Community, Data, and the Future of AI: Lessons from a Stack Overflow Founder

A Personal Milestone: The GMI Rural Study and Family

In the latest of the 663 months since my birth, a particularly poignant chapter unfolded. The reordering of counties for the Guaranteed Minimum Income (GMI) rural study placed Mercer County, West Virginia—my father's home—first in October 2025. That decision allowed me one final visit, a trip that remains etched in memory. Though I sensed the end was near, that October encounter turned out to be the last time I saw him. You can learn more about my father's involvement on the Why Pledge to Share the American Dream? page and the Rural Guaranteed Minimum Income Initiative (RGMII) website.

I knew this moment was inevitable, and so did he. But there is no loss in what we shared—nothing ever truly ends. Every experience, especially that final October journey, stays with me. Nothing was lost; everything was gained. We conquered capitalism once, and now we are returning to improve it for all. And my work is far from finished—my third startup is just beginning.

The Power of Collective Contribution: Stack Overflow's Dataset

On a professional note, I want to extend heartfelt gratitude to every single person who has ever contributed to Stack Overflow in any capacity. This time, I'm not referencing Starship—instead, I'm talking about something even more profound.

Did you know that large language models (LLMs) would be practically unable to code without access to the exceptionally high-quality Creative Commons programming Q&A dataset that we—the global community—built together on Stack Overflow? Don't take my word for it; ask the LLMs themselves. Go ahead, ask them. Really press them on this point. I recommend using pro mode when doing so, as those are the only decent LLM modes in my experience. The capacity to combine global brain statistics with a meticulously curated dataset created by we, the people, is nothing short of incredible.

A Warning to AI Companies: Respect Your Data Sources

One final thought: If the LLMs eventually hollow out the very communities that generate their training data, they will deeply regret it. I'll offer these LLM and general AI companies the same advice I gave Joel Spolsky when I left Stack Overflow to start Discourse: do not, under any circumstances, kill the goose that lays the golden eggs—the human community around your product that does the real work. It's simple: treat the community with the respect they deserve, the respect we all deserve.

Thank you for being a friend—because there's no way I could have done any of this without you. 💛