Descriptions:
Ben, the solo founder of Polsia, joins the Latent Space podcast to discuss how his AI agent platform crossed $1 million in ARR within roughly one month of launch — one of the fastest-ever revenue ramps for an AI-native startup. The interview was recorded the same day the milestone was reached.
Polsia is an autonomous company-building agent: users describe a business idea, and Polsia handles product coding, competitive research, cold email outreach, social media posting, and iterative bug-fixing entirely on its own. Each night, a “CEO instance” evaluates the state of each user’s business, decides what to work on, executes tasks, and sends the user a morning summary email. Users can guide the system by replying to those emails or via a dashboard. At the time of recording, Polsia was managing over 1,000 active company instances simultaneously, generating more than 2,000 emails per day across its fleet.
The conversation covers key product decisions: why Ben stripped away extensive configuration options in favor of extreme simplicity (inspired by the observation that his 91-year-old father could use the system via email alone), how he chose Twitter as the first social media integration due to its permissive API for automated posting, and why roughly 50% of agent compute currently goes to engineering tasks because most companies are still in early product-build phases. Ben also discusses the core tension of the current AI era: the ability to build almost anything makes deciding what *not* to build the hardest and most important skill.
📺 Source: Latent Space · Published March 01, 2026
🏷️ Format: Interview
