A Vision for Walking Robots
“Marc Raibert. Or just call me Dr. Raibert.”
“Yoo Jae-won.”
“I know who you are. Youngest Fields Medalist, and now one of the richest men on the planet thanks to ID Group. Been a few years already, hasn’t it?”
From the very first greeting, Marc Raibert’s personality came through loud and clear. Whether you called it blunt or simply incapable of small talk was up for debate.
“I caught your vision presentation this year. Something about system integration and artificial intelligence, if I recall.”
“Wow, you’re well-informed.”
“Of course. When a major investor finally shows up after all this time, the least I can do is look into him. Though it seemed like your interest was only in software. I didn’t expect you to care about our company.”
“Did it come across that way?”
Looking back, if you separated system integration from artificial intelligence, it was true that the focus had been on the software side. By that measure, Dr. Raibert was a hardcore hardware enthusiast. His dream was to build the perfect bipedal robot that moved exactly like a human.
Many roboticists dismissed bipedal locomotion as inefficient. Implementing two-legged walking was simply too difficult. Making the legs sturdy enough, ensuring the durability of the power systems and gears, and developing a reliable posture-control system were all monumental challenges. It would be far simpler to stick with proven, time-tested wheels.
When colleagues raised these criticisms, Marc Raibert would counter: the world is built around the human form. That is the essence of the human interface. Rather than creating specialized robots for every task, a machine built in our own image could immediately integrate with existing devices and facilities.
“How could anyone not be drawn to a bipedal robot that walks on two legs? And if you layer on an artificial intelligence with specialized expertise… that would be something else entirely.”
“Oh?”
Yoo Jae-won’s slightly excited tone made Marc Raibert’s expression shift. Moments earlier he had been faintly guarded, as if wondering what this meeting was really about. The moment he recognized a kindred spirit, that wariness vanished. The idea of giving a finished robot true artificial intelligence went one step beyond even his own vision.
“Artificial intelligence, huh. Noble goal, but the computational requirements would be enormous.”
“If we’re talking about a standalone AI, yes. But why go that route? All we need is a system capable of running a lightweight client program and a high-speed wireless connection. The actual AI can live in the cloud and communicate remotely.”
The robots Marc Raibert designed were fundamentally independent entities—machines that could move on their own once powered on. What Yoo Jae-won envisioned was different: robots wirelessly linked to a massive cloud-computing AI.
“That’s the most interesting thing I’ve heard in a while. Artificial intelligence running over wireless links that crawl like snails.”
“All technology improves. Ten years ago the internet was something only a tiny handful of people even knew about, and the speeds were terrible. Now? Most households in North America have 8 Mbps service. In a densely populated country like Korea, 50 Mbps VDSL is already available.”
“What? 50 Mbps? Six megabytes per second?”
“It’s asymmetric, so only the download speed reaches that rate.”
Marc Raibert looked genuinely stimulated by the numbers. But that was as far as it went. At the moment he was still building the prototype of BigDog, a quadrupedal robot modeled after an animal, while the bipedal version had only its lower body completed and required external power and posture-stabilizing cables just to move. In short, it was still a distant dream.
Still, it wasn’t a bad thing. He had confirmed that his new backer possessed a clear and ambitious vision.
“Doctor! You’ve left our distinguished investor standing out here in the cold far too long. He’ll catch a chill!”
While they had been deep in conversation, one of the researchers who had come out with Marc Raibert finally reminded them of the obvious. They had been talking at the entrance for quite some time. Now that he thought about it, it was rather cold. Unlike the mild winters of San Francisco, Boston’s wind cut straight through you.
“Let’s go inside.”
Marc Raibert seemed to feel the chill at last and led Yoo Jae-won and his party into the building.
His expectations proved correct. The investment contract Yoo Jae-won presented exceeded Marc Raibert’s hopes by three hundred percent.
“Fifty-one percent equity for 5.1 million dollars. In other words, you’re valuing Boston Dynamics at ten million dollars.”
“Yes. I valued it generously. Is it insufficient?”
At the question, several thoughts flashed through Marc Raibert’s mind. In truth, he had privately valued Boston Dynamics at around three million dollars. The company’s assets consisted of a few prototypes built with a few hundred thousand dollars and the software running on them. The building they occupied had been rented cheaply from MIT and wasn’t factored into the valuation. The research staff were simply former students he had kept an eye on during his time at MIT. Patents—the most important asset in valuing a venture—were fewer than expected. The robotics field already had established players. European companies were formidable, and Japanese firms held significant market share. While Boston Dynamics would one day grow enormously, right now it was still taking its first steps. Real patents would only emerge once the Atlas bipedal prototype was complete. At present, it was essentially a one-man operation led by Marc Raibert himself.
That was why the offer of 51 percent equity for 5.1 million dollars felt like an extraordinarily generous valuation to both Marc Raibert and his employees. The reason was impossible to discern. It even made him wonder whether Boston Dynamics possessed some hidden value he himself didn’t know about. Naturally, that thought led to the idea that he might be able to negotiate a higher price.
It was, of course, a complete miscalculation. Both Marc Raibert’s own assessment and the due diligence conducted by ID Group’s secretariat had placed Boston Dynamics’ current value in the high two millions. The fivefold premium was simply a matter of conscience—though Yoo Jae-won’s conscience was already riddled with compromises made in the name of ID Group’s growth. He had sworn there would be no illegal activity in running the group, yet he had still used future knowledge to seize technologies, products, and promising companies ahead of everyone else. Boston Dynamics was no different. It would one day grow larger than Google. Even 5.1 million dollars for 51 percent was still a bargain, but Yoo Jae-won intended to more than make up for it with overwhelming support after the acquisition. The 5.1 million was also his way of honoring Marc Raibert’s passion.
