The Apache Iceberg Meetup Bay Area ran on the evening of September 23, 2026 at the Silicon Valley AI Hub, 135 Constitution Drive in Menlo Park: dinner and networking from five, talks from a quarter to six, and more networking until eight. Three hundred and fifteen people registered.
Russell Spitzer gave a talk whose title is a RoboCop joke and a real argument: Part Man, Part Machine, All Open Source. He is on the project management committee of several Apache projects, and his subject was how large language models have changed the way he does open-source work.
He opened with a complaint. LLMs ruined Advent of Code for him — the annual December puzzle calendar, one problem a day, each harder than the last, the kind of thing he looks forward to. He walked through the blog posts that measured what happened: by the 2024 calendar, Claude 3.5 Sonnet was solving roughly eighty per cent of the first ten days, the easier half.
The talk goes from there into what that shift means for someone who maintains a project rather than competes in a puzzle. The part worth carrying away is what he now asks of a pull request. Not the prompt. A clear summary of why the problem matters, what solving it is worth, and what the impact will be on other users of the project — the things that get missed, he says, a lot these days.
The recording is below.