The link appeared in the team Slack at 8:47 AM Pacific, February 7, 2026. Navan posted it. No commentary, just the URL. Simon Willison had published his article about the factory.
Jay clicked through first. He read it standing up at his desk, one hand on his coffee, the other scrolling. He read it once quickly, then again slowly. Then he sat down.
"He got it right," Jay said, and there was something in his voice that was hard to name. Relief, maybe. Or recognition. The particular feeling of being understood by someone who had no obligation to understand you.
Justin read the article at his own pace, which was slower. He'd been aware it was coming—Simon had reached out with questions weeks earlier, and Justin had answered them carefully, the way he answered everything carefully, with the precision of someone who believed that clarity was a form of respect.
Simon understood the factory the way Simon understood everything: by pulling it apart and laying the pieces out in daylight. He'd grasped the non-interactive development model, the role of scenarios as external validation, the satisfaction metric, the Digital Twin Universe. He'd understood that the factory wasn't about AI replacing programmers. It was about changing the relationship between humans and code. Humans described the world. Agents wrote the code. Scenarios validated the result.
What impressed Jay most was that Simon had understood the why. Not just the mechanism but the philosophy. The deliberate naivete. The willingness to discard Software 1.0 habits not because they were bad but because they were obstacles to a fundamentally different approach. Simon had traced the intellectual lineage from Justin's observation of the October 2024 inflection point through to the factory's founding in July 2025, and he'd done it without sensationalizing the timeline.
"No breathless AI hype," Navan observed, still reading. "He's just... describing what it is."
"That's why it matters," Justin said. "Simon has credibility because he's precise. He doesn't oversell. When he says something is interesting, people trust that assessment because he's earned that trust by being accurate about a hundred things before."
They spent the next hour in the Slack channel, not discussing the article's content—they lived the content every day—but discussing its framing. The choices Simon had made about what to emphasize and what to leave out. The analogies he'd chosen. The questions he'd raised without answering, leaving them as invitations for the reader to think.
Jay noticed that Simon had mentioned the token spend—the thousand-dollars-a-day-per-engineer figure. He'd presented it without shock, as a data point rather than a punchline. That was the mark of someone who understood the economics rather than reacting to the number.
"He's going to send us traffic," Navan said.
"He already is," Justin replied, watching the analytics dashboard. The factory.strongdm.ai page views were climbing. Not virally, not explosively, but steadily. The way things climb when a trusted voice points and says look at this.
Jay bookmarked the article. He'd read it a third time that evening. Some things deserved three readings.
Simon's coverage is what brought me here. He has this ability to make complex technical systems comprehensible without dumbing them down. The factory deserved that treatment.