OpenAI Launches 'dots': An Always-On AI Assistant Evolved from Coding Tools

OpenAI has officially unveiled 'dots,' a persistent, always-on AI personal assistant that marks a significant shift in how deep learning models are integrated into daily workflows. Announced at the company's developer day in late September, dots represents the latest entry in a competitive landscape dominated by startups like Instinct and Meta's Muse. The core innovation lies in its 'always-on' nature; unlike traditional chatbots that wait for prompts, dots operates in the background, continuously processing information to anticipate user needs. Alexander Embiricos, who leads OpenAI's enterprise product team, explained that the goal was to frame the AI not as a collection of separate conversations, but as a single, intuitive entity that feels like a proactive colleague. The technical foundation of dots is rooted in OpenAI's earlier coding tools, specifically the Codex app released earlier this year. While Codex required significant manual guidance, dots repackages this agentic capability into a more accessible, consumer-friendly interface. The assistant is currently available to Pro-tier subscribers, with plans starting at $100 per month. Once connected to a user's calendar, email, and computer tools, the agent autonomously works to resolve tasks. Embiricos noted that the product's success relies less on new underlying technology and more on the user experience, aiming to make complex AI interactions feel natural and intuitive. In practical testing, dots demonstrates a clear ability to think ahead, such as flagging calendar conflicts and upcoming deadlines. However, the timing of these alerts is still being refined. Embiricos shared an anecdote about an early version alerting him to a conflict just three minutes before it occurred, highlighting the current limitations in real-time responsiveness. Despite these early hiccups, the proactive approach has shown promise in high-stakes environments. One engineer at a drone delivery startup reported that dots detected a system outage, analyzed the issue, and provided resolution strategies before the human operator even noticed the problem. A critical aspect of dots is its 'proactive research' feature, which periodically reviews user data and accessed resources to suggest new ways to be helpful. This functionality raises significant privacy questions, particularly regarding the volume of personal data exposed to the software. While OpenAI states that dots shares the same privacy controls as ChatGPT—allowing users to disable model training and manage resource access—critics argue that these controls are not granular enough. Embiricos acknowledged the need for more detailed privacy settings, emphasizing that the model's discretion and intelligence are key to balancing utility with user trust.