Where I End and You Begin

Posted on June 17, 2024. Filed under: Uncategorized |

The interface between users and autonomy is getting interesting. I’ve had a few experience recently that have made it clear that thoughtful UX in AI applications is going to center around the question of “where I as a user end and autonomy begins.” The concept of a copilot acknowledges this interface, but how we collaborate with autonomy, and the friction associated with that dialogue is very expressive surface area for great design and magical experiences. I was in California a month or two ago and my friend Jeff Weinstein let me drive his Tesla with the new full self driving enabled. It was obviously very cool, but not in the same way as the Waymo self driving taxi experience I had the next day. I’d argue it was more magical, despite needing to sit in the front seat…so why? The thing Tesla got really right was the handoff between FSD and me taking control of the wheel. There’s no button to press, no need to change states, if the Tesla is driving and I want to provide input I just start guiding the wheel and I’m seamlessly back in control. It’s subtle, but even my Rivian is a pretty jarring experience moving from Level 2 cruise control (stay in line + speed control) to human guided driving. Tesla made it frictionless for me to define where I want to end and where I want automation to begin and it was delightful.

Fast forward a week or two and I started using what has become one of my absolute favorite AI applications: Granola. I had tried other meeting capture / summarization type products before, and nothing came close to clearing the bar. So why did Granola stick so hard for me? The answer, obviously, lies in the interface between myself and autonomy. During a meeting, the notes that I take are a form of thinking. I pick out what matters. I jot down something really only legible or useful to me, and then I rely on my memory of the conversation or meeting to fill in the rest of the context around my note. So a typical note I might take during a meeting would be:

Stanford undergrad
Knows John Lilly
38% D30 retention
designing for autonomy
big picture = assistant
5 people
raising $4M, Patrick invested

Something like that…super legible to me right after the meeting, but decaying in value as I forget the rest of the context inbetween these points.

Granola’s brilliance is that they acknowledge the truth which is that I don’t want to outsource my thinking to AI. A summary of the meeting, without my input, isn’t what I value or want to reference. Granola gives me a note pad, to jot down my shorthand, but then utilizes the transcript from the meeting to “enhance” my notes with the surrounding context from the meeting. The output is fully flushed out, highly legible notes, where I’ve defined the focal points (where I end) and it has filled in the blanks (where it begins). It’s a very subtle interplay between the user and autonomy that once again strikes this magical balance similar to Tesla.

Anyway, these are just two examples that I think highlight a battleground where many more AI applications will distinguish themselves from the field. Each use case that contemplates autonomy calls for a different interaction paradigm between a user and automation, and those that elegantly enable the baton of control to be passed back and forth seamlessly within a given use case will thrive. If you are a product designer and this line of thinking resonates, I’d love to try your product, join your testflight, maybe invest… I’m jordan@pacecapital.com

*Disclosure: I’m not an investor in Granola (thought I wish I was), I am an investor in Tesla