Hey there,
In late 2020, HERE Technologies reorganized its developer relations team. The company had been moving toward a functional structure for years, and we were one of the last teams to make the transition. When our turn came, the team was split up. Some of us went to product management, some to marketing, others to engineering. At the same time, the company explicitly de-prioritized the developer community work we'd been focused on for about three years.
Nobody was fired. Nothing was formally canceled. The work just got absorbed into other priorities.
To understand what was lost: for three consecutive years, we had exceeded our goal of doubling the developer community. We had a very large, growing community and were working on innovative ways to engage them including new toolkits, community-seeded projects, creative approaches to developer engagement that were genuinely pushing boundaries. All of it stopped at the reorg.
Within six months, the four developer evangelists who were the company's relationship with its developer community had all left. My boss, the head of developer relations, left. Several product people left. The institutional knowledge, the relationships, the momentum, it all quietly walked out the door. The developer community that had been doubling year over year immediately flattened, then declined.
Then, to add insult to injury, about a year later, leadership recognized the mistake and they re-prioritized developer relations. But by then, there was almost nothing to rebuild from. The people were gone. The community relationships had atrophied. The organizational muscle that had taken years to develop had dissolved in months.
They didn't lose a strategy document. They lost a capability.
"They didn't lose a strategy document. They lost a capability."
Last week, four astronauts headed to the moon for the first time since December 1972. Artemis II launched last Wednesday, and as I write this, the Orion capsule is approaching the moon for a flyby of the lunar far side. It's an incredible and historic moment.
But throughout all of this excitement and wonder, I've been quietly bothered by the fact that it's been 53 years since we went to the moon.
We didn't stop going to the moon because we failed. We stopped going because we succeeded. President Kennedy's goal was met. The Soviets were beaten. The political urgency evaporated, and the budget followed the motivation. NASA leadership argued that the goal had been achieved and there was no reason to keep accepting the risk. Three planned Apollo missions were canceled. Then everyone moved on.
And what happened next? Nothing...nobody dismantled anything, the Saturn V rockets weren't destroyed, the engineering blueprints weren't burned. But the skilled machinists who built the hardware retired, the supply chains dissolved, and the institutional knowledge scattered. By the time we decided we wanted that capability again, it was gone. NASA couldn't restart Apollo any more than HERE Technologies could restart developer relations by simply re-prioritizing it on a slide deck.
They had to start over. It took decades and billions of dollars.
I've written before about the Fresh Start Trap — the enthusiasm cliff that kills most transformation initiatives around Week 4. But there's a quieter failure mode that hits the initiatives that actually survive that cliff.
It looks like this: Last year, I trained two product teams at the same company on AI capability building. Same material, same frameworks, same level of investment. I also trained an AI champion on each team, someone responsible for keeping the momentum going after I left.
One of those teams is still pushing boundaries today. They're using the training material as part of onboarding for new hires. Their capability is compounding.
The other team? I haven't heard from them. I don't know exactly what happened, but I can guess. A priority shifted. The champion got pulled onto something else. The Slack channel went quiet. Nobody canceled anything. It just slowed, then stopped.
Twelve months from now, someone at that company will ask why one team uses AI effectively and the other doesn't. Someone will suggest a pilot.
This is the Apollo Trap. The most dangerous moment for a capability isn't when it fails. It's right after it succeeds.
"The most dangerous moment for a capability isn't when it fails. It's right after it succeeds."
The conventional framing of AI transformation is about building capability. Getting executive buy-in. Running a pilot. Proving ROI. And all of that matters.
But HERE Technologies had built the capability. NASA had built the capability. In these examples, the capability wasn't the problem. The problem was that nobody designed a system to keep it alive after the original urgency faded.
Capability isn't something you build and then have. It's something you feed or lose. The people who hold the knowledge leave. The context that made the workflows effective decays. The organizational attention that gave the work oxygen moves on to the next priority. Not with a dramatic cancellation, but with a quiet, perfectly rational reallocation of resources.
We're doing it differently this time around, though. Artemis is designed differently than Apollo. Apollo was a flag-planting mission — get there, prove we can do it, come home. Artemis is building infrastructure for sustained presence. It's creating reusable systems, international partnerships, and a foundation designed to survive changes in administration and priority.
That's the question worth asking about your AI initiative. Are you building Apollo, or are you building Artemis?
Because if you're building Apollo — if the goal is to prove AI works and then move on — you'll succeed. And that success will be the most dangerous thing that happens to your team's AI capability.
Break a Pencil,
P.S. If your organization has already proven AI works and you're wondering how to make it stick, that's exactly the problem I help product teams solve. Let's talk.
P.P.S. Forward this to a product leader who's about to celebrate a successful AI pilot. They might need it more than they think.
P.P.P.S. This is the 53rd issue of Broken Pencils. Sometimes the universe writes its own transitions.