Rodney Brooks on the Nature of Intelligence, Computation, and the Future of Robotics

This episode features a deep dive with Rodney Brooks, a titan in robotics, exploring his journey from childhood fascination to co-founding pioneering companies like iRobot and Rethink Robotics, and his current venture, Robust.AI, focused on teaching robots common sense. The conversation begins with a discussion of the aesthetically and mechanically beautiful robot Domo, highlighting the importance of a robot's appearance not "over-promising" its capabilities, a principle Brooks applied to his collaborative robots Baxter and Sawyer with their expressive, intent-communicating eyes. Brooks shares his early inspirations from "How and Why Wonder Books" and his rudimentary experiments with chemical-based learning systems, setting the stage for his lifelong pursuit of building intelligent machines.

A central theme of the discussion revolves around the fundamental question, "Can machines think?" Brooks asserts that humans are machines, and thus machines can think, but questions whether humanity possesses the intelligence to build such machines. He introduces his upcoming book, "Not Even Wrong," which critiques the prevailing computational metaphor for intelligence. The book traces the history of computation from early mathematicians to Alan Turing's 1936 paper, arguing that Turing's definition of computation was a human-derived construct, not a universal given, which later became incredibly powerful due to its silicon implementation and Moore's Law. Brooks highlights how four key disciplines—Artificial Intelligence, Artificial Life, Neuroscience, and Abiogenesis—all adopted computation as their primary metaphor between 1945 and 1965, often with overlapping researchers like Warren McCulloch and John von Neumann.

Brooks challenges the notion that computation, as currently understood, fully captures the essence of human intelligence, suggesting a gap between the computational machine and the conscious, intelligent mind. He posits that intelligence might not reside solely in the brain but also in the entire body and, crucially, in social interaction and the ability to externalize and share knowledge. He emphasizes that all known intelligences, from humans to octopuses, perceive and act in the world in complex ways, constructing reality (e.g., color constancy) rather than merely processing raw data. This leads to a discussion of the "symbol grounding problem" and why deep learning, while effective for labeling, hasn't truly solved it.

The conversation culminates in an exploration of "Maravec's Paradox," which states that reasoning is easy for machines, but perception and mobility are hard. Brooks aligns with this, attributing it to evolution's long investment in developing sophisticated perceptual and motor skills over hundreds of millions of years, compared to the relatively recent emergence of higher-level reasoning, language, and agriculture. He illustrates the complexity of manipulation with an anecdote about his 16-month-old grandson intuitively operating a window handle. Ultimately, Brooks suggests that the hardest parts of robotics are all of them, emphasizing the need for a deeper understanding beyond reductionist approaches to truly replicate or create intelligence.

"if your robot looks like albert einstein it should be as smart as albert einstein"

"I think it's a big fear I think any other philosophical position is sort of a little ludicrous what does think mean if if it's not something that we do and we and we are machines so yes machines can but do we have a clue how to build such machines that's a very different question"

"turing in that paper set out to define what sort of machine could do that mechanical machine where it could produce an arbitrary number of digits in the same way a human computer did"

"it's not given by the universe it was this is what we're going to call computation"

"the word computation very quickly starts doing a lot of work that it was not initially intended to to do"

"our desire for metaphor and combined with our limited cognitive capabilities gets us into trouble"

"when deep learning came along and started labeling images people said ah the grounding problem has been solved no the labeling problem was solved with some percentage accuracy which is different from the grounding problem"

"what did evolution what did evolution spend its time on yes it spent its time on getting us to perceive and move in the world that was you know 600 million years as multi-cell creatures doing that"

"my 16 month old grandson was in his new house first time right first time in this house and he'd never been able to get to a window before but this had some low windows and he goes up to this window with a handle on it that he's never seen before and he's got one hand pushing the window and the other hand turning the handle to open the window he he knew he two different hands two different things he knew how to how to put together yeah and he's 16 months old"