Why AI researchers should publish on LessWrong: insights from Logan Riggs

Logan Riggs, a researcher at CBAI, is collecting experiences on publishing research via LessWrong or Twitter/X to advise a research fellowship program. In his own account, he highlights three key benefits. First, blog posts force crystal-clear explanations, which then improve in-person communication, reuse in presentations/papers, and sharpen technical writing skills. Second, they attract the right people: he was contacted by three startups (though not his interest) and proactively reached out to 10+ researchers after reading their posts. This led to co-authoring an SAE paper with Lee Sharkey and collaborating with Thomas Doom on tensor networks—work that is both performant and enables interpretability. Third, he finds feedback on the posts themselves mediocre; the real value comes from 30+ minute calls or in-person chats sparked by the posts.

Riggs is particularly interested in hearing from others with different experiences, including those on Twitter, and wants to compile tips for fellows. The post has minimal comments so far—only Elizabeth asking if he's focused solely on AI. The overarching message: publishing on LW or X is a powerful networking and communication tool, even if the feedback isn't always deep. It's a pragmatic strategy for early-career alignment researchers to build connections and refine their thinking.