OpenAI Dots vs. Meta Muse: Two Different Visions for the Future of Personal AI

OpenAI has unveiled "Dots", a new line of persistent AI agents designed to handle tasks over time without needing constant prompts. These agents can operate across multiple applications, perform ongoing work, and delegate actions with increasing autonomy. OpenAI's Dots are currently being developed using the GPT-6 Astra model, and they will have their own cloud computer and browser.

They will be able to connect with over 4,000 applications through OpenAI's plugin ecosystem and communicate with users through various platforms, including ChatGPT, Slack, and Microsoft Teams.

Meta, on the other hand, has introduced "Muse", an AI agent that can take on a broader range of personal tasks such as travel arrangements, shopping, and planning. Like Dots, Muse operates through a dedicated cloud-based virtual machine with its own browser and can continue tasks even after the user closes the app. Users can access Muse through Meta's dedicated application or WhatsApp.

Meta's Muse also supports payments through Stripe's Link, allowing transactions to be carried out without exposing a user's primary card number directly to the agent or merchant.

However, while both Dots and Muse share similar technical foundations, each company has emphasized different aspects of the personal AI experience. OpenAI is focusing on ongoing knowledge work with Dots, while Meta is presenting Muse as a versatile agent capable of handling a wide range of personal errands, purchases, travel, and planning. The key difference lies in the areas of focus: Dots seem to be more oriented towards work-related tasks, while Muse appears to cater to a broader mix of personal tasks.

Both AI agents, nevertheless, aim to reduce the back-and-forth typically associated with generative AI tools. Instead of requiring a new prompt for every step, these persistent agents can handle ongoing assignments and return when work has progressed or a decision is needed. This shift in responsibility from the user to the AI could change the skills required from workers, as they would need to define outcomes, set boundaries, review results, and decide when human judgment is still necessary.

OpenAI and Meta have also implemented approval and monitoring controls around their agents to ensure user safety and maintain necessary human oversight.