Muscles humans coordinate in real-time for simple motions like hammering
The Master of Escape
In a well-documented laboratory experiment, an octopus is placed inside a sealed jar with a screwed-on lid. Without prior knowledge or training, it explores the container with its arms, generates torque, and unscrews the lid from the inside to escape. This invertebrate, whose parents die before it hatches, operates not on inherited knowledge but on embodied instinct and real-time, trial-and-error problem-solving. Its intelligence is not centralized in a brain packed with experience but distributed across a perceptive, adaptive body. The octopus, as marine biologist Peter Godfrey-Smith argues, represents an entirely different evolutionary pathway to consciousness—one where “mind” is inseparable from direct, physical interaction with the world. Its success forces a radical question for human engineers: have we over-indexed on brain-centric, knowledge-based intelligence at the expense of a more ancient, bodily wisdom?
Beyond the Brain’s Bounds
The human engineering mindset has long been dominated by a “brain as CPU” metaphor. We prize knowledge—the compressed experience of the past—and build systems to store, process, and replicate it. This worked brilliantly in a stable, predictable world. But in a world of sharp, unpredictable changes, knowledge has a half-life. The octopus reveals an alternative: wisdom. Wisdom is the capacity for adaptive action in novel situations, derived not from data recall but from integrated perception and embodied response. Engineer and philosopher Shuichi Fukuda frames this as a shift from Pattern Recognition (classifying inputs based on known data) to Pattern Cognition (understanding a situation holistically to inform action). The latter is the domain of instinct, honed by evolution for survival. Our challenge is to design systems that don’t just execute pre-loaded instructions but can perceive, adapt, and “learn to swim” in uncharted waters.
The Architecture of Adaptation
Human movement science reveals our own latent capacity for this embodied wisdom. Russian physiologist Nikolai Bernstein’s famous “cyclogram” studies showed that human motion trajectories (like hammering) are wildly variable at the start but converge to a precise path as the hand nears the target. We don’t execute a pre-programmed script; we use continuous sensory feedback to coordinate and balance over 600 muscles in real-time, reducing degrees of freedom until the goal is achieved. This “motor control” is tacit, nonverbal, and largely inaccessible to our conscious brain—the epitome of Michael Polanyi’s “tacit dimension.” Yet, most machine “motion control” is built only to replicate the final, reproducible trajectory, ignoring the essential, adaptive search process that precedes it. We automate the result but discard the capability.
The Peril of Perfect Information
Modern Artificial Intelligence, particularly deep learning, excels in domains of “perfect information”—games like Go or Chess with fixed rules, known players, and clear victory conditions. Here, brute-force computation on historical data can yield superhuman strategies. This mirrors the old engineering paradigm: a closed world where backward induction from a known goal is possible. However, as Herbert Simon’s concept of “bounded rationality” established, most real-world decisions must be made with limited time, information, and computational power. In an open world—be it diagnosing a patient in an emergency or navigating a business through a market shock—we cannot optimize. We must “satisfice”: find a good-enough solution that works now. This is the realm of pragmatic, instinctive action, where the octopus thrives and where our current AI models often falter.
Engineering for the Tacit Dimension
The path forward requires a synthesis. It demands engineering that respects and leverages human instinctive intelligence rather than seeking to replace it with brittle, knowledge-based automation. This means designing support tools for human judgment, not autonomous black boxes. It involves creating interfaces and systems that enhance our innate ability for pattern cognition and satisficing. For example, a diagnostic tool for a doctor shouldn’t just output a single probability but should visually pattern multi-sensor patient data (vitals, imaging, labs) in a way that aligns with the clinician’s holistic, instinctive assessment process. The goal is not to make the doctor think like a computer, but to make the computer support the doctor’s human genius for synthesis under uncertainty. The cathedral builder trusted his embodied skill; the modern engineer must build the tools that restore that trust in our own innate capacities.
References
Bernstein, N. A. (1967). The co-ordination and regulation of movements. Pergamon Press. Fukuda, S. (2019). Self engineering: Learning from failures. SpringerBriefs in Applied Sciences and Technology. Godfrey-Smith, P. (2016). Other minds: The octopus, the sea, and the deep origins of consciousness. Farrar, Straus and Giroux. Polanyi, M. (1966). The tacit dimension. University of Chicago Press. Simon, H. A. (1947). Administrative behavior: A study of decision-making processes in administrative organization. Macmillan.