The lament that the 'dead internet theory' has become a reality as AI conversations become commonplace on social media and message boards.
Dead Internet Theory - Dmitry Kudryavtsev
https://kudmitry.com/articles/dead-internet-theory/
Kudryavtsev doesn't generally use social media, but he does read Hacker News, a platform for technology enthusiasts, entrepreneurs, and internet trolls. Hacker News helps him keep up with the latest technology news, rarely discusses politics, and doesn't have the social media aspect of individual accounts connecting with each other.
One day, while browsing Hacker News, Kudryavtsev came across someone sharing an open source project. 'It's great when people decide to work on a project and share it with the world,' he said. 'I think people underestimate the fear of actually shipping something and sharing it with the world.'
However, upon looking into the thread, many users pointed out that much of the project was generated by AI, questioning the validity of the project itself, with users criticizing it for its incomprehensible commit timeline and comments generated by a coding AI.
While Kudryavtsev himself doesn't mind if the code is written with AI, he believes it's fair to disclose the use of AI, especially in open source software. Kudryavtsev points out that large-scale language models, which are the foundation of AI, are merely probabilistic token generators, so while they're very good at most simple tasks, they can cause problems with some complex tasks.
The developers of the project responded to these criticisms by claiming that they were '100% written without the use of AI' and tried to reply to all the comments. As he watched the exchanges, Kudryavtsev realized that the developers' comments themselves were likely generated by AI.
For example, the
In addition, the comments were also based on evidence that they were generated by AI, including repeated use of the phrase 'you are absolutely right,' which Kudryavtsev had never seen before, and repeated use of AI-like endings such as 'let me know if you want to [do that thing] or [explore this other thing].'
'I sat there refreshing the page and saw that both the code and the comments pointed out the use of AI, despite the author claiming that he hadn't used any AI at all,' Kudryavtsev said. 'I honestly thought I was going crazy.' 'Is this person actually a real person? Are these people real? And then I realized: We've reached the 'dead internet.''
The Dead Internet theory holds that most online interactions are between bots, and that most content is machine-generated to sell products. When Kudryavtsev was young in the early 2000s, there were very few bots on the internet, and people could learn a lot from forums and anonymous people who gathered there. But now, Kudryavtsev laments, 'We no longer know what reality is.'
While Kudryavtsev doesn't use social media, he says he has seen real technology companies posting 'AI-generated photos of their offices and employees' on the business-focused social networking site LinkedIn. In fact,Facebook and YouTube have reportedly become overwhelmed with poorly generated AI videos (AI slop).
Survey results show that YouTube is being dominated by low-quality AI-generated videos, with 63 billion views and over 221 million subscribers - GIGAZINE
'I was honestly sad that day. I can only describe it as hopeless. AI is readily available to the general public, and it's possible to generate massive amounts of AI slop. People no longer need to write comments or code; they simply feed it into an AI agent, which will generate the next 'You're Absolutely Right' masterpiece,' Kudryavtsev said. 'What I'm seeing now makes me question whether technology will be useful in the future we're heading towards.'
Related Posts:
in AI, Posted by log1h_ik