How Small Parameters Create Big Shifts

The web you’re using right now exists because of a cascade of unlikely decisions. Not grand plans or corporate strategies, but specific choices by small teams and individuals like you that triggered emergent properties — new behaviors that couldn’t have been predicted from the original components.

The internet was never designed. It emerged. It was predicted. It was manifested. From scattered machines in research labs to billions of people connected worldwide, the web grew out of improbable decisions and accidents of timing. Like galaxies forming from dust, small choices aligned into something vast — a system no one foresaw but all of us now inhabit.

And it’s still emerging. Every click, every link, every choice you make nudges the web. Its future is tipping now, and your choices decide the direction.

Think of it this way: when Tim Berners-Lee wrote the first web browser in 1990, he couldn’t have predicted TikTok, Wikipedia, or cryptocurrency. Yet each of these emerged from the foundational parameters he set in motion. This is emergence in action — simple rules creating complex, unpredictable outcomes.

Today, we’re witnessing a critical shift in those parameters. The same forces that once pushed toward an open web are being reconfigured, nudging us toward something more enclosed. Understanding how emergence works — and how small actors influence it — reveals that the future isn’t predetermined. You have more power than you might think.

The Unlikely Architects of Openness

The free web wasn’t inevitable. It emerged from specific decisions by people who often had no idea their choices would reshape civilization.

April 30, 1993: CERN physicists made a decision that would define the next three decades. They released the World Wide Web protocols royalty-free, then issued an open license to “maximize its dissemination.” This wasn’t the obvious choice — they could have licensed the technology, created a revenue stream, maintained control. Instead, they removed all tolls on implementing web servers and browsers.

1994: Tim Berners-Lee founded the World Wide Web Consortium (W3C), establishing a critical emergent parameter: core web standards must be implementable without paying royalties. This Royalty-Free Patent Policy, formalized in 2003, meant that companies contributing to web standards had to give up patent fees on essential technologies.

1995: The U.S. National Science Foundation made another pivotal choice: shutting down the NSFNET backbone and opening the network to commercial traffic. Before this, the internet was restricted to research and education. This single policy change unleashed an emergent property — the commercial web — that no one fully anticipated.

1998: When Netscape open-sourced its browser code, creating Mozilla, it seemed like a desperate move by a failing company. Yet this decision kept a standards-driven, open browser in the ecosystem, preventing any single company from controlling how we access the web. Today’s Firefox, maintaining about 3% market share, still serves as a crucial check on browser monoculture.

These weren’t coordinated moves by a grand alliance. They were independent decisions that created Constraint parameters — the boundaries within which the web could evolve. Low barriers to entry (anyone could run a server), open standards (anyone could build a browser), and Network effects (more sites meant more value for everyone) combined to generate an emergent system no one designed: the open web.

The stage was set and shared; many recognized its potential and used it, creating a feedback loop that collectively shaped its evolution

These decisions weren’t random footnotes — they defined the parameters of the system itself.

Think of the web operating through four key emergent parameters:

• Constraint Severity
• Tool Use
• Network Effects
• Recursive Self-Improvement

Emergence and the Web’s Parameter Space

The Open Web operates like a city with public streets. Anyone can set up shop (run a server), anyone can visit (type a URL), anyone can give directions (share links). It’s messy, diverse, harder to control — but fertile for creativity and long-term resilience.

The Closed Web functions like a shopping mall. Controlled access points (app stores), curated experiences (feeds), centralized authority (platform policies). It’s sleek, controlled, monetized — but risks monopoly, censorship, and innovation stagnation.

The Four Governing Parameters of Web Emergence:

These parameters don’t just influence the web — they determine what can emerge from it. Change them, and you change what becomes possible. Your daily choices affect all four.

These parameters aren’t theoretical. They show up in every stage of the web’s story — from how we search for information, to how knowledge is shared, to how intelligence itself has emerged online.

From Information to Intelligence

Each emergent property of the web became the substrate for the next emergence. This is how complex systems evolve — through hierarchical emergence where each layer enables unprecedented phenomena at the next scale.

1998: Google emerged by exploiting an emergent property of the open web — its link structure. PageRank treated hyperlinks as votes, using the web’s own topology to measure relevance. This only worked because the Tool Use parameter was open — anyone could crawl and index the web.

2001: Wikipedia emerged from radical openness in the Recursive Self-Improvement parameter. By allowing anyone to edit, it created a knowledge system that improved itself using its own outputs. By 2005, it had 750,000 articles. This massive corpus would later become critical training data for language models.

2015-2016: The open source release of TensorFlow and PyTorch represented another parameter shift. Machine learning tools became freely available, setting Tool Use to maximum openness for AI development.

2017: The Transformer architecture paper — published freely on arXiv rather than patented — provided the blueprint for modern AI. Combined with decades of open web content scraped by Common Crawl, this enabled the emergence of large language models.

