This chapter introduces Beau, a public robot that uploads daily text posts through a pseudo-social media platform. I describe the design and implementation of a series of research prototypes that explore the experience of interacting with a robot through social media posts and how such interactions shape perceptions and experiences of the robot. The prototypes consist of three components: Beau, HeyBeau, and IamBeau. Beau is a public robot that measures visitors’ body temperature at a building entrance and shares text posts written from the robot’s perspective with users. Because the study took place during the COVID-19 pandemic, Beau was used very frequently. HeyBeau is a curation system designed to support valid and effective user studies. It allows researchers to select appropriate topics from the data collected by Beau and to review automatically generated posts for contextual errors or ethical concerns before they are shown to users. Finally, IamBeau is a social media-style Web application that enables people to read these automatically generated posts. These three elements were carefully designed with attention to appearance, a realistic research context, and data collection practices that respect privacy. During a 16-day field deployment, 12 participants read a total of 32 robot perspective posts through IamBeau. Changes were tracked through four administrations of the RoSAS questionnaire and post-study semi-structured interviews. Quantitative analysis showed that robot perspective text significantly increased perceived Warmth among the robot’s social attributes, while no significant changes were found for Competence or Discomfort. Qualitative analysis revealed that these changes were linked to the way intelligence, consciousness, and emotion were expressed through text, how accumulated posts came to constitute the robot’s identity, and how users began to accept the robot as an acquaintance or a potential friend. At the same time, the study identified mismatches that arise when expectations generated by text are not fully met through physical interaction, tensions between the efficiency of automatic generation and the demands of authenticity and ethical filtering, and the need for deliberate data and model design to shape social perception intentionally. This chapter suggests that when a thing generates and shares its own narrative through an LLM, relationships between humans and things can become more sociable.