Hans Moravec

• 1. Introduction

  ›

  Hans Peter Moravec was born on the 30th of November 1948, and he grew up to ask a question that most scientists would not dare put on paper: what happens when machines become smarter than we are? Moravec is an Austrian-born Canadian computer scientist, an adjunct faculty member at the Robotics Institute of Carnegie Mellon University in Pittsburgh. He built robots that could feel their way through cluttered obstacle courses, co-founded a company to make them navigate entirely on their own, and published predictions about the coming transformation of intelligence that attracted praise from Arthur C. Clarke and sharp rebuke in equal measure. He also quietly published an argument about consciousness seven years before a more celebrated philosopher made a similar case. This is the story of a man who spent decades working at the edge of what machines can do, and then spent equal time imagining where that edge was heading.

• 2. Building the Stanford Cart

  ›

  Stanford University granted Moravec his PhD in computer science in 1980, and the vehicle that earned it could negotiate a room full of obstacles without a human hand on the wheel. The Stanford Cart was a TV-equipped robot remotely controlled by a large computer. It was a demonstration that a machine, given the right sensors and processing, could build a picture of a physical world and move through it. Moravec also developed techniques in computer vision for determining the region of interest in a scene, a practical tool for deciding which part of an image deserves a machine's attention. That work in spatial awareness extended further into what he called 3D occupancy grids, a new approach to representing physical space so that a robot could know where objects were and plan a path around them. Moravec joined the newly established Robotics Institute at Carnegie Mellon that same year, 1980, as a research scientist. He became research professor in 1995 and has held an adjunct professorship there since 2005.

• 3. Mind Children and Consciousness

  ›

  In 1988, Moravec published Mind Children, a book that traced Moore's law and then pushed past it into territory most engineers left to philosophers. His argument about consciousness, which he called the neural substitution argument, proposed something precise and testable in principle. If each neuron in a conscious brain could be replaced one at a time by an electronic substitute behaving identically, the biological consciousness would transfer seamlessly into an electronic computer. The conclusion he drew was that consciousness does not depend on biology and can be treated as an abstract, computable process. That argument appeared in Mind Children seven years before David Chalmers published a structurally similar case in his paper "Absent Qualia, Fading Qualia, Dancing Qualia," which is sometimes cited as the source of the idea. Mind Children also sketched a timeline for robot evolution. Moravec set the start of a new series of artificial species at around 2030-2040, a horizon he believed Moore's law made inevitable.

• 4. Forecasting the Mind Fire

  ›

  Moravec's 1999 book, Robot: Mere Machine to Transcendent Mind, extended his earlier predictions and introduced the phrase "mind fire" to describe a coming period of rapidly expanding superintelligence. To reach that conclusion he generalised Moore's law beyond the integrated circuit, applying it to earlier technologies and extrapolating forward. Arthur C. Clarke called it "the most awesome work of controlled imagination I have ever encountered," and said Moravec "stretched my mind until it hit the stops." David Brin praised the book as blending "hard scientific practicality with a prophet's far-seeing vision." Colin McGinn, writing for The New York Times, reached a different verdict. McGinn wrote that Moravec "writes bizarre, confused, incomprehensible things about consciousness as an abstraction" and that his speculations "spiral majestically into incoherence." The gap between those two readings captures the central tension in Moravec's work: the same ideas that inspired scientists and science-fiction writers struck some philosophers as fundamentally confused about the nature of mind.

• 5. The 2020s Prediction

  ›

  In a 1998 paper titled "When will computer hardware match the human brain," Moravec put a number on the question. He estimated that human brains operate at roughly a fixed number of instructions per second, and that if Moore's law continued its trajectory, a computer matching that speed would cost only 1,000 US dollars in 1997 terms by the mid-2020s. His stated conclusion was that computers suitable for humanlike robots would appear in the 2020s. The estimate was not a vague gesture toward the future. It was built from a comparison of measured computational cost against a projected price curve. Whether the curve has held, and whether humanlike robots have arrived on schedule, is the question his 1998 paper leaves hanging in the air for any reader picking it up today.

• 6. Seegrid and the Bush Robot

  ›

  In 2003, Moravec co-founded Seegrid Corporation in Pittsburgh, Pennsylvania. Its stated goal included developing a fully autonomous robot capable of navigating its environment without human intervention, the same problem he had chased since the Stanford Cart days, now applied inside factories and warehouses. Alongside that commercial venture, Moravec also developed the theoretical concept of the bush robot, a speculative design in which a trunk branches into smaller and smaller limbs, each capable of independent manipulation. He also worked on space tethers, a different class of engineering problem involving orbital mechanics rather than ground-level navigation. His path from a BSc in mathematics at Acadia University in 1969 to an MSc from the University of Western Ontario in 1971, and then to Stanford and Carnegie Mellon, placed him at the centre of several research communities at once, each shaped by a different set of constraints and possibilities.