During the Manhattan Project, while working to build the first atomic bombs, several mathematicians came up with a calculation that warned a critical chain reaction which unleashed the power of the atom might not stop until the entire world was destroyed.
The scientists' fear was allayed when their colleagues showed that while such a nuclear doomsday was possible, the probability of it happening was close to zero.
Because the project was highly classified, the public would not know about the doomsday debate for decades. But super-secrecy has its virtues - the warnings and rebuttals were all dealt with by experts, so public hysteria and the imaginings of sci-fi writers and Hollywood filmmakers never intruded.
A similarly apocalyptic warning was sounded about the nascent nanoscience and technology in 1986, when US futurist K Eric Drexler imagined that tiny self-replicating robots - built to manufacture products from clothes to computer chips - might reproduce uncontrollably and consume everything in their path for energy until the entire world was turned into 'grey goo'.
This scenario was modified by bestselling US science fiction writer Michael Crichton two years ago, when he published Prey, in which out-of-control nanobots, programmed for predatory behaviour and self-replication, turned to killing their human creators.
Nanotechnology has been plagued by public fears and Hollywood paranoia from the start. But while nuclear physics was already a mature science in the 1940s, today's nanoscience is still in its infancy. There is a gulf between what people expect from it and what is actually on offer.
The Hong Kong University of Science and Technology (HKUST) leads the world in some areas of nano-research and consumer product development, and these projects give a good idea of what is currently available.
But first, what exactly is nanotech? Well, one nanometre (nm) is a billionth of a metre - about the width of 10 hydrogen atoms, or 1,000 average-sized bacteria, laid side by side. It is also the exact diameter of a sugar molecule - a minor discovery made by a young Albert Einstein in 1905, his so-called annus mirabilis.
So nanoscience can refer to any scientific study involving objects measured in nms, and nanotechnology to any practical application or manipulation of objects on that scale. What is intellectually exciting - and virgin territory for new discoveries - is that nano-sized objects are too large to follow exactly the laws of quantum mechanics, yet too small to be completely governed by the classical physics and chemistry that rule the macro world with which we are familiar.
Kwok Hoi-sing, the head of HKUST's Centre for Display Research, leads a team that has developed the world's fastest liquid crystal display, which eliminates shadows and shades on screen to rival the high-resolution images of plasma television. A nano-technique developed by the team enables the crystals to align quickly to form moving images. The centre is currently in talks with several LCD producers overseas to license its technology.
'The LCD TV world market is worth billions a year, so just taking a very small slice of that would be quite significant,' Professor Kwok said in a Post interview.
Back in 1991, Terminator 2 featured a seemingly indestructible, liquid-metal android called T-1000. The university has also produced a 'smart liquid'. University physicist Wen Weijia and his team have created an oil solution with suspended nano-sized silicon particles which can solidify in a split second when an electric field is applied to it. Between its liquid and solid states, varying degrees of hardness can be generated by varying the intensity of the electric field, making it perfect for such things as car suspensions - without the usual wear and tear.
Meanwhile, a chemical engineering team at the university is working to produce drugs stored in uniformly tiny particles that can be absorbed through the skin, a project that may make the hypodermic needle obsolete.
And overseas, nano-sized zinc oxide particles that deflect ultra-violet light at optimal angles are being used to make sunscreen.
Exxonmobil uses tiny zeolites, a type of mineral, to break down large carbon molecules to form gasoline, while IBM has built nanoscale layering into disk drives to increase data storage capacity. Almost all of this takes what researchers term a 'top-down' approach, while the real glamour and danger is in working 'bottom-up'.
Chun Zhang, technical manager of the university's Institute of Nanomaterials and Nanotechnology, says that bottom-up nanotechnology is years, if not decades, away. 'Talk of self-assembly, or self-replicating robots refers to what we call the bottom-up approach - something in the future,' he told the Post. 'What we can do now is mostly top-down - mass-producing nano-scale products at relatively cheap prices. There is a timeline for this [bottom-up] technology to develop, and people are actively pursuing it, but we are not there yet.'
