World's Fair 2025
Dispatch from the Future: building an AI-native Company – Dan Shipper, Every, AI & I
Overview
This talk explores the emerging playbook for building AI-native companies, emphasizing that the process is currently being invented collaboratively. The core thesis is that achieving 100% AI adoption among engineers creates a significant, tenfold difference in organizational capability. This shift enables a single engineer to build and maintain complex production products, fundamentally altering how software development is approached.
Who should watch
- AI Engineers
- Product Managers
- Builders experimenting with AI in software development
- Those interested in organizational productivity gains through AI
- Teams considering full AI adoption for engineering workflows
Key takeaways
- A 10x difference in organizational capability is observed when 100% of engineers adopt AI tools, compared to partial adoption.
- A small team of 15 people at Every.to successfully runs four software products, demonstrating the potential for lean, AI-native operations.
- Compounding engineering is proposed as a new paradigm where each feature built makes subsequent features easier to develop, creating a positive feedback loop.
- The compounding engineering loop involves four steps: plan, delegate, assess, and codify, with codifying learned knowledge back into prompts being a crucial step.
- AI adoption facilitates easier tacit code sharing, allowing developers to learn from and adapt existing solutions across different projects and tech stacks.
- New hires can become productive on their first day due to AI agents setting up environments and providing guidance on best practices.
- Developers can easily contribute to other products within the organization by leveraging AI to understand and fix bugs or implement minor improvements.
- Organizations can avoid standardizing on a single tech stack, as AI simplifies translation and adaptation between different languages and frameworks.
Notable quotes
*The playbook is actually being invented right now.*
*There is definitely a huge there's a 10x difference between an org where 90% of the engineers are using AI versus an org where 100% of the engineers are using AI.*
*In traditional engineering, each feature makes the next feature harder to build. In compounding engineering, your goal is to make sure that each feature makes the next feature easier to build.*
Unofficial community note. Prefer the recording for nuance.