2022: ChatGPT’s release represented a new phase transition — AGI- like intelligence emerging from the web’s accumulated information. But here’s the critical point: this emergence was only possible because the web had remained open for 30 years. Wikipedia’s freely licensed content, Reddit’s discussions, Stack Overflow’s answers — all became the training data. A closed AOL-style system would not have produced this emergence

The Gradual Enclosure

But emergence never stops. Change the parameters, and you change what emerges. Today, we’re watching those parameters shift.

2017: The W3C made Encrypted Media Extensions (EME) a web standard, essentially embedding DRM into the open web. The Electronic Frontier Foundation resigned from the W3C in protest, calling it a “betrayal.” This wasn’t just a technical decision — it changed a fundamental parameter. Now, parts of web content could be locked behind proprietary decryption modules.

2023: Reddit dramatically increased its API pricing, effectively killing most third-party apps. Apollo, Reddit is Fun, and other clients that had enabled alternative ways to access Reddit’s content were forced to shut down. This exemplifies a pattern: platforms that once exposed free APIs to encourage an ecosystem now wall them off, changing the tool-use parameter — who can build on top of existing systems.

2024: Google’s AI Overviews keep more user attention on the search results page, reducing outbound clicks to actual websites. An Enders Analysis report found that major UK publishers lost up to 80% of their Google search visibility since 2019, with the decline accelerating sharply since April 2024. This shifts a critical Network effect — the reciprocal relationship between search engines and publishers that sustained the open web.

Apple’s App Store fees and controls (even as DMA-driven changes roll out) illustrate how distribution chokepoints reshape Constraint parameters. Perhaps most telling: Apple’s Eddy Cue testified that Safari search volume declined for the first time in 22 years, attributing the drop to users switching to ChatGPT and other AI tools.

The Next Emergence: AI Inflection Point

We’re now witnessing the parameters shift again, approaching another critical threshold. The same system that enabled AI’s emergence is reconfiguring in ways that may prevent future emergent breakthroughs.

The web is nearing a phase transition: when accumulating parameter changes flip system behavior. Dial-up to broadband birthed YouTube; mobile birthed the app economy. The next winners are unknowns.

On one side, AI pushes toward openness. Apple’s Private Cloud Compute and on-device processing reduce reliance on centralized servers. Standards like C2PA (Content Authenticity Initiative) create open protocols for verifying content provenance. The EU AI Act establishes interoperability requirements rather than platform-specific rules.

But AI also enables new forms of enclosure. When Google’s AI provides direct answers, it breaks the fundamental exchange — publishers provide content, search engines send traffic. Users no longer need to click through, and publishers lose traffic. Meanwhile, exclusive training data deals — like OpenAI’s agreement with News Corp — create new competitive moats around AI capabilities.

Signals of Resilience: The Open Web Isn’t Done

Before accepting that we’re witnessing the web’s inevitable enclosure, consider the push back already emerging in trying to restore it. Google’s overall search volume continues growing even as Safari search declines—people are searching more, not less. Publishers like The New York Times, Wall Street Journal, and Financial Times are successfully building direct subscriber relationships that reduces dependence on both Google and social media entirely.

Counter-movements like Fediverse‘s ActivityPub protocol enables federated social networks without central control, while Meta’s release of LLaMA sparked an open-source AI revolution democratizing AI capabilities.

The creator economy demonstrates another adaptation path: independent journalists on Substack, YouTube creators, and podcast hosts are monetizing through direct audience support rather than advertising. Meanwhile, AI tools themselves remain critically dependent on high-quality source material—if they destroy their training data ecosystem, they undermine their own effectiveness.

Perhaps most importantly, users are beginning to recognize AI limitations. The initial novelty of chatbot answers is giving way to awareness of hallucinations, bias, and the loss of source transparency. This growing skepticism may drive people back toward trusted, attributable sources—exactly what the open web provides best.

Understanding Your Leverage Points

This is where understanding emergence becomes practical. Complex systems have sensitive dependencies — small changes in the right parameters can cascade into large effects. The web’s emergent properties arise from millions of individual choices compounding over time.

Consider how Recursive self-improvement — a system’s ability to enhance itself using its own outputs — determines whether something becomes a dominant platform. Google used search data to refine search algorithms. When you contribute to systems with this property, you’re fueling their growth. When you withhold participation, you’re constraining it.

The web’s future isn’t being decided in boardrooms alone. It’s being determined by millions of micro-decisions compounding every day. Understanding emergence means recognizing that the tomorrow’s web is sensitive to today’s choices — including yours. You are not a passive observer of this phase transition. You are an agent within the system whose actions influence its evolution.

I think we’re approaching what physicists call a bifurcation point — where the system must transition to one of two attractor states. Either the web reorganizes around closed platforms (the shopping mall) or it maintains enough openness for continued emergent innovation (the city).

Convergence is inevitable. The physics of information systems drives toward universal platforms where everything interconnects. But the phase state of that convergence remains undetermined.

Every click, every link, every choice nudges the web. The question isn’t whether you influence what emerges — you already do. The question is whether you’ll choose to do it intentionally.

– Sail