> StackOverflow now gets fewer than 1000 questions a month

As a public community forum, Stack Exchange (and specifically Stack Overflow) is effectively dead in terms of active community growth and daily engagement, but the parent company remains highly profitable as an AI data supplier.
Public data shows that by the end of 2025, monthly new question volumes plummeted back down to 2008 launch levels—effectively erasing 17 years of user growth.
The platform's decline and current status are driven by a mix of cultural flaws and the shift toward generative AI:

Why the Community Aspect Died

• The AI Migration: Developers have overwhelmingly migrated away from public forums in favor of AI tools like ChatGPT, Claude, and GitHub Copilot. AI provides immediate, customized, and context-specific code debugging without long wait times or human gatekeeping.
• A Toxic Moderation Culture: Long before the rise of AI, users were already abandoning the platform due to a notoriously hostile environment. The heavy-handed use of the "duplicate hammer," rapid question closures, and downvoting alienated junior developers and frustrated veterans trying to post about modern framework updates. Many developers openly celebrate the shift to AI simply because "AI doesn't judge you for asking a dumb question".
• Fragmentation: Active technical discussions have largely moved away from static Q&A sites and into private, fluid ecosystems like GitHub Issues, Reddit, and dedicated Discord servers.

How the Company is Surviving (and Thriving)

While developers have stopped posting, Stack Exchange Inc. has successfully pivoted its business model from a community forum into an AI infrastructure company:
• The Data Moat: Stack Exchange sits on over 15 years of meticulously curated, human-verified engineering data. Because large language models require high-quality, human-written edge cases to prevent "model collapse," this data has become incredibly valuable.
• API Licensing: Instead of relying entirely on ad traffic, the company has signed highly lucrative data-licensing agreements with major AI labs—including OpenAI and Google—to legally use the platform's historical archive for model training.
• Enterprise Products: They have actively scaled Stack Overflow for Teams (and private internal knowledge bases), selling securely hosted versions of their Q&A infrastructure directly to corporations.
Ultimately, Stack Exchange is no longer the thriving, active town square of the internet it once was; it has transitioned into a highly profitable, static archive used to power the very AI systems that replaced it.
If you are looking for alternative places to get development help or want to talk about how this affects code quality, let me know:
• Are you looking for alternatives to get human programming help?
• Do you want to know how the decline of human data affects future AI models?
• Are you curious about how other Stack Exchange network sites (like Math or Sci-Fi) are holding up?