Legible EventsYour Agent Is Only as Good as Its DataNov 12
← Newsletter

Muse, Instinct, and Dots: The Rise of the Personal Assistant

Hi All,

The big topic here at Legible this week has been the rise of personal AI assistants aimed at everyone, not just knowledge workers. What’s notable about these products is their ambition. This is an attempt to equip every human on the planet who has an internet connection with a personal AI Assistant.

Meta shipped Muse, which reached 5M downloads in just 22 days (compared to 56 days for ChatGPT). OpenAI shipped Dots yesterday. And an independent company called Instinct has been growing 10% day over day while still in private beta and raised well over $1B in funding. All of them should probably give a nod to OpenClaw as an inspiration.

We think (and write) a lot about mental models for working with AI, and the personalized assistant that knows your world represents a Third Way, after the tools-based approach of Claude and ChatGPT, and the AI Teammate approach of Skydive and others.

Personal Assistants are great for booking a dinner reservation, ordering flowers, filling out a form, and doing some quick research. That is what Instinct, Muse, and Dots are all going after. I write about my experiences with Instinct here: The rise of the personal assistant for everybody else.

Before you run out and try one of these, I would recommend that you read Brandon’s post on the security implications of these products. He includes things you can test yourself. OpenAI Dots and what it means for security.

The other thing we have been chewing on this week is what companies choose to do with the AI productivity dividend. When AI saves your team five hours a week, the default behavior is to fill those five hours with more work. I have met CEOs who are thinking about it differently, choosing to share that dividend with their employees so they can live more balanced lives. I wrote about this tension here: Your AI manifesto needs to reflect your values.

Finally, if you missed our session on AI Policy and Acceptable Use, you can watch the recording here.

Lawrence

Originally sent by email. View on Beehiiv ↗