Nano-scientists are exasperated - they need public interest to sustain government funding and private investment, yet find it hard to explain what they are doing, according to Neal Lane, once science adviser to former US president Bill Clinton, who was speaking at a nano-tech conference at HKUST last month.
Also, the majority of nano-scale products and research are nowhere near as interesting, glamorous or hazardous as some people imagine.
'There is always the danger of hype, that you promise too much and deliver too little,' Professor Lane said. 'It's almost impossible to strike a balance ... when you have Michael Crichton writing about nano-particles eating up the world.' Currently a physics professor at Rice University in Texas, Professor Lane has helped to spearhead the United States' nanotech initiative, a project begun during the Clinton administration that has managed to sustain about US$1 billion a year in federal funding for American researchers in the past three years under US President George W. Bush.
The excitement centres around tiny, extremely rigid graphite cylinders called nanotubes, which have unusual electrical properties.
Nanotubes are being used by the golf-loving Japanese to make more durable clubs, while General Motors is mixing nanotubes with paints to make them adhere to the surfaces of cars more readily when the tubes are electrified. But they have far more exciting applications, including the potential to replace semi-conducting silicons to make transistors and wires. Because of their tiny size, they promise incredible shrinking circuits that would replace microelectronics with nano-electronics, making computer storage capabilities and processing powers virtually unlimited.
The entire contents of the US Library of Congress, for example, might be stored in memory chips no bigger than a pinhead.
With nanotubes, we are finally entering the territory of Messrs Crichton and Drexler.
HKUST is not the first to produce nanotubes, but it can claim to have made the smallest in the world using a technique pioneered back in 1999 by two of its physicists, Tang Zikang and Wang Ning.
Most scientists say that making computer circuits small and dense enough is the key to building a machine that can think of its own accord. Silicon-based microelectronics, they agree, will never give rise to circuitry complex enough to allow the emergence of real cognitive abilities to rival our own. Nano-electronics is likely to make this possible.
Also speaking at last month's nano-tech seminar, Matthew Tirrell, the dean of engineering at the University of California in Santa Barbara, referred to mother nature as the perfect bottom-up nanotechnologist from which researchers should draw inspiration.
He was talking about genetic processes by which DNA transcribes into RNA, which in turn translates linear strands of amino acids which then fold at specific angles to direct the formation of new three-dimensional proteins.
Scientists only recently came to appreciate this protein-making assembly line in which errors are rarely made and, if they occur, are quickly identified and corrected because they damage or kill healthy cells. Protein misfolding is the root cause of such diseases as type II diabetes, Alzheimer's and Parkinson's diseases, cystic fibrosis, mad cow disease, and some types of cancer.
'Inspired by molecular biology, studies of advanced nanotechnologies have focused on bottom-up construction, in which molecular machines assemble molecular building blocks to form products, including new molecular machines,' wrote K Eric Drexler in a recent article in Scientific American.
'It would be a natural goal to be able to put every atom in a selected place with no extra molecules on the loose to jam the works. Such a system would not be a liquid or gas, as no molecules would move randomly, nor would it be a solid, in which molecules are fixed in place.
'Instead this new machine-phase matter would exhibit the molecular movement seen today only in liquids and gases as well as the mechanical strength typically associated with solids. Its volume would be filled with active machinery.'
Mr Drexler says such molecular nanotechnology is one to three decades away, but we can already see glimpses of what it might be like. Back in 1994, US computer scientist Leonard Adleman used the base-pairing of DNA to build a computer to solve the so-called travelling salesman mathematics problem, used by phone companies to construct efficient routing, and by energy companies to build power grids.
Most experts believe DNA computing will probably merge with nano-electronics and information technology. The late evolutionary biologist Stephen Jay Gould liked to quote the Marxist-Hegelian law - that great quantitative changes may produce a quantum change in quality - as a metaphor of scientific progress. Nanoelectronics may be just that - a leap so vast that we may be face to face with our cognitive successors in this century.
And that is a truly scary thought.