For the past month, the Buffer blog has been experimenting with new share options: Share to ChatGPT and Share to Claude. These buttons allow users to open articles in their preferred AI model with a prompt to summarize the content. The outcome is striking: interactions with these LLM share buttons have surpassed those from traditional social media platforms like LinkedIn, X, and Threads. This suggests a shift in how people consume information, moving from sharing and reading to seeking quick, digestible summaries. Marketers and content creators should consider a few key strategies. First, content needs to be structured for LLMs, which means using clear headings and strong metadata. Second, optimizing for the "moment of help" is essential. If users frequently turn to AI for content processing, making that experience seamless can enhance value. Finally, LLMs should be seen as channels for interpretation rather than just social sharing. This shift creates a new layer in how content is engaged with online. The conversation raises questions about the impact on writing, journalism, and research. While scientists may benefit from this format, the implications for creative writing, especially poetry, are less clear. There's a concern that focusing on LLMs could dilute the depth and nuance of human expression. Critics point out that relying on AI for summaries might indicate a lack of genuine engagement with the material, potentially alienating serious readers. Some express skepticism about the business value of sharing to LLMs. Concerns include losing direct engagement with the audience and whether those seeking summaries are the best customers. There’s a call for content creators to consider providing their own summaries or interactive features that keep users on-site, rather than relying on LLMs to relay their work. This highlights a tension between convenience and thoughtful engagement with content, suggesting a divide between those using AI to enhance understanding and those using it to bypass deeper thinking.