Description

In this first episode of 2026, Nicklas Berild Lundblad and Richard Allen explore how AI tools are transforming the craft of public policy. Their core argument: if you're regulating technology, you need to understand it firsthand—and the barriers to experimenting have never been lower. 

Get comfortable with the command line. Every operating system has a terminal (Command+Space then "Terminal" on Mac, search "Terminal" on Windows). Pair it with your favorite chatbot for a 30-minute crash course. This direct communication with your computer is foundational for everything else.

Install an IDE. VS Code or PyCharm (both free) provide clean environments to run code. The goal isn't becoming a programmer—it's achieving code readability so you can understand and guide AI-generated code.

Learn Python basics. Twenty hours of investment unlocks a career-long skill set. Focus on reading code, not writing it—LLMs handle the heavy lifting while you become the quality-control human in the loop.

Run local models with Ollama. This free tool lets you download and run LLMs entirely on your machine—no API costs, no data leaving your computer. Start with small models like Tiny Dolphin, then explore Llama or Gemma. Running the same prompt hundreds of times reveals how probabilistic these systems really are.

Explore GitHub. A vast repository of ready-made code. Search for whatever interests you and load it into your IDE.

• Use chatbots as tutors. Stuck? Ask. The days of getting blocked by technical hurdles are over.
• Be blunt. Tell the chatbot directly when it's wrong. Start fresh chats to escape context traps.
• Start with a problem statement. Before coding, articulate exactly what you need. This forces clarity your team probably lacks.
• Build, don't buy. The economics have shifted. Custom internal tools—like parliamentary monitoring systems—are now cheaper and more tailored than off-the-shelf products.
• Research APIs first. Run a deep research query on any API before coding against it to get current documentation.

The bottom line: spend a few days learning these tools and you'll understand AI technology at a level that transforms how you think about regulating it.