Rodney Brooks Challenges the Billion-Dollar Investment in Humanoid Robotics
Rodney Brooks, a distinguished expert in robotics, cautions investors pouring billions into humanoid robot startups that their money might be squandered. He specifically critiques companies like Tesla and Figure, which attempt to teach robots manual skills by analyzing videos of humans performing tasks. Brooks dismisses this approach as unrealistic and overly optimistic.
The Complexity of Human Dexterity: A Barrier for Robots
The human hand’s extraordinary sensitivity and dexterity stem from approximately 17,000 tactile receptors, a level of complexity no current robot can replicate. While machine learning has revolutionized fields such as speech recognition and image analysis, these advances rely on extensive, well-established datasets. Brooks points out a critical gap: the absence of comprehensive tactile data for robots to learn from, which severely limits their ability to develop fine motor skills.
Safety Concerns and Energy Demands of Humanoid Robots
Another significant challenge is safety. Humanoid robots of human size consume vast amounts of energy to maintain balance and mobility. Their potential to cause harm increases exponentially with size; for instance, a robot twice as tall as current models could deliver eight times the impact energy if it falls, posing serious risks in real-world environments.
The Future of Robotics: Moving Beyond the Human Form
Brooks predicts that within the next 15 years, successful robots will likely abandon the humanoid shape altogether. Instead, they will feature wheeled bases, multiple arms, and specialized sensors tailored for specific tasks. He argues that the current billions invested are largely funding costly training experiments that lack scalability for mass production.
AI’s Mixed Impact on Productivity: Insights from Recent Studies
Brooks has also expressed skepticism about the hype surrounding generative AI, noting that its promise often exceeds its practical benefits and can even increase workload. Supporting this view, a recent study by the AI research nonprofit METR involved 16 top-tier developers tackling nearly 250 real-world coding problems with and without AI assistance. Surprisingly, developers took 19% longer to complete tasks when using AI tools, despite perceiving a 20% acceleration in their work.
Debunking AI as an Existential Threat
Contrary to fears voiced by figures like Elon Musk, Brooks has long maintained that AI does not pose an existential risk. In a 2017 discussion at MIT, he highlighted the emerging trend of companies specializing in creating machine learning datasets, a development that continues today. He also emphasized that despite Big Tech’s vast data resources, their dominance in robotics is not guaranteed, though many leading robotics firms are increasingly partnering with these tech giants.
Collaborations Between Robotics Startups and Tech Giants
For example, Apptronik, a humanoid robotics startup, counts Google among its investors and recently teamed up with Google’s DeepMind to integrate advanced AI with cutting-edge robotic technologies. Similarly, Figure, backed by Microsoft and the OpenAI Startup Fund, partnered with OpenAI in early 2024 to merge OpenAI’s research capabilities with Figure’s robotics expertise. Figure’s latest funding round raised over $1 billion, valuing the company at nearly $39 billion.
