The Pattern
Stack Overflow questions dropped from 400,000 monthly in 2022 to roughly 150,000 by late 2024, according to data cited by The Pragmatic Engineer. ChatGPT's November 2022 launch correlates with the steepest decline, though the platform's trajectory was already downward.
Meanwhile, DEV.to reports growth in both content and engagement. The divergence isn't just about AI - it's about community mechanics and what developers actually need from knowledge platforms in 2026.
What Changed
Stack Overflow built its reputation on expert contributions and rigorous quality control. That same gatekeeping now works against it. New developers face high bars to participation. Experienced developers increasingly turn to ChatGPT for quick answers, even if those answers occasionally hallucinate or reference outdated patterns.
DEV.to took a different approach: lower barriers, badge systems critics claimed rewarded quantity over quality, but challenges that emphasize substantive content. The platform's curation - weekly digests, follow recommendations - surfaces quality without requiring contributors to navigate complex reputation systems.
Enterprise Implications
For CTOs and engineering leaders, this matters beyond which platform developers prefer. Three implications:
Internal documentation strategy: If public platforms struggle with quality and discoverability, internal knowledge bases face the same challenges. The Stack Overflow model - strict quality gates - creates maintenance burden. The DEV model - loose contribution, heavy curation - requires different investment.
Platform engineering context: DORA 2025 data shows 55% of organizations adopted platform engineering practices, with elite teams hitting multiple daily deployments. Those teams need fast answers, not research projects. AI tools provide speed; human-curated communities provide context.
DevEx investment priorities: Companies spending on developer experience need to understand how their engineers actually find answers. If it's ChatGPT plus selective community engagement rather than comprehensive platform expertise, training and tooling strategies should reflect that.
The AI Factor
Blaming AI oversimplifies. Wikipedia maintains quality through aggressive curation and nonprofit focus. Medium declined behind paywalls and stock imagery. Gutefrage.net in Germany reportedly dropped below 2.0 on TrustPilot as quality collapsed.
AI accelerates existing dynamics - it didn't create them. Platforms with poor UX or misaligned incentives were already struggling. ChatGPT just made the alternatives more obvious.
What We're Watching
Whether DEV.to's growth continues as it scales. Whether Stack Overflow can adapt its model - or whether that model served its era and we're moving on. And whether enterprises build internal knowledge systems that learn from these platform experiments or repeat their mistakes.
The 2010s might have been a golden age of free collaboration. What comes next depends on whether we figure out sustainable models for quality knowledge sharing, with or without AI in the loop.