The order of magnitude future
December 9, 2011 1 Comment
Around 1970, the law known as Moore’s Law was formulated: the simple statement that processing power would double, and price halve (or such kind of guidelines) every 12 to 18 months. A phenomenal prediction at the time, but one that has held amazingly well so far. We see a new range of amazing processors come out every year, but recently, the push has been for parallelism rather than sheer speed of silicon: for density over yield and integrity. This does mean that markers such as transistors per chip go up, but has the problem that actual user experience does not necessarily go that way (after all operating systems and programs cannot be custom fitted to a particular architecture of parallelism while keeping general enough for consumers). The problem is: we have hit a wall: Silicon just isn’t fast enough. It pretty much tops out at around 8.3GHz (or so the over-clockers seem to have proved). Now while this might be pretty fast, it means that to get more power, one has to have more die area. To add to this we are hitting the limits of just how small you can make a single transistor: about 5nm. So what the world really needs is a completely new technology base: something that can go orders faster, orders smaller, orders lower power… And we are starting to see such technologies emerging in labs around the world, some are even beginning to make real progress, so much so we may see them in a decade or so…
A similar problem is that of space travel: its been pretty common for a few decades now but in no way is it cheap or easy, it takes thousands of tons of fuel to propel tons of satellite into orbit costing millions. To say take men off Mars or other feats, we dont need a little more energy and some clever design, we need loads more energy and loads less weight, we need systems which run off pretty much no power. We need an order of magnitude.
Labs at the moment, like for semiconductor design, are also discovering new materials (weirdly based off carbon again) which give strength, weight and properties way higher than what aluminium and carbon fiber currently delivered. But this brings with it a whole new set of accessibility problems: Who could drill something as strong as diamond with a hand drill? Who could grow crystals of insanely complicated solution in their garage? What start up could afford very high cost manufacturing just to keep up with big companies? At the moment the technology industries are split, but in hardware enthusiasts can mostly match big companies in terms of materials, maybe not in reproducability, but in ease of prototyping certainly.
So what will the future of engineering hold? Probably a few very able companies with patented super materials, licensed to medium-sized manurfactures…Looks pretty bleak, especially thinking how many massive developments have come from little companies.
Energy use too needs to be massively reduced if we are to continue living. The current trend is to insulate a bit more, make some renewable power stations and charge more for fuel and these approaches gets one some way towards the targets set. But it doesnt go anywhere near the the problem that the amount of energy and material the average person in a developed country uses is way, way more than is ever sustainable. In energy we need an order leap to the user using less, not reducing the impact of each unit, but reducing the units. Our way of living is totally unsustainable and we need to reduce impacts by tens or even hundreds of times.
I leave you with this thought: “If humanity survives the next 300 years, we will survive the next 1500”