Episode #469: Can Tesla Teach a Bot to Bachata?

In this episode of the Crazy Wisdom Podcast, I, Stewart Alsop, sit down with returning guest Brian Ahuja to explore a thought-provoking idea he’s been stewing on—could we one day build a robot capable of true partner dancing? From the biomechanics of salsa to the possibilities of AI embodiment, we unpack what it would take to engineer fluid, responsive movement and how that intersects with everything from artificial muscles to the intimacy of tactile feedback. We also touch on Brian’s long-term vision for a potential lab or foundation to tackle this challenge. You can follow Brian and future developments on Twitter @brianahuja.

Check out this GPT we trained on the conversation

Timestamps

00:00 – Brian Ahuja returns to discuss AI embodiment, sparked by his experience in ballroom dance and curiosity about translating physical intelligence into robotics.
05:00 – They explore robotics in partner dancing, touching on the difference between choreographed motion and improvisational, responsive movement.
10:00 – Brian breaks down human biomechanics, emphasizing that hip motion in dances like salsa originates from knees and feet—not the hips directly.
15:00 – The conversation shifts to balance, proprioception, and ocular reflexes, linking them to movement stability in dance.
20:00 – They compare robot vs. human movement, noting robots’ jerky motions and the absence of muscle-based initiation.
25:00 – The need for haptic feedback is discussed, with Brian detailing how partner dancing depends on tactile signals and real-time response.
30:00 – They touch on robotic form factors, questioning whether humanoid robots are the best approach and pondering the design of artificial muscles.
35:00 – Brian proposes the idea of the Ahuja Test, gauging if a robot can move so fluidly it's indistinguishable from a human, using dance as the standard.

Key Insights

1. Partner Dancing as a Frontier for Robotics: Brian Ahuja proposes that partner dancing could be a benchmark for robotic embodiment, where success would indicate a robot’s ability to replicate fluid, responsive human movement. This task is far more complex than solo choreography—it requires real-time tactile feedback, improvisation, and nuanced physical communication.
2. Movement Origin in Humans vs. Robots: A critical difference lies in how movement is generated. Human motion begins with muscle contraction, not at the joints. Robots, however, typically initiate movement at joint points, missing the layered interplay of muscles, tendons, and fascia that create smooth, lifelike motion.
3. Haptic Feedback and Improvisation: Real partner dancing involves subtle cues, like pressure through fingertips, to signal direction and timing. For a robot to follow or lead a dance, it would need a highly sensitive haptic feedback system capable of interpreting and responding to these nonverbal signals in real time.
4. The Limits of Current Robotics: Even with advanced robots like the Tesla bot, current movement still appears jerky and lacks the fluidity needed for partner dancing. The mechanical design—especially the lack of artificial musculature—may impose fundamental limits on how closely robots can mimic human motion.
5. Applications Beyond Dance: The implications of this inquiry stretch beyond dance into fields like physical therapy, elder care, and domestic robotics. A robot that could move like a human could handle tasks requiring adaptability, precision, and physical sensitivity.
6. Vision and Systems Thinking: Brian frames the challenge as a systems problem that might start with a lab or foundation. He emphasizes not needing to do everything alone, recognizing the value of building knowledge iteratively through conversations, research, and community.
7. The Ahuja Test: Inspired by the Turing Test, Brian coins the idea of the “Ahuja Test”—a way to measure if a robot can move indistinguishably from a human. He suggests partner dancing could serve as the ultimate proving ground for such a test, given its demand for embodied intelligence and nuanced coordination.