The AI paradox: More automation, more humans, more work | Dan Shipper

1. Adopt AI as Your OS

Integrate AI environments like Codex or Cloud Cowork as your primary operating system for daily work, including email, documents, and research, to leverage their full capabilities.

2. Ride the AI Models

Actively use new AI models and tools in your work to extend your capabilities and stay relevant, as this is the primary way to adapt to the changing job landscape.

3. Cultivate AI Curiosity, Play

Regularly experiment with new AI models and tools, even for tasks they couldn’t do before, to discover novel applications and continuously extend your capabilities.

4. PMs: Enhance Technical Skills

Product Managers should deepen their technical understanding and embrace AI tools to directly build and ship products, leveraging their strong product sense to drive innovation.

5. Designers: Learn Full-Stack

Designers should acquire full-stack development skills and use AI tools to directly build their designs, enabling them to create unique interactions and ship products independently.

6. Maintain AI Agents Actively

Ensure a human is consistently monitoring and caring for AI agents, as they require ongoing attention and maintenance to remain useful and effective.

7. Prioritize Creativity and Novelty

Focus on using AI to create new and interesting things that stand out, as AI commoditizes existing competence and leads to a lot of generic output.

8. Design SaaS for AI

Build SaaS products that facilitate seamless collaboration between humans and AI agents, allowing users to bring their own AI tokens and simplifying product features by offloading tasks to agents.

9. Implement Company Super-Agent

Establish a single, general-purpose AI agent for your entire company, managed by a ‘forward-deployed engineer,’ to centralize delegated tasks and ensure consistent functionality.

10. Leverage Multi-Agent Interactions

Utilize multiple AI agents that can communicate with each other, as this allows for richer context exchange and more efficient task completion than direct human-to-agent interaction.

11. Prepare for Output Review

Anticipate a significant increase in output (e.g., pull requests, data analysis) from non-technical roles using AI, requiring more effort from technical staff to review, integrate, and maintain quality.

12. Expect Human Oversight in Automation

Recognize that increased automation with AI will likely lead to more human work in overseeing and refining automated processes, rather than a reduction in overall human effort.

13. Embrace AI-Assisted Writing

Utilize AI for generating documents and emails, especially for routine or factual content, but always ensure you understand and stand behind the AI’s output to avoid ‘slop.’

14. Move Beyond CLIs

Shift away from Command Line Interfaces (CLIs) as your primary work surface, especially for non-programmer tasks, as Graphical User Interfaces (GUIs) within AI environments offer a more effective and user-friendly experience.

15. Consider Forward-Deployed Engineer

Explore or train for the role of a ‘forward-deployed engineer’ who manages and optimizes internal AI agents, ensuring they effectively serve the entire company.

16. Solve Problems with AI

Actively identify problems in your personal or professional life and explore how AI tools can provide solutions, fostering a deeper understanding and appreciation for AI’s utility.

17. Shift SaaS AI Costs

For SaaS companies, consider a model where users bring their own AI tokens when interacting with your product via an agent, potentially improving your margins by reducing your AI infrastructure costs.

18. Invest in SaaS Stocks

Consider investing in SaaS stocks, as AI agents are predicted to increase the number of SaaS users, improving company margins and demand.