You Cannot Figure This Out Alone
I spent an embarrassing amount of time in the first months of 2025 trying to figure AI out by myself.
I read articles. I followed researchers and practitioners on social media. I signed up for newsletters. I listened to podcasts (so many podcasts).
And at the end of all of it, I had a lot of information and not very much actual capability. I couldn't have told you, concretely, how to apply any of it to a real problem I actually cared about.
Part of what made this so disorienting was the sense that everyone was talking about AI and nobody was actually helping. The gap between the volume of the conversation and the usefulness of the guidance was vast. And I don't think that was just my experience.
Look at what's actually being offered to most workers right now. Large organizations, small organizations, public agencies — the interventions range from nothing at all to what a friend recently described as "AI ping pong." Someone drafts an email using AI. The output is so bloated and incomprehensible that the recipient needs to use AI to summarize it before they can respond. Nobody learned anything. Nobody's work got better. The tool just got inserted into an existing workflow and called it a day.
That's not upskilling. That's exposure with a press release attached.
And if you go looking for coverage, for local journalism or civic blogs or even the candidate forums that take this seriously as a workforce question for the people who actually live here — good luck. Erin and I started this blog series partly because we couldn't find that conversation happening anywhere. We searched local news. We looked through smart-growth and development publications. We checked what local leaders were saying. There is very little that speaks directly to DC's working population about what AI means for their careers, their communities, their daily work lives. The national tech press covers the technology. The political conversation gestures at trade schools. The space in between — where most people actually live — is largely empty.
The greater DC region leads the country in job loss and in unemployment rate. Last week, a WJLA story reported on a new workforce analysis finding that DC, Maryland, and Virginia are among the places with the highest AI exposure in the country. The analysis, which looked at more than 800 occupations and rated how easily AI could perform or assist with the tasks involved, found that more than 55% of DC workers are in jobs where AI threatens or transforms the work — the highest rate in the country. Maryland came in at nearly 40%. Given what this region's workforce actually looks like, those numbers are absolutely worth paying attention to.
The vacuum around this discussion (and the lack of action) is part of why people feel so alone with this. It's not just that the technology feels overwhelming to keep up with. It's that the institutions that are supposed to orient us — employers, local government, news organizations, civic leaders — have mostly opted out of that responsibility. Which leaves individuals feeling like the confusion they're experiencing is personal, when it's actually structural. It's not that you're slow to adapt. It's that you've been handed a tool, pointed at your desk, and told to figure it out.
What changed things for me wasn't more reading. It was other people.
It was sitting in a room — literally, it started in a room and evolved into a team project over weeks — with people who brought different things to the same questions. A procurement leader. A corporate strategy leader. An enterprise systems administrator. A product owner. An attorney. A customer relationship manager. An international development consultant.
None of us were experts in AI. All of us knew something the others didn't. And the process of working through problems together, arguing about what we were seeing, pushing back on each other's assumptions — that produced something that individual reading simply cannot: grounded, tested, contextual understanding.
This is not a new insight about how learning works. It is an old one that we keep having to rediscover. Deep expertise gets built in communities of practice, not in solitude. We learn by doing, and we learn faster and better when doing happens in relationship with others who are working on the same things.
The specific challenge with AI is that the landscape is changing fast enough that individual learning loops can't keep up. By the time you've read enough to feel confident, the tool has updated, a new one has arrived, the use case you were planning for has changed. Trying to achieve fluency through individual consumption is like trying to study a map of a city that's being demolished and rebuilt while you read it.
What actually works is building a community that learns continuously — where each person is tracking different corners of the landscape, where insights get shared and tested in real time, where the collective intelligence of a group compounds, instead of each person starting from scratch. Where you can bring a question you've been embarrassed to ask anywhere else and find out that four other people have been sitting with the same one.
That's not a theory. I've watched it happen. A few days of experimenting, asking questions, sharing results with a team led to more learning than months of individual research had. Because the format required people to work together on something real. Problems that mattered to specific people. Solutions that had to actually function. Disagreements that had to be resolved, not just observed.
The people who figure this out together will fare better than the people who try to figure it out alone. And the first step is finding the room where other people are also trying — and being honest, in that room, about how much none of us has figured out yet.
