When Bobak Tavangar explains why a small company might have a chance in augmented reality, he reaches for a Hobbit. There are armies marching across the landscape, equipped for a grand confrontation. Then there is someone taking another path up the mountain. In his telling, the founders of Brilliant Labs collectively occupy the Frodo-and-Samwise part of the expedition. The analogy eventually raises the question of who gets to be Gollum. Hardware strategy can accommodate a little mischief.
It is a revealing choice of story. Tavangar is attempting to build AI eyewear in a field that attracts companies with enormous resources. His answer has been to narrow the task: make something light enough to wear, give it a useful relationship with the world around it, and let other people build on it. There is plenty of ambition here. Much of the work involves deciding how little machinery that ambition requires.
An economist walks into a hardware company
His route into the business took a few turns. Born in Philadelphia, Tavangar studied in Washington, DC, and later at Cambridge. His academic background was economics and management. Before the eyewear came an interest in how systems work, how information connects, and how technology might increase what people can do. A circuit diagram was hardly the only possible destination.
He calls himself a generalist. The description fits someone whose education continued across institutions, countries, and companies. At Microsoft, an internship offered an early look inside a technology business. Later, he chose entrepreneurship over an anticipated PhD place. He had an adviser lined up and a next academic chapter within reach. Instead, he started a company. A career can turn on the acceptance letter one declines.
For Tavangar, learning has an element of voluntary discomfort. He has described arriving at the swimming pool and needing to jump into the deep end. Risk makes the learning interesting to him; a steep curve is part of the attraction. That is a useful piece of autobiography for a founder entering hardware, where even a promising idea must eventually negotiate with a physical object.
Asia became central to that education. Learning Chinese gave him a way into the practical business of how things get made. He talks about manufacturing with the interest another founder might reserve for strategy. Someone has to understand the roots before enjoying the branches. His subsequent career would give him ample opportunity to inspect both.
A lesson in making less
Before Brilliant Labs, Tavangar built Infograph in Hong Kong. Its work brought together graph search, computer vision, and geolocation. Experiments with augmented reality exposed a particular inconvenience: a phone could reveal information about the world through its camera and screen, but holding it up was an awkward way to go through that world. Eventually, the company sold its technology, and Tavangar went to Apple.
His Apple work placed him in Shanghai and inside operations. Here was the large-scale relationship between software and the supply chain: production, information, and people had to fit together. The experience offered an education in what it takes for a design to become an object. In 2019, he left to start Brilliant Labs.
The appeal of Apple extended beyond its finished products. Tavangar has spoken about the company teaching its culture deliberately, including a course called “What Makes Apple Apple?” Focus and simplification were principles people were expected to learn. A company could build shared judgment alongside its hardware. For someone preparing to assemble another team, that was an education worth taking away.
“The idea is that it leaves you hands free.”
Bobak Tavangar, describing Frame
There is a pleasing difficulty in applying that lesson to a new category. Before a product exists, it can do almost anything in conversation. Once it must sit on a face, every capability starts submitting an invoice. More processing needs power. Power needs a battery. The battery needs somewhere to live. The wearer has a vote in all of this, even without understanding a single component.
The frame has to leave the laboratory
Brilliant Labs’ founding group included designer Benjamin Heald and engineer Raj Nakarja. Its first product, Monocle, clipped onto existing eyewear. The name was unusually literal for a technology product. It also made the proposition visible: begin with a small device that people could experiment with, rather than requiring an entirely new apparatus around the head.
Frame followed in 2024, pairing round glasses with an AI assistant called Noa. The assistant brought a conversational element to the device. Information could arrive in a see-through display while the wearer kept looking outward. The name Frame was another reminder that an everyday object had to carry this unfamiliar computer.
The circular rims had their own cast of characters. Tavangar has discussed references ranging from John Lennon and Steve Jobs to Groucho Marx. That last name lends a welcome levity to the usual technology ancestry. The frame belonged to a recognizable cultural history before it contained electronics. Borrowing from that history gave the object something to say before anyone asked what processor was inside.
He treats appearance as a substantial constraint. Glasses are part of a wardrobe, highly visible and close to the identity of the person wearing them. You can take another gadget from a pocket for a brief task. Eyewear asks for a more public commitment. That makes thinness and weight part of the product’s purpose, rather than details to address once the electronics work.
There is a lighter example of this preoccupation: will.i.am tried Frame, and Tavangar described the musician as looking professorial in it. He also conceded that Frame looked better on will.i.am than the musician’s glasses looked on him. A founder can argue about artificial intelligence and still recognize when someone else has won the fitting.
Leave the door open
Tavangar’s other recurring choice is openness. He wants people outside Brilliant Labs to inspect and modify the platform. His argument includes a straightforward recognition that developers have ideas his own team will not anticipate. Put a new medium within reach, and the eventual uses may be more varied than the demonstrations prepared for launch.
The early customer base for Frame reflected that approach. In an October 2024 interview, he described it as largely developers, with applications shaped by their particular fields. He also acknowledged that mass adoption would take time. That distinction matters. A platform for experimentation and a device that fits effortlessly into millions of routines have different work ahead of them.
He has discussed the balance between openness and a coherent product experience, too. Releasing a platform does not remove the need for design judgment. People still need tools they can understand and a device they can use. The invitation to build becomes more convincing when there is a workable starting point.
The philosophical argument is more personal. A device that sees and hears near its wearer occupies an intimate position. Tavangar links the ability to examine and change technology to trust in it. The business question and the human question meet at the same object: how much control should someone have over a computer they carry through their day?
The next pair carries the previous lessons
Halo, announced in July 2025, extended the project toward conversational assistance and memory. Brilliant Labs described a system intended to learn from what a wearer sees, hears, and says. The ambition moved beyond answering a question in isolation toward retaining useful context. Remembering where something was left is an ordinary enough problem to give the idea a recognizable shape.
The company’s own account of developing Halo was candid about the preceding work. Frame had taught hard lessons in development and manufacturing. Before the next device, Brilliant Labs revisited its culture, team, and supply chain. A product generation can conceal those revisions behind a new name. Tavangar’s company chose to describe the organizational work as part of the story.
Wearability + useful context + room to build
The hardware decisions followed that intent. Brilliant Labs described avoiding WiFi, bulky optics, and large image sensors oriented toward recording, while adding another microphone and speakers around a low-power processor. Each choice reflected a particular job for the glasses. Even a component left out becomes part of the design when it saves the space or energy another function needs.
By May 2026, Tavangar was describing a processing arrangement spread across glasses, smartphone, and cloud, with speech and image inference running locally on the phone. He also said Foxconn was mass-producing Halo. Those were his accounts of the system and its manufacturing, rather than independent performance tests. The practical questions remain the ones a wearer will settle: does it help, does it last, and does it deserve to come along tomorrow?
As of October 2026, Halo’s product listing carries a $399 price and production language, alongside shipping statements that still require careful reading. Tavangar’s project remains a story of making and delivering a new kind of everyday object. His small-company strategy has brought him through several designs; the daily usefulness has to be earned again with each one.
The Hobbit analogy is enjoyable because it leaves room for an uncertain journey. Tavangar has chosen a smaller object, an open platform, and a route through developers who can supply their own ideas. Somewhere between the drawing, the factory, and the face, that argument has to become a pair of glasses someone reaches for without much ceremony. For a computer with such expansive ambitions, ordinary use would be a meaningful destination.