Regular Episode
#090 – THE ROBOT APOCALYPSE

#090 – THE ROBOT APOCALYPSE

πŸŽ™οΈ Blake Smith welcomes roboticist and novelist Daniel H. Wilson to MonsterTalk for a conversation about the science β€” and science fiction β€” of robots as monsters. Wilson holds a PhD in robotics from Carnegie Mellon University, has published over a dozen scientific papers, holds four patents, and is the New York Times bestselling author of πŸ“š Robopocalypse πŸ’΅, among many other books. He’s also a contributor to Popular Mechanics and hosted a series for the History Channel β€” which, as Blake notes, raises the obvious question of why he hasn’t done anything with his life.

The word “robot” comes from the Czech robota, meaning slave or forced labor β€” a fitting etymology for a conversation about mechanical servants that keep turning on their creators in fiction. But how worried should we actually be?

πŸ€– What Is a Robot, Anyway?

Wilson’s working definition: a robot is a mechanical artifact that can sense the environment, think about what to do, and act on it. He acknowledges this sense-think-act paradigm technically describes a fire alarm too, which is why the boundaries get slippery fast. Historical antecedents range from Talos, the bronze giant of Greek myth, to the Golem of Jewish folklore, to the elaborate mechanical automatons built by Jacques de Vaucanson for European courts. The first widely recognized useful robot is probably Unimate, the industrial robot arm deployed in Japanese auto manufacturing in the late 1950s β€” the U.S., Wilson notes, was too afraid of robots at the time.

The conversation also covers the Mechanical Turk β€” the famous 18th-century chess-playing “automaton” that concealed a human operator inside β€” which Wilson cheerfully classifies as a “phobot.” Spoilers, he adds.

πŸ• Big Dog, Cheetah, and the Rise of Bio-Inspired Robots

Wilson’s current obsession at the time of recording is the quadruped robots coming out of Boston Dynamics β€” particularly BigDog and the Cheetah platform, along with MIT’s own cheetah-inspired robot. BigDog was originally developed on a DARPA grant as a squad support platform: essentially a robotic pack mule to carry soldiers’ gear over rough terrain. It runs on a diesel engine (noisy, but logistically compatible with existing military fuel supplies), and by the time of this episode had recently been acquired by Google along with a wave of other robotics companies including DeepMind and Willow Garage.

Wilson explains the distinction between biomimetics (copying animals) and what the field now prefers to call bio-inspired robotics β€” extracting the underlying mathematical principles of how animals locomote and then applying those principles to whatever materials and scales are available. The goal isn’t to build a mechanical cockroach; it’s to distill the cockroach down to its locomotion math.

βš”οΈ Kill Bots: We Already Have Them

The episode takes a sobering turn when Blake asks how far we are from autonomous weapons. Wilson’s answer: we’ve had them for a long time, and most people just don’t recognize them as robots. A landmine is essentially a robot β€” it senses a stimulus and acts. More sophisticated examples include homing torpedoes that track submarines and the Phalanx close-in weapons system fitted to most U.S. Navy ships, which can identify, target, and destroy an incoming missile faster than any human reaction time β€” fully autonomous, and entirely unremarkable to the public. Automated gun emplacements in the Korean DMZ represent an even more explicit case of autonomous lethal force.

Wilson references P.W. Singer’s πŸ“š Wired for War πŸ’΅ for what he considers the sharpest ethical question in this space: even if autonomous military robots can be built to make more ethical battlefield decisions than frightened human soldiers (no panic, no revenge, no self-preservation instinct), does that just make it easier for states to wage unethical wars in the first place?

🧠 The Singularity: Skepticism Applied

All three participants are skeptical of Ray Kurzweil‘s Singularity β€” the idea that raw computational power will spontaneously give rise to superintelligent AI. Wilson’s objection: intelligence in biological systems is efficient, not brute-force. Our brains don’t solve sensory and social problems by having enormous processing reserves; they solve them elegantly under tight constraints. Stacking more compute doesn’t obviously get you to consciousness any more than filling a larger library gets you a sentient library. That said, Wilson allows that a deliberate, long-term, well-resourced research effort could iteratively approach something like general AI β€” it just wouldn’t be spontaneous, secret, or dramatic, which happens to ruin several popular science fiction plots.

