AI Humanoid Robots Are Hitting the iPhone Moment | 1X, Bernt Børnich

Key Concepts

• Accepting failure as part of innovation
• Importance of culture in fostering innovation
• Humanoid robots and their potential impact
• Tendon-driven robotics vs. traditional robotics
• Low-energy systems for safe human-robot interaction
• Social aspects of robots working among humans
• Consumer-first approach for scaling innovative products
• Deep tech development and solving fundamental problems
• Importance of passion and excitement in a long journey

Innovation and Culture

Bar Borick emphasizes that accepting failure is crucial for innovation. He argues that most groundbreaking endeavors will initially be "plain out wrong," even "borderline stupid" in hindsight. However, this is a natural part of the innovative process. The only unacceptable behavior is a lack of effort or failure to learn from mistakes. He states, "The only thing that's not acceptable from a culture point of view is you didn't really try...or you didn't actually reflect over why did it fail so you can learn."

Borick stresses the importance of maintaining this culture of accepting failure even under pressure to meet deadlines and milestones. He believes that doing so leads to "great innovation."

OnX and Humanoid Robots

Bar Borick is the founder of OnX, a company that creates humanoid robots. The initial motivation stemmed from the observation that robotic videos online were often sped up, highlighting the unnatural movement of robots. OnX aims to create robots that move at human speeds and naturally. Their mission is to "create an abundance of labor through these intelligent machines." The vision is a future where humanoids assist with everyday tasks, freeing up humans to focus on more meaningful activities.

Borick's inspiration came from Honda's Asimo robot, particularly its ability to interact with people. He envisioned robots performing chores, similar to the robots in Star Wars. However, he also analyzed Asimo's failures, attributing them to the "classical robotics regime" used in factories, which relies on precise environmental assumptions that don't hold true in real-world scenarios.

Technical Approach: Tendon-Driven Robotics

OnX takes a different approach to robotics, focusing on tendon-driven systems (also known as cable drives). This approach is inspired by how humans are built and move. While not entirely novel, OnX has invested significant effort in making it work effectively. This creates a "moat" because it's difficult for others to catch up.

The key advantage of tendon-driven systems is their low energy and compliance. Traditional industrial robots use high gear ratios (e.g., 100:1), resulting in enormous energy in their movements. Borick explains, "If your arm is moving like this, something inside here is spinning 100 times faster...something here is spinning at 20,000 RPM and then when your arm hits something, this needs to immediately stop." This makes them dangerous and unsuitable for unstructured environments.

Low-energy systems are crucial for safe interaction with humans and the environment. They allow robots to explore and learn through trial and error without causing damage or injury. Borick uses the analogy of kinetic energy, stating that "if a car moves twice as fast, it's not twice as dangerous, it's four times as dangerous because it falls to square."

Social Intelligence and Human-Robot Interaction

Borick highlights the social aspects of robots working among humans. Even simple tasks like getting something from the fridge involve social interaction and communication of intent. He argues that "any kind of labor that you do among people is kind of social labor."

Manufacturing and Scaling

OnX is taking a vertically integrated approach, building not only the robots but also the machines and automation equipment needed to manufacture them. This is because the required components for their tendon-driven system don't exist off-the-shelf. They are focused on simplifying the product design to minimize complexity, reduce part count, and avoid the need for special alloys. The goal is to reduce the complexity from that of a car to something closer to an electrical appliance.

Consumer-First Approach

Borick argues for a consumer-first approach to scaling humanoid robotics. He notes that historically, innovative products rarely start in enterprise due to risk aversion and bureaucratic hurdles. He uses the example of ChatGPT, which gained traction through consumer adoption before facing resistance in enterprise settings. He believes that consumer adoption creates "bottom-up pressure" that eventually forces enterprise adoption.

Deep Tech and Solving Fundamental Problems

OnX is characterized as a deep tech company, focused on solving a fundamental problem: turning energy into any kind of labor. Borick states, "If you cure cancer, you're not wondering whether or not people will actually buy your product." He acknowledges the importance of product-market fit but emphasizes that the initial focus is on solving the core technical challenge.

Robot Learning and Teleoperation

The robot learning process involves a combination of:

1. Internet data: Leveraging existing datasets.
2. Synthetic simulated data: Training in virtual environments.
3. Robot data: Gathering data from real-world interactions.

Teleoperation is used to bootstrap the robot's learning. This involves a human embodying the robot, seeing through its eyes, and controlling its movements. Borick describes it as a "magical experience" and a way to transfer knowledge from humans to machines.

Passion and Excitement

Borick concludes by emphasizing the importance of passion and excitement in a long and challenging journey. He states, "Focus on the things that make you happy because it's going to be a long journey and if you're not having fun you're not going to make it." He advises choosing a problem that excites people to attract top talent. He says, "The best cheat code I have is pick a problem that really excites people...We do human robots. It's really freaking cool."

Conclusion

The video provides insights into OnX's approach to developing humanoid robots, emphasizing the importance of accepting failure, fostering a culture of innovation, using tendon-driven robotics for safety and compliance, focusing on social intelligence, adopting a consumer-first scaling strategy, and prioritizing passion and excitement in the long journey of deep tech development.