Muse vs Instinct: how entry point and safety architecture shape retention
A product leader compares Meta's Muse and Instinct, arguing that entry format, security architecture and feature pacing decide which assistant keeps users.
Here’s a great example for product leaders on why every design and architecture decision in building a product really matters even in the AI era. Taste matters. Judgement matters. Understanding user behavior patterns matter. Product sensibility really matters. Let me share why I think Meta has rocked it the past few weeks. So Muse started with an app and Instinct started as a conversation thread on either WhatsApp or SMS. There was something magical about the conversation thread just being in WhatsApp but it was a single thread for all your conversations which if you created a separate thread in WhatsApp would make it super clunky. Neither decision was wrong but they were very different design choices. I started with Muse, but loved Instinct much more but then got a bit concerned about security and tools access with Instinct so went back to Muse and was pleasantly surprised with Muse on the safety and security architecture as well as how they were doing massive growth hacks to drive adoption. Why is that important. Because over time this app benefits from network effects. In any event, Muse just launched their WhatsApp connector. The amazing thing that Meta has done is really nailed feature roll out. So I got a notification. Clicked on it and instantly it activated my Muse on WhatsApp. And the design choice that they have made in the app is to show a separate thread for what I do in WhatsApp. These are thoughtful decisions made in the product that do add more complexity than Instinct but it was very elegantly implemented. So congrats @Muse on getting it right and all of a sudden in a couple of weeks if Instinct doesn’t move much much faster on some dimensions, what started as a better product than Muse will have a hard time catching up because of the mass consumerization that Meta has been able to do. Overall, just a super dynamic market and why it’s not just about the engineering and coding being fast but product judgement being well implemented. Lastly, what was impressive is how much of the Meta ad inventory they utilized to obtain early user traction. Now if you read this and are asking what either Instinct or Muse are, you are really missing out since it is the next ChatGPT moment with personal agents and you must download it. What Meta has done in the past three weeks is nothing short of remarkable at so many different levels.