I spent this week in DC.
My hotel, a 15-minute walk from the Capitol, was surrounded by hip coffee shops and Ethiopian diners (the diaspora is strong, here!). The heat was brutal; mornings would see a huddled mass of suited-businessmen and lobbyists, all waiting in the air-conditioned lobby for their Ubers to arrive.
The Federation of American Scientists had brought me out for a panel on Thursday, as part of their Day One sprint on AI policy. I decided to stay for the week, and set up a bevy of meetings with folks working within and alongside government.
I was supposedly there to talk about my memo on data sharing standards for healthcare (so important! Another blog on this coming soon!!). But, selfishly, my ulterior motive was to understand how an idea becomes a law, and how external experts should participate.
Here’s my takeaways.
Cold Outreach Works
Perhaps I’m used to Silicon Valley, and its brutal sales pipelines, where 10,000 emails might be met with crickets. Instead, about half of my meetings this week came in cold; it turns out the “contact me” form on Members’ websites actually gets read!
Rep. DelBene (D-WA) was nice enough to sit down with me to talk about nerdy issues in AI + healthcare, as were folks from the offices of Sen. Wyden (D-OR), Sen. Brown (D-OH), and others. This was true for folks off the Hill as well - I sat down with think-tankers from Third Way, the Bipartisan Policy Center, Alliance for Healthcare, & more.
Our meetings felt substantive - we used the memo as a starting point, but often digressed into the current priorities of the member, and how work at the intersection of AI and healthcare might fit in.
But with that being said…
Expertise is Needed
DC is a transient town. Representatives are on two year election cycles; many staffers are there on year-long fellowships. Oftentimes, staffing is seen as this stepping stone before law school or a job at a think tank - few peoples’ horizons stretch longer than the next year or two. At the same time, Members in safe districts may stay in Congress for decades - their legislative priorities may move much slower.
This constant turnover creates significant knowledge gaps, especially in rapidly evolving fields like AI. Meanwhile, member offices are criminally understaffed, and each staffer has a lot on their plate - I met with folks who oversee not just AI, but all of technology, cybersecurity, defense, science, and innovation.
The complex nature of each of these topics demands deep, nuanced understanding that's difficult to develop in such short timeframes. During meetings, it became clear that many officials and their staff were eager to learn, but struggled with the intricacies of AI applications. Generally, policymakers often lack the technical expertise needed to craft effective, forward-thinking legislation; they serve as idea evaluators, rather than idea originators. That may seem obvious, but I do think this role is why think tanks and lobbyists exist - an industry built on creating compelling ideas, and having them evaluated favorably.
Ideally, this should be a good thing - over time, good ideas should rise on their own merits. In practice, people are being pulled in dozens of directions at once, and it becomes very hard to push any particular thing past the finish line. Every organization is trying to do something about it - BPC is launching AI 101; every think tank feels like it’s hosting lunch-and-learns for offices on AI. But we’ve still got plenty of room to grow.
Socials, Not Suits
It’s possible that more policy work gets done at the Tatte Cafe across the way, than in the Capitol building. Monday through Friday, you’ll sit down with your latte, and overhear suited-up policy wonks talking about carbon credit design on your left, and NATO security guarantees on your right.
Connections are currency, and, in many ways, the papers and memos everyone writes are excuses to chat. I doubt a third of the people I met with have read, or even will read, the memo I put out. Indeed, it was the happy hour afterwards where staffers would come up, and we could get to know each other, trade contacts, and plan for future work together. Each event is a way to dive into another event - staffers operate in a loud, overpouring, cacophony of ideas, and choose which conversations to follow up on depending on their legislative priorities.
It’s a small world
At the end of the day, there’s only 535 Members of Congress. There can’t be more than, say, 10,000 staffers working on the Hill. I found the AI contingent of bright thinkers was far smaller - you could probably fit them all into a single chamber.
And it really felt like a small world. A solid number of folks working on AI went to college with me; the staffer who co-sponsored our panel was the roommate of someone I had met freshman year. In the audience of our afternoon panel, I ran into the professor who taught me CS 103. Half the people at the IAPS happy hour on Wednesday showed up at the FAS happy hour Thursday.
Thanks to all the folks who helped me on this adventure! Everyone who took time to chat, and who put me in touch with others working hard on the Hill. Everyone who helped refine my pitch, clean up my ideas, and reminded me to bring a suit. I’m excited for my next trip there!
Great insight into the world of DC policy! Would love to see an article on what your pitch is and how you went about developing it