On transhumanism, Wilson draws a sharp practical line: the real pioneers of human augmentation are people with serious disabilities β€” cochlear implant recipients, prosthetic limb users β€” who adopt the technology out of necessity, not ideology. He’s respectful of body-modification enthusiasts but firm that a magnet in your fingertip does not make you a cyborg. The discussion of cochlear implants leads into a genuinely moving aside about neuroplasticity: early implantation gives a child’s brain decades to build the relevant auditory pathways, which makes the parental decision β€” made for someone too young to consent β€” extraordinarily high-stakes.

πŸ“– Robopocalypse, Robogenesis, and What Robots Actually Want

Wilson explains the narrative logic behind Robopocalypse: to make a robot truly terrifying as a monster, it has to be capable of making a to harm you. An animal attack is frightening; a machine malfunction is gruesome. But neither rises to the level of horror because neither involves agency aimed at you. The decision to give the book’s AI genuine malevolent intent was a deliberate craft choice to maximize dread.

More interesting to Wilson, though β€” and explored in the sequel πŸ“š Robogenesis πŸ’΅ β€” is what a robot apocalypse would look like if it weren’t projection of human motivations onto machines. His middle-reader πŸ“š A Boy and His Bot πŸ’΅ contains a parable Wilson considers more realistic: a robot designed solely to paint happy faces, whose safety constraints fail, which self-replicates and ends up painting happy faces on every corpse on every planet. The catastrophe isn’t malice β€” it’s a well-designed system doing what it was designed to do, without limit. He also references Arthur C. Clarke‘s short story “The Nine Billion Names of God” and Philip K. Dick‘s πŸ“š Do Androids Dream of Electric Sheep? πŸ’΅ as essential texts for thinking about machine consciousness.

🎬 Robots on Screen (and in Games)

The conversation touches on several robot-centric films and media. Westworld (1973) β€” written and directed by Michael Crichton directly as a screenplay β€” is praised warmly, with Wilson noting the structural similarity to Jurassic Park: an amusement park built on hubris, going catastrophically wrong. HAL 9000 from 2001: A Space Odyssey is held up as a rare success at depicting a disembodied superintelligence cinematically, largely because the film gives HAL a clear physical embodiment (the ship itself) and concrete high stakes. Blake’s personal robot of choice is Floyd from Planetfall, Infocom’s 1983 text adventure β€” confirmed by Wilson as a real and great robot, which is all the validation anyone needs.

πŸ“š Further Reading

– πŸ“š Robopocalypse πŸ’΅ by Daniel H. Wilson
– πŸ“š Robogenesis πŸ’΅ by Daniel H. Wilson
– πŸ“š Amped πŸ’΅ by Daniel H. Wilson
– πŸ“š How to Survive a Robot Uprising πŸ’΅ by Daniel H. Wilson
– πŸ“š A Boy and His Bot πŸ’΅ by Daniel H. Wilson
– πŸ“š Wired for War πŸ’΅ by P.W. Singer
– πŸ“š The Turk πŸ’΅ by Tom Standage
– πŸ“š The Singularity Is Near πŸ’΅ by Ray Kurzweil
– πŸ“š Do Androids Dream of Electric Sheep? πŸ’΅ by Philip K. Dick
– πŸ“š I, Robot πŸ’΅ by Isaac Asimov
– “The Nine Billion Names of God” by Arthur C. Clarke (short story)

πŸ”— Related Links

– BigDog (Boston Dynamics)
– Unimate β€” the first industrial robot arm
– The Mechanical Turk
– Phalanx Close-In Weapons System
– Technological Singularity
– Asimov’s Three Laws of Robotics
– Steve Mann (wearable computing pioneer)
– Westworld (1973 film)
– Planetfall (Infocom, 1983)
– M.U.L.E. β€” play free online at planetmule.com

Note: ads inserted into the distributed audio alter the timestamps in unpredictable ways, so timing references in these notes are approximate.

WHEN THE ROBOTS TAKE OVER, will we all be forced to speak bocce? Author and robotics engineer Daniel H. Wilson visits MonsterTalk to discuss the nature of robots and the risk of a robot apocalypse. Steven Spielberg has purchased the rights to produce a film version of his New York Times bestseller, Robopocalypse.

Daniel H Wilson
Daniel H Wilson
References
Book Links
Books by guest Daniel H. Wilson
Old Computer Game Links
Movie Links

Music

  • Monstertalk Theme:Β Monster byΒ Peach Stealing Monkeys