PyCon US is coming back to California next month, and for the first time in over a decade, West Coast developers won’t need to fly across the country to attend. The conference runs May 13-19 in Long Beach, with core talks from Friday the 15th through Sunday the 17th.
What’s more interesting than the location is what’s happening with the programming. PyCon has added two dedicated tracks this year: an AI track on Friday and a Security track on Saturday. This isn’t just schedule shuffling. It’s a recognition that Python has become the lingua franca of machine learning infrastructure, and the community needs dedicated space to discuss it properly.
The AI Track Signals Something Bigger
The AI program is being run by Silona Bonewald from CitableAI and Zac Hatfield-Dodds from Anthropic. That’s not an accident. These aren’t academic researchers giving theoretical talks. These are people building production systems that handle real-world AI workloads at scale.
I’ve been attending PyCon for over twenty years now, since 2005, and the vibe has always been different from other tech conferences. Even with 2,000+ attendees, it manages to avoid feeling like a corporate trade show. The lightning talks are packed. The PyLadies auction is chaotic and fun. The sprints let you sit next to maintainers of libraries you actually use and contribute patches in real time.
But adding a dedicated artificial intelligence track feels significant. Python won the AI framework wars not through corporate mandate but through organic adoption. PyTorch, TensorFlow, JAX, scikit-learn, NumPy, pandas. The entire modern ML stack is Python-first. PyCon acknowledging this with dedicated programming is overdue.
Open Spaces Over Scheduled Talks
The scheduled talks are fine, but the real value at PyCon has always been in the edges. Open spaces let anyone reserve a room and start a conversation. It’s unconference-style programming baked into the main event. I’m planning to spend most of my time there this year, probably talking about Datasette and what I’m calling agentic engineering.
Agentic engineering is still a weird term that nobody quite agrees on, but the concept matters. We’re moving past the “wrap an API call in a chatbot” phase of AI integration and into something messier and more interesting. Systems that can plan, revise, and execute complex tasks without constant human guidance. That requires new patterns, new tooling, and new ways of thinking about system design.
These conversations happen better in open spaces than in scheduled 30-minute talks. You can’t workshop ideas in a lecture hall.
Why the Hotel Block Matters
The PSF is asking attendees to book in the official hotel block, and there’s actual infrastructure reasoning behind it. Conference economics are weird. Hotels give discounts on conference space based on room night commitments. If attendees book outside the block, the conference eats the cost difference. For a community-run event like PyCon, that directly impacts whether they can afford to keep registration prices reasonable.
I’m on the PSF board, and PyCon has historically been a major funding source for the foundation. That money funds things like PyPI infrastructure, security audits, and grants to Python projects worldwide. When people ask why conference logistics matter, this is why. The boring operational details cascade into whether the ecosystem can afford to maintain itself.
The West Coast location also changes who can attend. Flying from San Francisco or Seattle to DC or Pittsburgh isn’t cheap. Making PyCon accessible to California-based developers means more people from the companies actually building production Python systems can show up without needing corporate travel approval.
Maybe the most interesting thing about PyCon’s evolution is that it’s still genuinely community-driven even as Python has become critical infrastructure for half the internet and nearly all modern AI systems.