Narrative Goldmine

Death of the Internet

18 min read

Properties

Type
  • Note
Status
  • stable
Generated
  • by: process:vault-migrate/1.0 · at: 2026-09-22T12:41:58.55607396Z

The story arc of THE MISTAKES

  • make the internet

  • centralise it to monetise it

  • choose passwords instead of verified cryptographic end points

  • enshitification through greed on search monopoly

  • closed ecosystems

  • bots everywhere

  • 2FA everywhere, oh god it’s everywhere

  • pay for mitigation with increasingly intrusive gatekeeping and more enshitification

  • invent AI, before asking if we needed AI

  • immediately attempt to commodify it, bots proliferate

  • agree that search is now unusable

  • make people pay for essentially the late 90’s search experience via AI

  • remove viable search from 98.5% of the world

  • boil the oceans

  • everyone build an underground bunker (not you, them)

  • collect underpants

  • …

  • profit??

  • This isn’t a new problem, but seems perhaps to be accelerating.

    image.png|500

  • https://twitter.com/emollick/status/1741287767773163626?

    https://twitter.com/emollick/status/1741287767773163626?

AI augmented search is wasteful

Screenshot 2024-06-13 091134.png|600

Economic Exclusion

  • The nexus where payments, search, and AI converge marks a transformative period for the internet. The traditional model of free data access and ad-supported revenue is being challenged by AI capabilities necessitating paid data agreements and enhanced monetisation strategies. This evolution impacts everything from user behaviour and regulatory frameworks to financial viability and market competition, signalling a profound shift in the digital ecosystem.
  • AI-driven search engines are changing the fundamental economics of the web. Search GPT, developed by OpenAI, demonstrates this shift, relying less on traditional web crawling and more on curated, formalised partnerships with publishers to avoid legal issues over scraping. The new AI economy is increasingly creating a “two-tier” web, where access to valuable data and knowledge is controlled by those with the capital to enter into such partnerships, as noted by the Brookings Institution.
    • Global Accessibility
    • Cost and Accessibility Barriers:
      • The high energy costs associated with GENAI data centers are passed on to users, making these services less accessible to those in lower-income regions or with limited internet infrastructure.
      • The economic burden of GENAI adoption can further exacerbate existing digital divides, excluding those who cannot afford the necessary infrastructure or energy costs.
    • Alternative Approaches:
      • Researchers suggest that more specialized, less carbon-intensive models could be used for specific tasks, reducing energy consumption and making these tools more accessible to a broader audience.
      • The only other option is a phase transition to a new internet paradigm, with higher signal to noise, through aligned incentives and cryptographically assured end points.

Factors

Economic Models and Monetisation Strategies

  • The transition from free data access to paid partnerships reflects a deeper change in economic models within the digital landscape. OpenAI’s partnerships with major publishers for Search GPT indicate a movement towards a more closed, monetised web. These partnerships enable OpenAI to offer its AI-driven services legally and more sustainably, albeit at a higher operational cost, potentially offset by future advertising revenues or subscription models.

Regulatory and Market Dynamics

  • Regulatory environments are likely to influence the battle between traditional search engines and AI-driven alternatives. The European Union’s challenges in legislating competition within the search market underscore the dominance of established entities like Google. However, AI-driven search engines like Search GPT might benefit from regulatory shifts aiming to break monopolistic hold, thereby fostering competition and innovation.

User Behaviour and Adoption

  • A critical factor in the success of AI-driven search engines will be user adoption. While AI-enhanced search engines promise more accurate and contextually relevant results, there remains scepticism about whether users will transition from well-known traditional search engines like Google. Improvements in AI capabilities must demonstrate a substantial enhancement in user experience to effectively drive this behavioural shift.

Cloud Infrastructure and Competitive Dynamics

  • Microsoft’s role as a cloud infrastructure provider for both Bing and OpenAI highlights the strategic importance of cloud services in the AI and search engine landscape. The partnership dynamics between Microsoft and OpenAI, despite their competitive products, underscore the centrality of cloud infrastructure in supporting AI technologies.

Revenue Models and Financial Viability

  • The financial success of AI-driven search engines will hinge on their ability to generate revenue and offset operational costs. Traditional search engines like Google have perfected ad-based revenue models. AI-driven alternatives, however, must navigate establishing similarly effective monetisation strategies, potentially through a mix of subscription-based models and advertising revenue.