“Hmm. Wanting 51 percent means you intend to take control of the company.”
“Yes. As you know, when I invest in venture companies I generally secure management rights as well. Not to exert control for its own sake, but to maximize the development capabilities that are a venture’s greatest strength. Boston Dynamics will be able to focus entirely on research without worrying about secondary management concerns. We’ll create that environment for you.”
Yoo Jae-won handed over the prepared contract. It was the standard agreement ID Group used when investing in ventures. While minor adjustments had been made, the core terms Marc Raibert received were essentially the same.
“Ah. So once I sign, I’ll receive a steady salary. And the contract term is ten years—quite long-term.”
As expected of an MIT professor, he immediately grasped the key points. The reason Marc Raibert had hesitated despite the extraordinary sum offered was the transfer of management rights. Yet reading the contract, it didn’t seem so bad after all. The thought of never having to worry about payday again was genuinely appealing. Becoming a salaried researcher who could devote himself entirely to his work sounded ideal.
And that wasn’t all. The contract listed the benefits that would come with joining ID Group: technical support of every kind and a generous research budget. Until now they had barely kept the company running on alumni donations and government grants from agencies like DARPA. This would change everything. All it took was a signature.
Still, Marc Raibert hesitated. Terms this good usually meant there was a catch. He had received investment offers before and turned several down. There had been no concrete reason; his instincts had simply told him to wait. Because of that, the eyes of the employees watching the meeting from a distance were filled with desperate hope. They had never missed a paycheck, but they prayed this time Marc Raibert wouldn’t let his whims get in the way. Compared to the minor-league capitalists and investors who had visited before, Yoo Jae-won’s presence was on an entirely different level.
“Hoo. I’ll sign.”
Marc Raibert’s instincts had created hesitation, but no matter how many times he examined the offer, he couldn’t find a reason to refuse. Cold logic told him that a deal this good wouldn’t come along again even in another ten years.
“An excellent choice.”
Yoo Jae-won took a fountain pen from his inner pocket and offered it. It was a high-end pen reserved exclusively for signing contracts.
“Something like this suits an engineer like me better.”
After staring at the fountain pen for a moment, Marc Raibert picked up the ordinary ballpoint pen lying on his desk—the one with the Bic Round Stic logo. The writing instrument was plain, but the significance of the signature remained undiminished. With that stroke, the contract was complete. Fifty-one percent of Boston Dynamics, formerly Marc Raibert’s private company, now belonged to Yoo Jae-won.
Even after the signing, Yoo Jae-won did not leave immediately. There was still something he needed to discuss.
“What’s this drawing?”
Marc Raibert asked after examining the sketch Yoo Jae-won had handed him. Moments earlier his eyes had been filled with gratitude toward his savior; now they had done a complete one-eighty, as if to say, “So you’re that kind of person after all.”
“A robot vacuum cleaner.”
Even after the explanation, Marc Raibert’s expression did not soften. The square, flat robot vacuum in the sketch was equipped with wheels, after all. Marc Raibert didn’t hate wheels, but the robot he wanted to build was a humanoid android. The disappointment was palpable. Until just a moment ago, Yoo Jae-won had seemed to share his vision; now that the contract was signed, he was showing his true colors.
“I have no intention of abandoning bipedal robots. I simply want to demonstrate as quickly as possible how robotics can enrich our daily lives. That’s why I chose the robot vacuum cleaner as the first product.”
“It’s true that robotics still feels like a distant future to most people.”
“Besides, if we suddenly unveiled something like BigDog that moves exactly like an animal, or a robot that walks as naturally as a human, how would people react? Many would feel revulsion—the uncanny valley effect. So first I want to show them that robotic technology isn’t so far away, using something familiar like a vacuum-cleaning robot.”
At the same time, it would strengthen the home-appliances division of ID Electronics. Yoo Jae-won’s rapid explanation eased the tension on Marc Raibert’s face somewhat.
“Hmm. This little guy does look friendly.”
He even showed genuine interest. The sketch wasn’t a rough pencil drawing but a polished computer graphic designed to maximize cuteness.
“What are the essential features for a robot vacuum?”
“It has to move wirelessly. It needs ultrasonic sensors to detect obstacles, remember the paths it’s already cleaned to maximize efficiency, and return to a charging station automatically when the battery runs low. The user should only have to empty the dustbin when it’s full. Ideally, even that process would be automatic.”
As Marc Raibert considered the various components, he offered one comment.
“It’s going to be an expensive piece of equipment.”
“Heh, at first, yes. But once we start mass-producing a million units, the economics change completely.”
It was the confidence of a man who owned ID Electronics.
“For that to happen, Doctor, you’ll have to build us a proper robot vacuum.”
“Understood. I’ll make one that’s truly worthy of the name.”
Marc Raibert nodded with confidence. Core components such as ultrasonic sensors, high-performance cleaning motors, and mobile application processors were all readily available within the group. With the world’s foremost authority on robotics fully committed, there was every chance they would create a product that transcended its era.
As Yoo Jae-won moved on to his next appointment, his heart felt light. He had successfully brought one of the most critical figures in his twenty-first-century strategy inside the ID Group’s walls. Even though the final item on his East Coast schedule was somewhat burdensome, his mind remained at ease.
With that, Yoo Jae-won and his party headed for CIA Headquarters in Langley, Virginia.