# Web4.0 = Death of Internet ## A working note on AI, the web, advertising, and journalism --- ## 1. Starting point The problem here is the weakness of the view that “AI assists journalists.” This view confines AI within existing workflows: transcription, summarization, translation, drafting articles, organizing materials. All of these are useful, but they are not the essence of the AI era. The real question is not whether AI improves journalistic efficiency, but whether the profession of the journalist itself remains necessary. Statements like “AI is only a tool, human judgment is essential” often function as a way to preserve existing roles. But what AI is causing is not assistance — it is structural transformation. AI does not simply help journalists. It decomposes their work and redistributes it. --- ## 2. The decomposition of journalism Journalism is not a single function. It is a bundle of functions: - discovering signals - gathering information - structuring narratives - verifying claims - distributing content AI interacts with each of these differently. Some parts are automated. Some are displaced. Some become irrelevant. The key point is: there is no guarantee that the bundle stays together. --- ## 3. The collapse of the expression barrier What AI has fundamentally changed is not productivity, but the barrier to expression. Writing, video, images, coding — all forms of expression are now accessible. In Web2.0, expression was still constrained by talent and effort. Platforms enabled creators, but not everyone became one. In Web3.0, there was an attempt to remove platforms and directly connect creators and audiences through tokens. This largely failed. In what we might call Web4.0, expression becomes ubiquitous. Everyone can produce. --- ## 4. What happens when everyone can express? If everyone can express, then expression itself loses scarcity. When expression is no longer scarce, the bottleneck shifts elsewhere. It shifts to: - attention - filtering - interpretation - trust And increasingly: - machine readability --- ## 5. The rise of AI-mediated consumption The web is no longer primarily read by humans. It is increasingly: - read by AI - summarized by AI - filtered by AI - delivered by AI Humans interact with the output, not the raw web. This creates a shift: from “writing for humans” to “writing for systems that write for humans.” --- ## 6. Noise over curation Traditional media emphasizes curated, structured information. But AI does not necessarily need curation. AI can organize information itself. What AI needs is **raw input** — noise. Unstructured, fragmented, inconsistent data. This suggests: - edited information may be less valuable than assumed - “clean narratives” may be less useful than messy reality --- ## 7. The input layer problem There is a common idea: journalism becomes the “trusted input layer” for AI. This is not very convincing. If AI needs noise, then heavily edited information may not be optimal. The input layer may instead become: - social media - sensor data - continuous behavioral streams Possibly controlled by: - large AI companies - vertically integrated platforms --- ## 8. Vertical integration AI companies may not stop at models. They may control: - data input (platforms, devices) - processing (models) - output (interfaces, feeds) This resembles a vertically integrated media system. Traditional media organizations may not occupy a central position here. --- ## 9. Advertising transformation If humans no longer browse the web directly, advertising changes fundamentally. From: - banners - search ads - feeds To: - recommendations - embedded suggestions - agent-mediated decisions Advertising becomes: **indistinguishable from advice** --- ## 10. Web4.0 as “Death of Internet” “Death of Internet” does not mean disappearance. It means transformation: - from human-readable space - to machine-mediated environment The web still exists, but its primary users are no longer humans. --- ## 11. The role of the human In this system, what remains for humans? Possibilities: - observation - anomaly detection - maintaining continuity over time - being a persistent node Not necessarily “writing articles,” but maintaining a position within the flow. --- ## 12. Trust and persistence Trust may not come from legacy institutions. It may come from: - continuous presence - consistent output - long-term observation A kind of “madman strategy”: continuing, regardless of immediate visibility. --- ## Status Version 0.1 Continuously updated.