Moore’s Law is dead.
Intel co-founder Gordon Moore observed in 1965 that the number of components on an integrated circuit was doubling every year. He predicted that this growth would continue for another decade. In 1975, he revised the forecast to every two years.
In simpler terms, it can be said that computing power doubles about every two years, while the cost remains the same. Incredibly, this exponential rate of growth held true for more than 40 years. For the most part, you could expect that the computers available at any given time were about twice as powerful as those two years prior, while the cost was about the same. All told, computing chips available now are more than 2 billion times as powerful as those available in 1965, while the cost is about the same.
Our entire understanding of how technology works is based on this model. If it’s more than a couple years old, it’s probably time to replace it. In schools, we use Moore’s Law as a guideline for everything from planning replacement schedules to estimating depreciation. We replace desktop computers every six years. That’s the point at which they’ve lost about 90% of their value. It was predictable. We could plan for it. We could budget for it.
I got my first smartphone in 2010. I replaced it in 2012. That one would have been replaced in 2014 if I hadn’t dropped it in 2013. I’m making this third one last an extra year to get back on track, but it’s definitely showing its age now.
But Moore’s law is dead. Even the chip manufacturers have acknowledged as much. We’re still going to see growth in computing power, but that growth is going to be slower and less predictable. What’s that going to mean for schools? Lots of things:
We will keep computers longer.
We have to get over the idea that computers that are a few years old are too out of date to do anything useful. Most of the computers in our classrooms are now eight years old, and it’ll probably be another year before we replace them. It’s not that we’re trying to be cheap or that we don’t want to be cutting edge. The reality is that they still do most of what we need them to do, and we’d rather use technology resources to improve access for students.
We have to worry about durability.
When we bought the Acer netbooks in 2012, we expected to keep them for three years. We knew there would be problems with broken keyboards and cracked screens, but they were half the cost of desktop computers. If we could get three years out of them, we could leverage their mobility and still come out ahead.
But as we head into year five, those weak batteries and poorly designed keyboards are becoming more of a problem. While any life we continue to get out of them is just icing on the cake at this point, it’s a shame to throw them away when they still work fairly well. We’ll take the worst ones out of commission and use them for parts, and we’ll limp along for another year before retiring them. But as we move forward, we need to think about holding on to these things for more years. Build quality and durability will become more important, and we will have less patience for planned obsolescence.
Software will have to be more efficient.
In the early days, computer engineers were all about efficiency. They were working with some pretty tough constraints, and they would spend a lot of time working through performance and resource challenges. That art has largely been lost in the last generation. Applications use more memory / processing power / storage space to do the same things, because those resources have been unlimited for so long. There’s no reason why I should need 8 gb of ram in my computer to run a web browser and a terminal session. But I do because the software is developed without any consideration for hardware limitations.
When Windows Vista came out, there weren’t any computers on the market that could run it. In fact, Microsoft changed the certification from “Vista Ready” to “Vista Capable” so they could actually certify computers to run the new operating system. Within a few months, the hardware caught up, and soon just about everything could run Vista (whether they wanted to or not). The same thing is happening right now with the Oculus Rift. Very few computers meet the system requirements. So the early adopters have to buy new hardware, while everyone else will just have to wait for the industry to catch up.
But in a world where those hardware upgrades are NOT just around the corner, the software developers are going to have to find better ways to improve their products without boosting the system requirements beyond reach.
Costs are going to rise.
I expect costs to go down. That’s how we fuel sustainability. In 2011, we had 3 computing devices for every 7 students. Today, we have 11 computing devices for every 7 students. Over the same period, the school district has received no increase in funding. We’re not bringing in more money, and we’re not spending more money. We have moved things around a little. We’re spending a little less on textbooks, and a little more on computers. But for the most part, the financial side has been pretty flat.
The difference has been the cost of computing devices. When we started buying classroom sets of laptops, we were paying about half as much as we were for desktops. When we started buying Chromebooks, we halved that cost again. We didn’t spend less money, but we bought a lot more devices and improved student access to technology considerably. That, in turn, allowed teachers to better leverage technology to design instruction that meets the individual needs of each learner.
But this year, when I placed my order for Chromebooks for the incoming sixth graders, two things surprised me. First, the specs on the new devices are identical to those from last year. There’s no more memory or storage or processing power. The device is exactly the same as last year. Second, the price hasn’t dropped. Usually, if I buy something for $350 one year, I expect it to be $240 the next. Not this year. The new devices are a few dollars cheaper, but innovation has stalled and pricing is staying the same. That means the whole industry is slowing down. While the longer life cycles are going to help us keep technology longer, the cost stability is going to offset any savings we might have had.
The biggest change is going to be the mindset. The promise that there’s something newer and better right around the corner is a myth. We have to get over our fascination with the new and shiny, and focus a little more on doing great things with the amazing technology we already have.
Moore’s law chart by shigeru23 from WikiMedia Commons
Silicon Photonics 300mm wafer by Ehsanshahoseini from WikiMedia Commons