The Humane AI Pin never had a chance

Ten days from now, the Humane AI Pin will be able to tell you how much battery it has left, and essentially nothing else. To be fair, though, it couldn’t do that much before. And it doesn’t matter anyway, because you almost certainly didn’t buy one. But if you did, that’s the bad news: Humane is shutting down the AI Pin — almost exactly a year after it first started shipping the little chest-mounted device — and has sold some of its remnant technology to HP.

The details here are brutal for everyone involved. The $116 million HP is paying pales next to the $230 million the company raised since it was founded in 2018. (There was a rumor last spring that Humane was trying to sell to HP for somewhere near $1 billion.) The company’s founders, Bethany Bongiorno and Imran Chaudhri, have now gone from celebrated Apple product makers to… working on AI stuff for HP’s printers and conference room gadgets. Pin owners will be stuck with a $699 paperweight, since Humane is neither offering refunds outside the normal 90-day purchase window nor appears to have a plan to open-source or otherwise make available any of its software. It’s all just over.

Humane will go down in the annals of tech history as one of the most spectacular gadget failures of all time, up there with Quibi, the Fire Phone, and the Apple Newton. The company never seemed to put a foot right. It labored in secret for years, building relentless hype, before first showing off the device in an overwrought, thoroughly confusing TED talk in which Chaudhri mysteriously got the Pin to translate a sentence to French without ever asking it to, and answered a phone call by doing… nothing. Knowing what we know now, it seems virtually impossible that the demo wasn’t mostly or entirely a fake.

Humane was hellbent on convincing everyone that the Pin was the future, and that the future was here right now. It called that TED talk “The Disappearing Computer,” and pitched its device as the first true AI gadget — and maybe the successor to the smartphone. Humane made a big deal out of its announcement of the product’s name, and had its device declared one of Time’s “Best Inventions of 2023” before the thing ever actually shipped. (Surely it’s a coincidence that Salesforce CEO Marc Benioff both owns Time and invested in Humane, right?) The company delayed the Pin’s launch, before ignoring panicked warnings from early testers who said the device was woefully unfinished and shipping the thing anyway.

Even if the Pin had been, let’s say, pretty good, it would have felt like a disappointment after all that build-up. But the Pin was junk. I gave it a 4 / 10 in my review last year, and in retrospect I was too generous. There were certainly a few impressive things about it: it was small and sturdy and genuinely nice to hold and look at. As any good student of Apple design knows, though, design is how you use it. And using the Pin was a study in disappointment. The Pin couldn’t do much — it shipped without features as basic as a timer or an alarm, and was missing a number of the features Chaudhri demoed during his TED Talk. What it could do, it did sporadically and mostly poorly. The battery didn’t last very long, not that you could ever test it — the Pin overheated so quickly even in moderate use that it shut itself down long before it died.

Oh, and did I mention it was SIX HUNDRED AND NINETY NINE DOLLARS? Plus a mandatory $24 monthly charge for cell service? Forget pretty good, the Pin would have had to be great to be worth it. At least Rabbit had the good sense to only charge $200 for the similarly crummy R1 — and to make it orange and fun to fiddle with.

To its credit, Humane eventually updated the Pin to add some missing software features and fix a bunch of what was broken, and the Pin will certainly be a better device on its last day than on its first. But it never mattered. Pretty quickly, Pin returns started to outweigh Pin sales. Humane had to recall the Pin’s charging case, which was an overheating risk. The company had always planned to make its operating system, called CosmOS, into a platform for other developers and hardware manufacturers to use, but waited too long and found no takers.

Some of this, a lot of it even, was self-inflicted. Humane just tried to do too much, and reinvent too many things, that it would have had to execute at unheard-of levels for a first-time startup to pull it off. But Humane also made a huge bet on AI technology that, while it continues to improve, also continues to be woefully inadequate for most things. We’re only now seeing companies like OpenAI begin to slowly give their models these so-called “agentic” capabilities, so they can go book you Airbnbs or remind you to pick up flowers. When all that works perfectly, when AI models are integrated seamlessly with every app and service and real-time detail of our lives, maybe a totally interface-less device like the Pin will be useful. But if that’s ever going to happen, it’s not going to be soon.

What’s most damning in the Humane story, though, is how much the company was right about. The time was ripe for a first mass adoption of AI — but ChatGPT and Claude figured out the user interface people actually want right now. There are opportunities for new kinds of gadgets — but they’re going to look more like glasses than brooches. People do increasingly want faster, easier, less screen-based ways to interact with technology — but the voice assistant on their phone and in their ears is a lot more useful than a laser projector on their hand. Humane saw the big picture mostly correctly, and it utterly botched the details.

Since Humane’s failed launch, I’ve heard from multiple startup founders who were considering launching gadgets with AI features that decided to change course after seeing what happened to Humane. The company promised to kick off the AI gadget revolution, but if anything it set it back.