Humane's Ai Pin and Rabbit's R1 arrived with the same promise: escape the smartphone trap. Carry AI differently. Early reviews were kinder to the ambition than the daily experience, and that gap is the lesson.

Real tension, early execution

Both devices pointed at a real tension. Phones are distraction engines. Maybe ambient, voice-first, camera-assisted AI could delegate the busywork. Maybe a dedicated gadget would mean fewer thumb marathons.

Then people lived with them. Missed queries. Slow responses. Awkward social moments wearing a camera. Features that worked in controlled demos but struggled on a noisy sidewalk. Reviewers who wanted the idea to work, and still put the devices in drawers after a week. The story was post-phone. Daily behavior stayed with the phone, because the phone already works, and everyone around you understands it.

Humane pushed toward ambient, projector-based interaction; Rabbit pushed toward a teachable action model on dedicated hardware. Different bets, same bar: the phone is already a capable AI surface with a mature app ecosystem. Displacing it means clearing reliability, social acceptance, and daily habit at the same time.

Early is different from ready

Early can mean the stack still needs time: battery, connectivity, model reliability, developer ecosystem. It can also mean the job still needs sharpening. "General AI companion" is vague. "Translate this menu while I hold it still" is concrete. "Confirm my meeting and draft a follow-up in my tone" works when it's right 95% of the time.

Hardware punishes vagueness. Software can ship a beta and patch tomorrow. A $699 pin with spotty LTE teaches the market a story you don't control unless the daily loop already works.

What I'd take as a founder

First, separate keynote from habit. Would you use your product every day without a charging ritual you'd otherwise avoid? If daily use isn't there yet, vision can fund the work; retail needs utility.

Second, nail one loop before ten. Rabbit's teach mode was interesting because it hinted at personalization. Personalization on top of baseline reliability is homework; personalization after reliability is a habit.

Third, respect social design. Wearables are worn in public. If bystanders feel recorded, you lose, regardless of model size. Apple spent years training norms; new categories have to earn the same trust.

The product lesson

Humane and Rabbit were early bets marketed as arrivals. AI hardware will keep coming. The winners will likely look less like sci-fi pins and more like focused tools: a specific profession, a specific environment, a specific gap in what phones do well, solved almost boringly well.

Before you ship hardware, ask: is the daily job sharp enough, and is the first loop reliable enough, for someone to keep paying after the keynote fades? Honest answers save years.

Originally published on Product, AI & Business on LinkedIn.