Goliath Super Intelligence
IndustryOctober 27, 20262 min read

OpenAI launches “dots,” an always-on AI personal assistant

The new assistant builds on Codex technology, offers proactive research and continuous conversation, but raises privacy and scalability questions as it competes with Instinct and Meta’s Muse.

OpenAI introduced a new always-on AI assistant called dots at its September developer event, entering a market where startups such as Instinct and Meta’s Muse have already drawn attention. The product aims to keep working in the background without explicit prompts, offering a continuous conversational thread that users can rely on throughout the day.

Alexander Embiricos, head of OpenAI’s enterprise product group, described the emergence of similar assistants as a form of convergent evolution. He explained that framing the system as a single entity rather than multiple chats makes interaction feel more natural. “There’s kind of this convergent evolution happening,” he told TechCrunch Disrupt.

Dots builds on capabilities first seen in OpenAI’s Codex coding tool released earlier this year, but removes the need for users to guide each step. Embiricos recalled celebrating when his spouse began leaving her workstation on overnight so the assistant could continue tasks. He emphasized that the goal is to make the experience intuitive rather than to add new technical components.

A feature called proactive research lets the assistant periodically review its knowledge of a user’s preferences and recent conversations, then suggest actions it believes could add value. Embiricos summarized the idea as the system asking, “be more helpful to you?” This ongoing analysis is intended to surface reminders such as calendar conflicts before they become urgent.

OpenAI says dots inherits the same privacy settings as ChatGPT, allowing users to disable model training and to limit which local resources the assistant may access. However, Embiricos admitted that current controls lack fine-grained detail and that the app sometimes refuses read-only cloud permissions. “We need super granular controls,” he noted, highlighting a tension between functionality and data protection.

Scaling such agents raises questions about integration with business systems and payment infrastructures. Embiricos mentioned that OpenAI is engaging with payment processors but is cautious about rushing new capabilities. “We’re not always moving the fastest in terms of shipping the next ability,” he said, indicating a collaborative approach with the broader ecosystem to avoid overloading existing services.

At present, dots can manage email and calendar data but does not have permission to initiate financial transactions, a limitation that reflects both technical caution and user trust concerns. While the assistant can push notifications about upcoming events, users have reported that alerts sometimes arrive only minutes before the deadline, suggesting the feature is still maturing.

Sources

  1. How OpenAI turned its coding tools into dots, an always-on personal assistant TechCrunch

More reports

Industry · November 9, 2026 · 1 min

Apple Accelerates Siri Development to Compete with Emerging AI Agents

Apple aims to give Siri deeper app integration and autonomous task handling, targeting iOS 27.3 for internal controls and a spring rollout for third-party services.

International · November 9, 2026 · 2 min

Music Industry Body Warns Australia Against Diluting Copyright for AI

The International Federation of the Phonographic Industry cautioned that proposals to loosen Australian copyright rules for AI training could undermine global creativity and set a risky precedent.

Industry · November 9, 2026 · 2 min

Nvidia CEO Jensen Huang Faces 50-50 Survival Odds Amid Spy Allegations

Nvidia’s board opened a probe after CEO Jensen Huang admitted an affair with a suspected Chinese spy, and Yale professor Jeffrey Sonnenfeld rates his odds of remaining in charge at about fifty-fifty.