Stack Overflow.

For nearly fifteen years, Stack Overflow was the most important Q&A site in software development. Its decline since 2018 — and especially since the release of ChatGPT in November 2022 — is one of the most consequential shifts in the developer-platforms landscape of the last decade.

History

• Founded. 2008 by Jeff Atwood and Joel Spolsky.
• Launched publicly. September 2008.
• Sister sites. Server Fault (2009), Super User (2010), grew into Stack Exchange Network covering hundreds of topics.
• Acquisition. Prosus / Naspers acquired Stack Overflow in 2021 for approximately $1.8B.
• CEO history. Joel Spolsky (founder), Prashanth Chandrasekar (CEO from 2019 onward through Prosus acquisition era).

The rise (2008–2017)

Stack Overflow’s success rested on a small set of design choices that, in hindsight, defined the modern Q&A site:

• One canonical answer per question. The top-voted answer wins; duplicates merged.
• Reputation system. Points for upvoted answers; privileges unlocked at score thresholds.
• Strict scope. Off-topic questions closed; chat moved off-site.
• No threads on answers. Comments only; discussion happened in the answer-edit history.

At its 2014–2017 peak, Stack Overflow received roughly 200,000 new questions per month. The site was almost universally the first Google result for any technical search; entire careers depended on its answer archive.

The decline (2018–2025)

Question volume began declining from its peak in 2018 — four years before ChatGPT. Multiple causes contributed:

• Moderation severity. New users posting their first question often had it closed as “duplicate,” “too broad,” or “lacking research,” producing an unfriendly first-experience.
• Mod-and-user-base burnout. Long-tenured power users contributed less over time without commensurate inflow of new contributors.
• Documentation and chat fragmentation. Discord and Slack communities took over real-time help that Stack Overflow could not provide.
• Diversifying knowledge surfaces. dev.to, Hashnode, GitHub Discussions, project Discords, and others absorbed Q&A activity.

The decline accelerated dramatically after late 2022:

• ChatGPT launch. November 2022. Many developers’ first reaction was that ChatGPT answered programming questions as well or better than Stack Overflow searches.
• Copilot maturity. GitHub Copilot’s quality improvements through 2023 reduced the need for many Stack Overflow lookups.
• 2023 moderator strike. A multi-week strike against Stack Overflow’s policies on AI-generated answers reduced moderation throughput and surfaced governance tensions.
• API repricing. Stack Overflow restricted its API access and added paid commercial tiers, limiting third-party integration.

By late 2025, question volume had dropped approximately 95% from the 2014–2017 peak — back to roughly 2008-launch-year levels. Traffic was down by an estimated 75% from peak. The site remains a substantial archive — roughly 84% of developers still browse it, per ongoing reporting — but almost no one contributes anymore.

Stack Overflow Developer Survey

Despite the platform’s decline, the Stack Overflow Developer Survey remains the most-cited annual snapshot of developer practice. Conducted since 2011, the survey covers tools, languages, technologies, compensation, career topics, and (since 2023) AI tool usage. Results are widely cited and used by DevRel teams to track developer-population trends.

Implications for DevRel

The decline of Stack Overflow has shifted the discovery and learning surfaces developers use:

• AI assistants (ChatGPT, Claude, Copilot, Cursor) are now the first-touch for many technical questions.
• Project Discords and GitHub Discussions absorb peer-help conversations.
• YouTube tutorials absorb learning-style content.
• dev.to, Hashnode, and personal blogs absorb long-form written content.

For DevRel programs:

1. Stack Overflow is no longer reliable as a community help layer. Investing in answering tagged questions yields diminishing returns.
2. The archive still matters for search. SEO continues to surface Stack Overflow answers for many developer queries.
3. Optimising content for AI ingestion is the modern equivalent of optimising for Stack Overflow search.
4. The Annual Developer Survey is still essential market intelligence even as the platform declines.

Stack Overflow for Teams and AI

Stack Overflow’s commercial product (Stack Overflow for Teams) and AI-product initiatives (OverflowAI) are attempting to convert the archive into enterprise knowledge-management value. The success of this pivot is unclear as of 2026.

What replaced it

For pure Q&A function, there is no single replacement. The functionality has fragmented across:

• AI assistants for immediate answers.
• Discord / Slack / Discourse for peer help.
• GitHub Discussions and Issues for project-specific questions.
• dev.to / Hashnode for written explanations.
• Reddit / Hacker News for context, opinions, and learning what to learn.
• YouTube for visual / tutorial-style explanation.

The fragmentation has costs: knowledge is now harder to search across silos. The Model Context Protocol (MCP) and AI-mediated retrieval are partly attempts to re-aggregate this fragmented knowledge for AI-coding agents.

See also

• ./hacker-news.md
• ./reddit.md
• ./dev-to-and-hashnode.md
• ../11-trends/ai-and-llms.md