Watching the recent stories about the AI devices go nuts over the last three weeks—first with the Humane AI Pin, then the Rabbit R1—I’ve noticed a certain undercurrent in the coverage that I think needs to be picked at while it’s still fresh.

Yes, it’s the idea that these devices should just be smartphone apps. I’ve heard many variations of this over the last week, and I think it’s best exemplified in this video by Dave2D, who basically tries to cut away the Teenage Engineering design influence of the Rabbit R1 with a bit of clever 3D rendering:

Pretty convincing, right? Well, sure. Dave Lee is one of YouTube’s best tech commentators for a reason—with enough of a foot in the consumer realm and the enthusiast realm to speak capably about both, he straddles the line between MKBHD and Linus Tech Tips perfectly. He is the exact person to make this argument capably. But I think that while his point is well-taken, I worry that the tech sector is going to take the wrong lessons from this particular moment, where our tech commentators are doing their best to cut through the thickest cloud of hype they’ve ever seen. (Not helping: The Rabbit R1 software can be sideloaded onto a traditional Android phone, as a standard-issue app. Not a good look!)

But I think that there is a real risk we’re running into right now where, by questioning why devices don’t look like the existing popular thing, we’re stifling the future growth of hardware development. I’m not in the same rooms as the folks at Rabbit or Humane, and honestly that isn’t really my zone. But I think if you asked them, they would probably tell you that they each started from the angle of trying to not build against the obvious thing, that this is a solved sector. This is what Apple famously did—and you probably remember the criticism they got when they tried to lead with design over functionality.

The challenge is, by solving every single problem in the universe with software, we are tying one hand behind our backs in terms of trying to find new solutions that benefit consumers. We are essentially betting that phones are the final thing. The smartphone is such a dominant piece of the conversation that it makes sense why we have let it shape our view of what hardware should be.

But … honestly, I’m really struggling with the way we force convention on every single new idea. Don’t get me wrong—there’s a definite value to the convenience of being able to use the device you already use to access a chatbot. But hardware is more challenging, and it requires a higher level of thinking to get right. Apple and Samsung deal with all the chipsets and platform standards and board manufacturers so you, the software developer, don’t have to.

If we’re going to look at smartphones as a solved problem—and given the iterative nature of the devices over the last decade, we kind of have to—it’s natural that we have to give the hardware designers a way to make their presence felt, by experimenting beyond the smartphone’s limited parameters.

Look, I’m going to be the first to tell you that Humane and Rabbit rushed out skunkworks projects onto an unsuspecting public, so they could improve the firmware in the wild and let MKBHD get raked over the hot coals for having a broad viewership. They are beta testing in public—something that, as the Apple Vision Pro has proven, is in vogue right now.

With that in mind, I don’t think “design this in software, so this can be an app” is a lazy critique, to be clear. And I don’t think we should be especially forgiving to companies who have raised tens or hundreds of millions of dollars largely on hype. But I think convention, too often, is the enemy of innovation, and we’ve been especially hard on two of the few attempts to shake up said convention in recent memory. (Dear Jakob Nielsen: I apologize for this break in UX orthodoxy.)

I am surprised this thing isn’t still being made. It was great!

Back in the day, I was a huge fan of the original Jawbone Up device, a bracelet-style wrist wrap-around that looked unique and awesome. It is probably one of my favorite hardware designs of the past 15 years, and it didn’t have a screen. One problem: One day, I lost the cap for it. When I called Jawbone to get a replacement for this little cap, they said they were out of stock, and gave me some tiny Airtag-like thing instead, which wasn’t anywhere as cool or as interesting.

See, what happened in the meantime was that a competitor, Fitbit, had emerged in the fitness tracker space and was making the devices more conventional and watch-like. Jawbone essentially decided to move away from their cool design in favor of something seemingly more commercially palatable. I’d argue, in the end, the decision to experiment less with product design kind of did in their company.

Now, most of our fitness trackers have screens and are basically watches—a loss of creativity over convention.

Is it a bad thing that companies are trying to force us out of our touch-friendly comfort zone through design? I say no. I think e-ink phones, for example, are a really exciting idea that simply hasn’t met its moment.

I encourage everyone to look a little further out than Dave2D’s critique of the Rabbit R1, and all the other critiques floating around right now, and see what’s on the horizon—not in terms of AI, but in terms of the next weird idea that comes packaged in a new hardware context.

Do we really want a world where every time someone tries to develop a new type of hardware, the first criticism is, “why isn’t this just an app?”