Four Hundred Builders, One Day, No Filler
Inside the first Pragmatic Summit
You know you’re in the right place when the hardest decision of the day is choosing between a Martin Fowler and Kent Beck panel or a Simon Willison session. That was the Pragmatic Summit in a single moment.
Gergely Orosz hosted his first in-person Pragmatic Engineer event this week in San Francisco. A one-day summit, capped at 400 attendees (over 2000 applied), and a speaker lineup that had no business being assembled for a single day. The apply to attend format was a smart call. The room was full of active practitioners, people building with AI right now, not just following the discourse. It landed in a sweet spot between the intimacy of a meetup and the sprawl of a re:Invent.
The Middle Is the Tough Spot
A thread that surfaced across multiple sessions: mid-career developers are in the hardest position right now. It came up in the Fowler and Beck panel I attended, and Laura Tacho from DX made the same observation independently. The playbook mid-career developers followed got rewritten mid-chapter. Senior engineers have the judgment to leverage the tools. Junior developers never knew anything else. The middle is navigating without a map.
This is something I wrote about recently. Hearing multiple speakers land on the same theme was striking. Not because it validated the argument, but because it suggests the pattern is becoming visible enough that the most experienced voices in the industry are naming it.
DX Is More Important Than Ever
Laura Tacho, CTO of DX, surfaced what might be the sharpest irony of the AI era: leadership is finally paying attention to developer experience. Not because developers asked. Because agents are choking on low-DX ecosystems.
Humans quietly adapted to poor documentation, tangled dependency graphs, and undiscoverable APIs for years. AI tools don’t adapt quietly. They fail visibly. The result is that organizations are suddenly motivated to fix foundational DX problems they’d been ignoring. AI didn’t create these problems. It made them impossible to ignore.
The Moments That Stuck
The closing session was entertaining. Thomas Dohmke was visibly riding the high of announcing Entire, his new startup building tools to capture the context behind AI-written code. The company raised a $60M seed at a $300M valuation, the largest seed round in dev tools history. He spent part of the panel cheekily accusing Gergely of shit posting on X. Rajeev Rajan, CTO of Atlassian, admitted he had to buy a personal laptop to try Claude Code because enterprise IT wouldn’t let him install it on his work machine. We’re in a room full of people reshaping how software gets built, and one of the most senior engineering leaders in the industry can’t install the tools on his work laptop. That’s where we are. The ambition is miles ahead of the infrastructure to support it.
After Hours
The day didn’t end at the summit. That evening I attended the inaugural Hyper Engineering Meetup at Kernel Labs’ new space. Smaller, more casual. Three early-stage startups demoed tools they’re building or using internally. The energy was different from the summit but pointed in the same direction: builders, building, sharing what they’ve learned. I topped off the evening with my first Waymo ride back to the hotel. A fitting end to a day spent talking about the future.
The Pragmatic Summit confirmed something I keep returning to. The craft isn’t contracting. The problems are getting harder, the tools are getting cheaper, and the people in that room are figuring out what comes next. Four hundred of them applied to be there. That alone tells you something.