Looking back, my approach to interview preparation has truly evolved. A few weeks ago, I had the privilege of interviewing Jakob Tanner, a leading Swiss historian, about Switzerland during WWII.
Getting the chance to sit down with someone like Jakob and ask questions about a topic I’m genuinely curious about is why I happily spend my Sundays updating the website and editing audio. For those unfamiliar, Jakob is widely considered the top expert on Switzerland’s role in WWII. He played a pivotal role in advising the Swiss government, reviewing official documents, and helping determine how Switzerland’s wartime wealth, including stored gold and thousands of Jewish-owned bank accounts, should be handled.
After 27 episodes, I’ve really come to appreciate the value of proper interview prep. How much prep you do is an interesting question in itself. Larry King, famously did none, so he could approach each conversation with the same knowledge as the listener. Others ask the same questions every time, which lets them compare how different guests respond. Then there are people I truly admire, like Tim Ferriss and Lex Fridman, who do deep research to find the things that they are personally interested in.
I could simply ask ChatGPT to give me “10 questions about how WW2 affected Switzerland” and get a decent list. But if those questions don’t align with my actual curiosity, what’s the point? Conversely, researching too much can lead to already knowing the answers to questions I ask, turning the conversation into a rehearsed exchange rather than a curious exploration.
Now, I should say, that I do use ChatGPT/Gemini a lot to assist with preparation. Through some trial and error, I’ve found a way for AI to be effective while allowing me to maintain my curiosity and engage in authentic conversations. I view AI as a sparring partner: it helps me train, but I still have to step into the ring myself. For this dynamic to work, I need to know enough about a guest’s work or the topic to discover the specific angles that really pique my interest.
In Jakob’s case, I found that he and other historians had published the Independent Commission of Experts Switzerland – Second World War, a 600-page report on Switzerland’s role during the war… Jackpot.
I don’t have the kind of brain that can digest 600 pages of Swiss history in a few weeks. But I do have everything I need to prepare well. And here’s where Gemini becomes super useful.
I spend 10–15 minutes thinking about the specific questions I want answers to. Things like:
- What are the details around the Germans famously storing gold and other assets in Swiss banks?
- What was daily life like for Swiss people while the rest of Europe was at war?
- What was the immigration situation? I assume a flood of people tried wanted to get in to escape the war.
I feed those questions, and the report, into Gemini (I find Gemini consumes large documents better that ChatGPT) and ask it to pull out interesting insights. That gives me a high-level understanding and helps me decide whether to dig deeper.
For example, for the question related to Switzerland storing German assets, Gemini pulled this out for me:
“Investigations by the Swiss Clearing Office in 1945/46 found over one billion CHF in German assets in Switzerland, two-thirds of which had arrived after the war began. The Washington Agreement of 1946 led to a political agreement with the Allies regarding these assets ,however all further investigation halted due to the onset of the Cold War.”
I read that and I think, damn thats interesting, I have a bunch of follow up questions I am excited to learn about – These are what I use for the interview.
I also ask it to surface things I hadn’t even considered. “Give me the top 10 most interesting facts you found in this document” is a great prompt, and honestly, a few of those make it into my notes for the episode to discuss in more detail.
This is one of my favourite parts of the podcast process. Asking good questions is one way to scratch the curiosity itch. But before that, it’s about learning enough to decide which questions are worth asking.
Thanks for reading,
Mike