Algorithmic Capture and Decline of Human Exploration

  • We’ve become dependent on personalised, AI-driven feeds, narrowing our experience of the web. As Eli Pariser’s TED talk warned us years ago, AI-driven “filter bubbles” are trapping us in curated echo chambers, where discovery and serendipity are casualties of efficiency.
    • This reductionist approach to knowledge risks making us intellectually lazy. As MIT Technology Review notes, we’re being spoon-fed tailored, context-stripped answers that limit our engagement with the broader, more complex web.

Concerns Over AI’s Influence on Information

  • The rise of AI-generated content brings significant risks to the credibility of online information. The New York Times highlighted instances where AI chatbots have provided harmful or inaccurate information. This threatens to erode public trust in online content, as people increasingly struggle to discern between authentic human knowledge and AI-generated narratives, a concern echoed in Pew Research Center reports.
  • The World Economic Forum warned that AI’s role in content curation could deepen the spread of disinformation, leaving us with an increasingly “fake” internet, dominated by bots and automated systems, rather than genuine human interaction.
  • Algorithmic Bias: How Search Engine Results Can Be Skewed (talentsmart.co.in)

Rise of Bots Aligns With Dead Internet Theory

  • According to Imperva, bots made up 47.4% of internet traffic in 2022. This proliferation aligns with the once-dismissed Dead Internet Theory, suggesting that the web is increasingly driven by AI, bots, and algorithms rather than real human interaction. This isn’t just theory—evidence of widespread bot activity has been reported on platforms like Instagram and Twitter, where AI-driven manipulation of follower counts and traffic has become endemic.

Google search is broken. Google lied.

Computers breeding humans

  • It’s indesputible that dating app algorithms have been “matchmaking” humans against numerical criteria for a long time. It’s possible there’s are third generation “matched” users of the apps now, since computers were first applied to this in 1964.
    • “Cofounder of the RIZZ app here. We got inspired to build RIZZ a while back when we noticed the sheer volume of people asking for dating conversational help in the Tinder subreddit. Tons of people, both guys and girls, struggle with online dating. In fact, the statistics show that most are not successful with it. So how can they improve? Many go on Reddit asking for help. Or asking their friends for advice. Most simply can’t afford a dating coach. I believe AI can level the playing field, enabling people to enhance their flirting skills and boost confidence when interacting with new people. We launched just over a year ago and have learned so much up to this point. We still have so much to do to make our vision a reality.”
  • It seems not it’ll be bots talking to bots, under the guidance of algos, leading to dystopian babies.

Billionaires are monetising users

People have noticed

  • Science fiction author and all round thinker Cory Doctorow has been highlighting all this for a long time
  • Platform decay - Doctorow

https://pluralistic.net/2024/04/04/teach-me-how-to-shruggie/

https://www.benlandautaylor.com/p/the-ddos-attack-of-academic-bullshit

https://injuly.in/blog/darker-internet/index.html

https://www.wheresyoured.at/are-we-watching-the-internet-die/

Attention is too concentrated

  • We have become dependant on algorithmically generated personal feeds, we have lost the capacity to explore our digital society
  • These platforms are very vulnerable and exposed to manipulation, especially as the previous guardrails are removed for profit margins.
  • These technologies are reductionist, which feels like a shortcut, but is actually a simplification across the manifold of American English, and is likely making people intellectually lazy.
  • Increasingly there will be no other choice, trending toward Human vs AI.

Changes to information sources

  • Social media platforms, particularly Facebook and Instagram, are undergoing a significant shift by reducing the emphasis on political content. Meta, the parent company of these platforms, has introduced measures to minimize the presence of political content, such as the launch of the Threads app, which aims to de-emphasize news and politics (The Verge).
  • This move is part of a broader trend where social media giants are altering their algorithms to prioritize personal content over news and politics, potentially impacting public engagement and democracy (The New York Times).
  • Simultaneously, the influence of traditional or legacy media is experiencing a substantial decline. With the rise of social media and digital platforms, the reach and impact of legacy media have been radically reduced, leading to a transformation in the way information is disseminated and consumed.
    • Note that many are now signing deals with OpenAI etc
  • Twitter has also undergone significant changes and is perceived as a mere shadow of its former self. The platform has been criticized for various issues, including the spread of misinformation, harassment, and a decline in user engagement.
  • TikTok, a Chinese-owned social media platform, has been gaining prominence and is characterized by its opaque nature. As it ascends in popularity, concerns have been raised about its potential impact on shaping public opinion and its susceptibility to external influence.
  • Amidst these changes, there is a notable shift towards “the private internet,” including chat groups, Discord, fringe social media platforms, and Telegram groups. This trend has led to the fragmentation of views and discussions, raising concerns about the potential for echo chambers and polarization.

