The Heart and the Chip: Our Bright Future with Robots

Daniela Rus runs MIT's Computer Science and Artificial Intelligence Laboratory, has a MacArthur genius grant, and has spent decades building robots that navigate the gap between what computers can compute and what hands can do — which makes her the right person to tell you why the robot apocalypse isn't coming.

*The Heart and the Chip* is structured as a three-act argument: Dreams (what robotics could become), Reality (where the field actually is), and Responsibility (what we need to do about it). The first act is where Rus is most infectious. She describes robots built from soft materials — fabrics, gels, biodegradable polymers — that can be swallowed to perform internal surgeries, exoskeletons that restore mobility to people with spinal injuries, swarms of insect-scale machines that can navigate disaster zones. The central claim is that robots are best understood as human amplifiers rather than human replacements: tools that extend what a surgeon's hand can accomplish, not machines that stand in for the surgeon. The title's metaphor means that directionality matters; human intention leads, the machine executes.

Where the book earns its keep is in the Reality section, which drops the futurism and explains the actual engineering problems. Getting a robot to fold a shirt is harder than getting one to Mars: the physical world is chaotic in ways that resist formal modeling. Rus is good on why — the sense-process-act loop, the challenges of dexterous manipulation, the energy costs embedded in training large models. She cites a 2019 study showing that training a deep learning model can generate the carbon equivalent of five cars' lifetime emissions. This is the kind of specific, honest accounting that gives the optimism weight rather than just warmth.

The third section, on responsibility, is where the book loses its footing. The arguments against job displacement are weaker than the technology sections deserve. Rus points to individual firms hiring more workers after automation, which doesn't actually address what happens across entire industries when displacement outpaces retraining. Her recommendations — a robotics Hippocratic oath, better computational education for children — are reasonable and thin. The problems she identifies require structural answers; she offers vocabulary.

For a technical reader already inside AI literature, this book sits at a frustrating level of generality: broad enough to survey a lot of ground, not deep enough to test your assumptions on any one piece of it. For someone coming in with limited robotics background who wants to understand why this technology matters and where the genuine hard problems lie, it's useful. Rus knows the field from the inside, and that shows most clearly when she's explaining what isn't solved yet — which is rarer, and more valuable, than the optimism.