Lessons from Hoover

In 1936, Congressman John Hollis challenged J. Edgar Hoover during House appropriations hearings: “You’re asking for federal funding to build the FBI, but you’ve never made an arrest yourself.”

Hoover had no good answer. Because it was true.

In his early years as Director, Hoover understood something critical - he did the casework before building the systems. The 1924 photo of him in shirtsleeves, cataloging evidence, wasn’t for show. He worked cases before asking agents to work them.

But somewhere along the way, he stopped doing the work and started only directing it. And when Congressman Hollis called him on it, the credibility gap showed.

Last week I wrote about Troy - about choosing vulnerability over walls when facing AI transformation. Several of you asked - what does that vulnerability really look like?

It looks like making the arrest yourself before you ask for funding to scale.

I haven’t written a line of code in 20 years. But this year, I asked my organization to trust AI to transform how we work. So, I started building, too.

Claude Code for application development. Agents in Cursor – several productivity “assistants”, a newsletter agent, an agent that takes the stress out of travel days. Not because I’m trying to become an engineer again. Because I need to earn the right to ask for their commitment.

And here’s the thing - these tools let you build without writing traditional code. Which means anyone can do this, not just engineers.

When you build something yourself, you learn things you can’t learn from presentations:

• What “just build an agent” really means in practice
• Where AI actually saves time versus where it creates new friction
• Which fears are rational and which are just resistance to change
• How to ask for resources with credibility, and how to teach others along the way

This is what separates leaders who drive transformation from leaders who mandate it. You can’t credibly ask your organization to invest time, energy, and trust in something you’ve only experienced theoretically. That’s the Hoover trap - asking for appropriations without making arrests.

Do the work first. Build something. Break something. Learn why it’s hard.

Then when you ask your team to transform, you’re not asking them to trust your vision. You’re asking them to join something you’ve already proven possible.

Next week: what happened when Hoover finally claimed he made an arrest. And why that matters even more.