Elon Musk has made one of his more extraordinary predictions. Within roughly three years, he says, Tesla's Optimus robots could become better surgeons than the best human surgeons, and not merely as a laboratory curiosity. He predicts that they could eventually be manufactured at such scale that great surgical skill becomes available almost everywhere.
Three years? I would not bet the farm on that.
But forget Musk's notoriously adventurous timetables for a moment. The more interesting question is what happens if he is directionally right.
Surgery has always been one of the ultimate scarcity professions. Producing an excellent surgeon requires a human being with sufficient intelligence and temperament, years of university education, medical school, specialist training, supervised practice and then years more experience. At the end of this immensely expensive process, society has produced precisely one surgeon.
That surgeon has another unfortunate characteristic. He is human. He becomes tired. His hands age. He can become ill. His concentration fluctuates. He cannot simultaneously perform operations in Adelaide, Alice Springs and Addis Ababa. Eventually, after society has invested decades in creating his expertise, he retires or dies and much of the embodied skill disappears with him.
A successful robotic surgeon would change the economics of expertise itself. Suppose an artificial system eventually learned from millions of operations rather than thousands. Suppose every unusual anatomical variation, complication and successful technique encountered anywhere in the network could improve every subsequent machine. Suppose its hands could manipulate instruments with microscopic precision without tremor or fatigue.
Then, once the design worked, society would no longer have to spend decades producing each additional surgeon. It could manufacture one. That is the genuinely revolutionary part of Musk's prediction.
We should nevertheless distinguish that possibility from the present reality. Tesla's Optimus is not presently wandering into operating theatres performing heart bypasses. Tesla describes its objective much more modestly: a general-purpose autonomous humanoid robot capable of performing unsafe, repetitive or boring tasks.
Musk himself has recently made an important admission about the present state of humanoid robotics. Many of the astonishing robot demonstrations appearing online are pre-programmed or remotely controlled. Nobody has yet produced a humanoid capable of performing the full range of generalised everyday physical tasks that humans perform almost without thinking.
Surgery raises the difficulty another hundred levels. A surgical robot would need extraordinary dexterity, perception, force control, reliability and the capacity to recognise unexpected situations. Real bodies do not behave like engineering diagrams. Tissue tears. Blood obscures vision. Anatomy varies. Equipment fails. Patients suddenly deteriorate. Surgeons make decisions while manipulating a living organism whose condition can change by the second.
Then there is regulation. A machine that can stack boxes badly is an inconvenience. A machine that cuts the wrong artery is a catastrophe. Musk's three-year claim should therefore be treated as a prediction, not an accomplished technological fact.
But suppose it takes ten years rather than three. Suppose it takes twenty. The implications remain enormous. Medicine is expensive partly because expertise is scarce. Australia can put highly trained specialists into major metropolitan hospitals, but providing the same expertise throughout remote Australia is enormously difficult. Poor countries face the problem on a much greater scale. There simply are not enough highly trained specialists to give everyone access to the world's best medicine.
Manufactured expertise changes that equation. Once a robotic surgeon becomes demonstrably as safe and effective as an excellent human surgeon, the important economic question becomes the marginal cost of producing another machine. The first system might cost billions to develop. The ten-thousandth does not.
That is the same economic transformation that computers produced elsewhere. A mathematical calculation that once required a skilled human calculator became effectively free. Information that once required travelling to a major library became available instantly. Artificial intelligence is beginning to do something similar to portions of intellectual labour.
Advanced robotics could extend that process into the physical world. Medicine would be an especially dramatic case because human medical expertise is simultaneously enormously valuable and enormously difficult to reproduce.
Imagine a remote Australian community possessing a machine with surgical competence comparable to a major metropolitan specialist. Imagine the same machine in an African village or an Indian provincial hospital. It does not demand a million-dollar salary. It does not refuse to move to the country. It does not need twenty years of education before treating its first patient.
And, most importantly, improvements could propagate. When a brilliant human surgeon discovers a better technique, other surgeons have to learn it. They attend conferences, read papers, watch demonstrations and practise.
A network of robotic surgeons could potentially receive an improvement as software. That possibility turns medicine from a problem of training scarce individual experts into something resembling industrial production.
However, medicine is not merely mechanical competence. A patient frightened before cancer surgery does not merely need an efficient meat-repair machine. Medicine involves judgement, explanation, consent, trust, uncertainty and human relationships. Some patients will continue to want human doctors even if machines eventually outperform them technically. But this objection should not be confused with the technical question.
We still play chess against humans even though computers became overwhelmingly better chess players. Human participation retains value because we value the human activity independently of optimal performance.
Surgery is different. If I am having a brain tumour removed, I am considerably less interested in preserving surgery as a traditional human craft. I want whoever, or whatever, has the lowest probability of killing me and the highest probability of removing the tumour successfully.
If machines eventually establish that superiority through rigorous clinical evidence, sentiment will have difficulty resisting them forever. And here the robotic surgeon connects with a much larger transformation.
A machine capable of surgery is not merely a surgical machine. The dexterity required to manipulate delicate tissue is potentially applicable to thousands of other physical tasks. A robot capable of navigating the complexity of an operating theatre is approaching capabilities useful in laboratories, factories, workshops and eventually homes.
That brings us back to the economic question I discussed in relation to AI EAT, in another blog essay today. Artificial intelligence threatens cognitive scarcity. Advanced robotics threatens physical scarcity. Combine them and something historically unprecedented begins to emerge.
We have traditionally assumed that even extraordinarily intelligent machines would still require human beings to build things, repair equipment, mine resources, operate machinery and perform delicate physical work. Humanoid robotics is an attempt to remove precisely that limitation.
The robotic surgeon therefore matters even if Musk's three-year prediction proves spectacularly wrong.
It represents the ultimate ambition of embodied artificial intelligence: not merely a computer that knows what an expert knows, but a machine that can physically do what the expert does. Once intelligence and physical competence can both be manufactured, the economics of human expertise changes fundamentally.
There is an optimistic conclusion. A child born in a poor African village might someday have access to surgical expertise superior to that available today to a billionaire. Medical scarcity could collapse. Procedures presently rationed by specialist shortages could become widely available. The world's accumulated medical knowledge could potentially be embodied in machines distributed almost everywhere.
That would be one of the greatest technological achievements in human history. But there is also the darker question. If we can manufacture the surgeon, eventually what other human occupations can we manufacture? The truck driver, miner, electrician, farmer, builder, laboratory technician and factory worker cease being protected merely because their jobs occur in the physical world. Artificial intelligence supplies the cognition; robotics supplies the hands.
That future is not here yet. Musk's Optimus cannot presently replace the world's surgeons, and Musk's timetable deserves considerable scepticism. But the important question is not whether Elon Musk has correctly guessed the year. It is whether the underlying destination is technologically possible. If it is, the factory-built surgeon will be far more than a medical curiosity.
It will be evidence that one of the last great economic monopolies of the human species: the ability to combine intelligence with extraordinarily sophisticated physical action, has finally acquired a competitor.
https://www.thefocalpoints.com/p/elon-musks-factory-built-surgeon