I was in the US last week for a get together with all the DFJ folk from the 19 partner funds in the network, and as part of the proceedings Steve Jurvetson gave a presentation on innovation in which he showed the following chart.
It shows that Moore’s Law which has underpinned improvements in computer and network performance for the past four decades (including PCs and the internet) is the fifth such law. By generalising the definition from transistor density to the cost of processing power ‘computing’ performance has been on an exponential improvement curve since 1900 – doubling every two years.
There are two obvious conclusions from this:
- There is no reason to think this will stop any time soon, innovation and progress will continue, even if Moore’s law comes to an end. This article gives some thoughts as to what might be next.
- The progress over the last 100+ years has been even, continuing at the same rate through periods of economic boom and bust
One other point that Steve made which I liked concerns the nature of exponential curves (and forgive me for getting a bit geeky here). He noted that if you plot an exponential curve on a linear scale it always looks like you are just past an inflection point, but if you fast forward a couple of years, stretch the y-scale accordingly and plot again it will give the appearance that you previously called the inflection point too early. Hence it is only by plotting on a logarathmic scale and seeing a straight line that the true picture can be seen.
Companies attributing their success to being past an inflection point in something therefore need to be careful they are plotting on the right scale.