I tried Muse, Meta's new AI agent (meet Slime 🦖)
The speaker reviews Muse, Meta's new AI agent, highlighting its ability to perform browser-based tasks, connect to multiple data sources, and help users achieve personal goals through an approachable interface. Key features include connectors to email/calendar/health data, idea suggestions, artifact creation, and customizable avatars.
Summary
The speaker introduces Muse, Meta's personal AI agent designed to help users accomplish tasks across wellness, shopping, and family management. Unlike traditional chatbots, Muse has computer control capabilities, allowing it to perform any action a user could do in a browser—such as shopping, research, and form completion—which the speaker notes is particularly useful for parents.
The speaker outlines four core distinguishing features of Muse. First, connectors enable Muse to integrate with email, calendar, social media (Facebook and Instagram), health data, financial data, and other daily-use software programs. Second, Muse provides suggested ideas upon login, offering categories like shopping, selling on Facebook Marketplace, home care, health management, and entertainment. Third, Muse can create artifacts—documents, websites, images, and podcasts—such as custom party invites, brainstorming documents, or personalized news podcasts. Fourth, Muse includes a goal-tracking feature that helps users monitor personal objectives and receive regular progress reminders; the speaker shares personal examples including sleep tracking, concert ticket acquisition, and water consumption goals.
The speaker addresses data privacy concerns by recommending users review the terms of service, noting they personally reviewed and felt comfortable with Meta's data practices given the productivity benefits. The speaker emphasizes Muse's accessibility as an AI agent option for newcomers. Finally, the speaker highlights the customizable avatar feature, demonstrating with their own Muse named Slime, a dinosaur character that visually works alongside the user throughout the day.
Key Insights
- Muse differentiates itself from chatbots by having computer control capabilities, enabling it to perform any action in a browser that a user could do manually, including shopping, research, and form completion.
- Muse connects to multiple personal data ecosystems including email, calendar, social media, health data, and finance data to help users accomplish more across different life domains.
- Muse can generate artifacts such as custom party invites, brainstorming documents, and personalized local news podcasts that users can consume on-demand.
- The speaker reviewed Meta's terms of service and determined they were comfortable sharing personal data with Muse based on the personal productivity benefits offered.
- Muse features a customizable avatar system where users can select names and avatar designs that visually accompany them during work throughout the day.
Topics
Transcript
[0:00] I have a favorite new AI agent. He's chunky, he's fluffy, and he helps me get done. I'm talking about Muse, the new AI agent from Meta. If you haven't used Muse yet, it's a personal AI agent that's there to help you get stuff done. I use it for wellness, shopping, and getting things done with a family. What makes Muse different from a chatbot or other AI agents you might use? Well, your Muse has a computer, so it can do anything on a browser, on a computer that you could do. That means it can shop, research, and fill out forms for you. If you're a parent, you know this is super useful. The other thing…
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
Jev analyzed 1,700 PRs for 9 cents
A developer used AI (Gemini) to analyze 1,700 pull requests in 2 minutes for just 9 cents, extracting work allocation data across initiatives. This demonstrates how AI can help CTOs and CEOs quantify what percentage of engineering effort goes toward different products or projects.
I’m using Jev more than Opus 5.5 or GPT-6. Here’s why.
The speaker demonstrates why they're using Jev, a fast and inexpensive decision-making model from Type-Safe AI, more than other recent models like Opus 5.5 and GPT-6. Jev specializes in classification, clustering, and real-time decision-making tasks at a fraction of the cost of traditional LLMs, enabling complex data analysis and product features that would have been prohibitively expensive before.
Warp agents open PRs to fix the factory itself
Programming agents can autonomously improve factory systems by analyzing failed launches and proposing specific updates to agent definitions. A self-improvement loop enables observer agents to detect failures and generate evidence-based modifications that prevent recurring issues, such as changing specific steps in factory agent procedures.
Humans are still the bottleneck in Warp’s AI factory
Warp discusses how human code review has become the main bottleneck in their AI-assisted software development process, with a 3.5-hour delay from PR to first human review compared to 35 minutes from launch to PR. They're evolving their workflow to reduce human dependency by allowing requesters to review agent-generated code themselves, and plan to eventually skip review entirely for low-risk tasks by treating code review as a risk management exercise.
I Quit Claude Because It Was Annoying
The speaker explains why they stopped using Claude, citing frustrations with its tendency to produce nonsensical output and communicate in an unnatural, non-human manner. They mention considering a switch to Opus 5.5 but remain uncertain about fully migrating their work tasks.