch
- Founders and startup leaders looking to build lean, high-output organizations.
- AI engineers and developers focused on agentic systems and productivity.
- Product Managers and builders exploring new organizational paradigms enabled by AI.
- Anyone interested in the practical application of AI for scaling operations and output.
- Individuals concerned about job displacement due to AI, seeking to understand how human roles evolve.
Key takeaways
- Individual AI output can be orders of magnitude higher than human output for certain tasks, with one person using AI potentially doing the work of hundreds.
- The leverage in AI productivity comes not from the AI model itself, but from how work is organized and managed, akin to an organizational structure.
- AI-native companies can be built using "skill files" as employees, "resolver tables" as org charts, and "filing rules" as internal processes, all managed through markdown and code.
- The concept of a "company brain" is crucial, acting as a curated, compounding library of knowledge that agents can access, differentiating an AI colleague from a mere assistant.
- Effective AI integration requires distinguishing between latent space computation (judgment, vagueness) and deterministic space computation (code execution, structured data).
- Working memory limitations in humans (around 7 items) are overcome by AI's capacity (millions of tokens), but effective use depends on context engineering—deciding which information is relevant.
- The practice of "skillifying" work, turning one-off tasks into reusable skills, is essential for compounding organizational intelligence and avoiding repeated effort.
- AI-native companies can achieve unprecedented revenue per employee by encoding functions like sales, support, and operations into AI skills rather than hiring large human teams.
Notable quotes
*The leverage is not in the weights. It's in how you wire the work.*
*When you sit down with Claude Code, Codex, you're not writing software, you're hiring, training, and managing a workforce made of markdown.*
*Model quality is rented, but if you build your brain, you own that brain.*
Unofficial community note. Prefer the recording for nuance.