LessWrong vs. TikTok: 60 days of daily videos reveal attention hacks
After 60 days of daily TikToks, one creator learned how to capture attention in a non-rational space.
Taylor G. Lunt, a LessWrong user, spent 60 days posting daily TikTok videos and documented the experience in a viral essay titled "LessWrong vs. TikTok: Tips for capturing attention in a non-rational space." The experiment produced striking numbers: the highest-viewed video reached 78.6K views, while average views jumped from 563 across the first 10 videos to 2,071 across the last 10. Lunt notes that TikTok's algorithm gives creators near-immediate feedback — a brand-new account can go viral with millions of views, but even established creators get cut off at around 300 "test views" if the content fails to engage.
Lunt argues that TikTok is fundamentally different from LessWrong: where LessWrong audiences engage in reasoned argument, TikTok viewers are often in a "dissociative trance," reacting to social proof rather than logic. The essay highlights the low quality of comments, using examples from AI researcher Robert Miles' thoughtful videos, such as users claiming LLMs cannot be conscious "by design" without any attempt to persuade. Lunt also describes a personal evolution: learning to stop mumbling, improving lighting and expression, and eventually pulling back from persuasive political content due to toxicity. The piece ends with practical tips for anyone looking to reach a mass audience on TikTok, especially for AI safety communication.
- Average views jumped from 563 (first 10 videos) to 2,071 (last 10 videos) over 60 days of daily posting
- TikTok algorithm gives instant feedback — new accounts can go viral, while failed videos get cut off at ~300 test views
- Lunt warns that persuasive AI safety content attracts toxic, non-reasoning comments, requiring a thick skin
Why It Matters
AI safety communicators can reach enormous audiences on TikTok, but must adapt to a non-rational, attention-driven medium.