I recently rewatched two reviews of AI companion devices from Marques Brownlee (a.k.a. MKBHD), which both came out in April 2024 – the Humane AI Pin and the Rabbit R1:

Brownlee is a charismatic reviewer, and I won’t lie – there’s a certain amount of delicious schadenfreude seeing two products announced with so much unnecessary hubris faceplant so solidly.

But past all that, what stood out to me was the storytelling, and this is why I wanted to post these videos; the work feels good here, and deserves studying in how to talk about these things.

Brownlee manages to cut through the bullshit and candidly talk (but also show) what these devices should be all about. There are some nice editing choices; casually pulling up the phone and getting an answer while the AI Pin is still thinking is a pretty powerful moment. The reviews also come with higher-level reflections, and what stuck with me since the original watch years ago were the segments about how more and more often, devices like these – but not just these devices – come out of the gate profoundly unfinished.

The Pin and the Rabbit are also interesting to compare. They are both underbaked in some similar ways: the quality of the AI output, overfocusing on hardware at the expense of software, and the interaction systems that throw good money after bad and come out of the gate with strange complexity – AI Pin had the tricky-to-use laser projector menus, and Rabbit R1 managed to overgimmick a simple thing that is a rotary controller. On the other hand, one of the devices felt a lot more ambitious than the other. Also, both had fascinating differences in the packaging and their whole approach – it almost feels like there was a generational difference between the two endeavors.

(Humane Inc. has folded since. You can still buy the R1, which was since updated with a new UI, but it seems the company is pivoting to more generic-feeling software.)