The future of…

Click here for contact details, other articles and personal details

The future of engineering

Ian Pearson, BTexact, March 2004

It is an exciting time to be an engineer. We are changing the world faster than ever before, and the pace is accelerating. Engineering is subject to a positive feedback loop, which is becoming more prominent as time goes on. The better we make the tools, the faster we can make new and even better tools. This feedback loop has applied especially well in information technology, where faster and better computers and software have continuously accelerated the next generation. Eventually, we won't even be limited by human intelligence, since computers will become smarter than us, and design even smarter computers still. Then we will be on a very rapid acceleration indeed. Some of us are already looking forward to the rapid advance of capability called the Singularity, where technological development is so fast that it is like aliens landing and giving us all the gadgets out of their spaceship. Engineers disagree on when this is likely to happen, but with estimates varying from 2015 to 2040, that puts it within the working life of today's engineering students. And of course, the last several years before the singularity will involve very little input from people at all, being almost entirely machine driven. Engineers look forward to the new toys and capabilities, but most of us aren't looking forward to being cut out of the development cycle.

However, I would argue that the singularity is more likely to happen towards the near end of the band, perhaps around 2020. Recent progress in the fields of nanotech and biotech suggest that we will be able to assembling computing devices from molecular sized components using bottom up assembly, rather than lithography by 2010 or soon after. Much of the assembly may be done by custom designed proteins in living cells. Even if we are still mass-producing most chips by other means then, being able to make some prototype in this new way will bring large rewards. Making computers out of very small components, especially if there are large numbers of them, will produce very high processing capability. We might soon be able to make some serious steps towards conscious computers. Certainly, it should be possible to suspend trillions of optical neurons in a transparent gel, communicating with millions of wavelengths. By using evolutionary development and learning techniques to determine the appropriate structures, there is a very good chance that such a device could reach superhuman levels of intelligence, soon after 2010. Even the raw number crunching capability would be formidable by today's standards, let alone its likely ability to outthink humans in every department by several orders of magnitude. Tools such as these will greatly accelerate R&D engineering development.

What is far from certain is how much human involvement will still be required at any date in the future. Along the way, machines will give us increasing assistance in everything we do, increasing the productivity and capability of engineers, allowing us to do more and more with the same numbers of people. The engineer gradually becomes more of a system designer, with the detailed analysis and design being increasingly automated. Since this role is already fairly well staffed by industrial designers, the ultimate role of engineers will be challenged - we could quickly be reduced to a supervisory or managerial role. However, if machines can be built that are far smarter than people, we might conclude that people would almost all become intellectually irrelevant. We could just ask for something, and machines would do the research, design and fabrication. In most cases, even our needs would be anticipated so we shouldn't even need to ask very often. Engineers would have no more value in this process than ordinary people. We are already proceeding quickly along this road. How far we go will determine how many engineers we will need and what their role will be, but it is certainly clear that we are in danger of ultimately engineering our profession out of existence. With rapid development of ultra-high technology, we may not feel too concerned by this. Living in a Star Trek world might be adequate compensation for our enforced leisure, so long as socio-economic engineering ensures that we have adequate funds to enjoy it.

On the other hand, future engineering machines might be more like a musician's keyboard. The waveform synthesis, rhythm backing and so on can all be generated by the keyboard, but we still compose the musical piece (let's leave computer generated music out of it for the time being). Perhaps future engineers will still do the human side of engineering, making things appeal, adding the touches that make it all worthwhile, coming up with a really elegant solution once in a while. If the mathematics is done by machines, so what? We can hope that human engineers will always have a special kind of flair that is different, even if not superior to what machines do. We can hope, but I wouldn't bet on it.. Enjoy engineering while it lasts, but make sure you cultivate some human skills on the side as you go along. The ultimate consolation of course is that the range of new technology that we can invent and design in the next two decades will be more than all our ancestors did, so we will at least have something to look at and be proud of. If we feel too many withdrawal symptoms, we can hop into a virtual environment and engineer away at virtual technology until they go away.