We Are Trying to Break Your Heart

This chart is our opportunity at Rover.  

There are SO MANY AI agents out there, and they’re all focused on basically the same thing: Doing things for you that you in theory don’t want to do yourself. The common expression of this is “Only top executives can afford an assistant. Only rich families can afford a household chief of staff. Our AI agent does all of this for the low, low price of (all the data from you and your friends that our agent can harvest)* free!"

But I think that, while this is a real need among the kinds of people who fund tech companies, it’s actually a fairly small use case in the general population. Meta has the ability to convince a huge number of people this AI assistant is worth trying (me! Me! I’m one of those people!), and two weeks later only a quarter of them are still using Muse (also me!).

I think there are a couple of reasons for this:

1. This isn’t actually a huge problem for most people. Take airline reservations, the classic use case put forth for AI assistants. So the average American adult takes around one flight a year. This is not a big number! A small number of Americans (like VCs, for instance) take a lot more than that, and so to them this is a real problem. On the other hand, and this point is generalizable along a wide variety of things on the Internet, the last 20 years of Web development has been dedicated to solving this kind of problem! It’s actually quite easy for me to do a Kayak search and book a flight, or find something on Amazon to buy. How does an AI agent doing this for the average American adult once a year translate into a 10x better experience?

2. Ah, but maybe I’m looking at this wrong. It may be that no single task is all that compelling, but if you stack 5 or so things, the cumulative effect is that you have outsourced all of the small things that you don’t particularly like to do, freeing you up to, I dunno, play Minesweeper or something. 

Maybe that’s true, though if it is, the burden is on Meta to convince people to start stacking those things during that first two-week window.  But I don’t think so. I don’t think the point of life is to get rid of all friction and to reduce your daily cognitive load. To quote the ancients, “the to-do list is the way.”

I don’t want an assistant that thinks for me so I can zone out. I want a partner to help me think more deeply about things, to challenge me with ideas I haven’t considered, to hold my thinking and share it with others. I want something that matters, something that makes me engage with the world more, not less, something that makes me feel challenged, inspired, alive, even....heartbroken. The fact that the world is beautiful and can break your heart is what makes it worthwhile. To give your life over to a proxy is to be numb to all of that.

The problem a lot of agents seem trying to solve is “I want someone to do this for me.” So you get an agent that will, say, write your memo for you. What Rover does is to find the deeper thinking behind the last two years of your work on this topic, or show what others in your group have thought about this problem.

What stacks on Rover is your thinking over time, not your endless series of chores.  We don’t want to make it easy for you to think less. We want you to think better, and more.