In the mid-2000s, the term Web 2.0 emerged to describe a shift from static websites to interactive, participatory platforms. This era saw the rise of blogs, wikis, social networks, and other technologies that turned users into creators of content. Jakob Nielsen, a leading web usability consultant, observed a recurring pattern that became known as the 1/9/90 rule:
-
1% of users routinely created original content (e.g., posting on forums or publishing blog posts).
-
9% of users participated by commenting, editing, or lightly modifying existing content.
-
90% of users remained “lurkers,” consuming content without engaging in visible ways.
This concept, documented on Nielsen Norman Group, helped explain why the majority of activity on early platforms (e.g., Wikipedia, Reddit) stemmed from a relatively small group of dedicated contributors.
From User-Generated Content to AI Development
Fast forward to the present day, and the internet is no longer about passively reading blog posts or occasionally posting status updates. Instead, the focus is on advanced technologies like artificial intelligence—large language models, computer vision, speech recognition, and more. But the underlying participation dynamic remains strikingly similar.
-
1% (Foundational Model Creators) – AI’s equivalent to the power contributors in Web 2.0. These are organizations and researchers building fundamental AI models. Today’s large language models or sophisticated neural networks often come from this small group (e.g., OpenAI, Google DeepMind).
-
9% (Specialized Solution Builders) – They adapt and customize these foundational tools for specific industries and use cases, analogous to those who, in Web 2.0, took existing ideas and extended them (e.g., writing plug-ins or curating niche communities).
-
90% (Off-the-Shelf Users) – The majority of businesses or individuals who consume and deploy AI “as is” without major modification, much like how most users in Web 2.0 simply read posts or watched videos without actively contributing.
⠀
Customization vs. Commoditization
A key driver of AI adoption today is the tension between customization and commoditization, which also mirrors what happened during the Web 2.0 boom:
-
Customization: In the Web 2.0 context, customization might have meant building your own blog platform or cultivating a community around a particular niche. In AI, it means tailoring an existing AI model to one’s domain—like refining a language model for healthcare diagnostics or creating hyper-specific image recognition systems.
-
Commoditization: Once a few major players (e.g., WordPress, Myspace, later Facebook) dominated Web 2.0, publishing and engaging online became “plug-and-play.” Similarly, many AI tools are now provided via easy-to-use interfaces or APIs, making it cheaper and simpler to deploy standard solutions. The upside is accessibility; the downside is a reduced opportunity for competitive differentiation if everyone else is using the same off-the-shelf product.
Usage and Utility: Parallels from Then to Now
Cost Structure
-
Web 2.0: Hosting a blog or operating a small forum became cheap as infrastructure costs plummeted and platform providers (e.g., Blogger, WordPress) commoditized the creation process.
-
AI: Foundational model creators bear massive compute and R&D expenses, while most users pay for usage through subscriptions or API calls. This keeps entry costs manageable for the 90% and powers a long-tail of niche AI integrations for the 9%.
Participation Bottlenecks
-
Web 2.0: Users required motivation and some technical know-how to move from lurker to contributor.
-
AI: Companies need data engineering expertise and a level of AI fluency to go from off-the-shelf usage to specialized solution-building, leaving many to remain “lurkers” in the 90%.
Innovation Concentration
-
Web 2.0: A handful of platforms set the agenda (YouTube, Facebook, Twitter), while others followed their lead.
-
AI: A few major labs and tech giants shape the development of leading-edge AI technology. The rest depend on their breakthroughs, mirroring the same disparity in “creation vs. consumption.”
Looking Ahead
Just as Web 2.0 eventually evolved into social networks, ubiquitous mobile apps, and new media models, AI is poised to continue its metamorphosis toward broader accessibility:
Lowering Barriers
No-code and low-code platforms allow more organizations to adapt AI to their own needs. Over time, this may shrink the 90% category as more companies gain the expertise—or at least the toolkits—to do some customization themselves.
Ongoing Commoditization
Similar to how social media platforms became virtually indispensable, AI models will continue to integrate deeper into everyday products—from CRMs and ERPs to consumer-grade apps. This will reduce friction for adoption but also intensify competition among solution providers.
Ethical and Regulatory Considerations
With AI’s increasing ubiquity, the 1% of AI creators find themselves under scrutiny for data privacy, algorithmic bias, and potential societal harms. This mirrors content moderation battles in Web 2.0, but the stakes are higher given AI’s power to shape decisions at scale.
Conclusion
From user-generated content to cutting-edge AI, the 1/9/90 rule consistently appears wherever participation and expertise have a high barrier to entry. A small fraction invests heavily in creating the core technologies—whether that’s publishing a majority of the content or building foundational AI models—while a slightly larger group refines or adapts those innovations, and the bulk simply consumes them. Understanding how this dynamic has evolved from Web 2.0 to modern AI ecosystems can help companies and innovators decide where they want to stand on the customization-commoditization spectrum, and how they might move up the curve to unlock greater strategic advantage.