AI Advertising Trojan Horse Sites

The Dark Forest Thesis

  • The Dark Forest Theory of the Internet | by Yancey Strickler | Medium
  • The Expanding Dark Forest and Generative AI (maggieappleton.com)
  • In Liu Cixin’s sci-fi trilogy The Three Body Problem, the dark forest theory suggests that the universe is full of life, but it remains silent and hidden to survive, like animals in a dark forest at night when predators are out.
  • The internet is becoming like a dark forest:
    • In response to ads, tracking, trolling, hype, and predatory behaviors, people are retreating to safer, private spaces online
    • These “dark forests” of the internet are non-indexed, non-optimized, and non-gamified environments where depressurized conversation is possible
    • Examples include private channels, group chats, and newsletters sent to select groups
    • Podcasts are another “dark forest” where meaning is expressed through intonation and interaction, allowing for more forgiving communication compared to the internet at large

Bots Proliferate

Rise of Bots Aligns With Dead Internet Theory

Concerns Over AI’s Influence on Online Content

  • The collective impact of these developments creates a potentially more precarious environment for the manipulation of public opinion through artificial intelligence (AI). As political discourse becomes decentralized across various platforms and the emphasis on political content diminishes on major social media platforms, there is a growing risk of AI being exploited to manipulate public sentiment and political narratives.
  • Europol raised concerns about AI’s overwhelming presence in online content (reference).
  • Experts warn internet is saturated with bots, disinformation and biases that undermine perceptions of genuine interactions (reference).
  • The internet is becoming balkanised as large data aggregators like reddit close their APIs and prevent scraping. This is already having a meaningful effect on internet search.
  • ‘Background tokens’ is a term used on the latent space podcast to describe pre LLM unpolluted data. This legacy data is the last state of human discourse.
  • A fascinating unexpected second order impact of this is that future models may skew more right wing as those sites welcome the scraping.
  • AI and Leviathan: Part III by Samuel Hammond (secondbest.ca)

Mental health Social contract and jobs

  • Fraudulent studies are undermining the reliability of systematic reviews – a study of the prevalence of problematic images in preclinical studies of depression | bioRxiv Death of the Internet Deepfakes and fraudulent content

  • Jonathan Haidt Wants You to Take Away Your Kid’s Phone | The New Yorker

    • Jonathan Haidt, a social psychologist and the author of the book “The Anxious Generation: How the Great Rewiring of Childhood is Causing an Epidemic of Mental Illness”. The main points covered in the interview are:
      • Haidt argues that a whole generation has been damaged by growing up with unrestricted access to social media and an overprotected childhood, leading to a sharp increase in anxiety, depression, and self-harm among teenagers, especially girls, starting around 2012.
      • He attributes this to the rapid adoption of smartphones and social media platforms between 2010 and 2015, which radically changed childhood by replacing real-life interactions and play with excessive screen time and exposure to harmful online content.
      • Haidt presents evidence from correlational and experimental studies to support his claim that social media use causes mental health issues, while acknowledging the need for more research.
      • He argues that the benefits of social media are outweighed by its negative impact on child development, as it deprives children of essential real-life experiences, such as play, adventure, and healthy risk-taking.
      • Haidt advocates for changing social norms and implementing restrictions on social media use, such as banning phones in schools and limiting access to social media platforms for children under 16, rather than an outright ban on technology.

Tik Tok tick tock Death of the Internet Digital Society Surveillance Politics, Law, Privacy

Potential Green Shoots

Forcing function to trusted end points

  • AI will permanently put an end to the “post truth” era: this is the subject for another post, but clearly our prior epistemic standards no longer apply. the cost of creating arbitrary image or video content is effectively 0, so unsigned content will no longer by considered reliable (once people have learned to lately mistrust online content). to be considered reliable in the future, content will have to be signed, attested to, and timestamped (likely on a blockchain). - Carter

A handy simulator

Graph View

Backlinks