Exponential thinking

Hi Future Fans,

There has been a ton of great news out there in the last weeks. I will start posting some of the content I’ve earmarked, but today I wanted to touch on this useful & interesting Singularity Hub article about how exponential progress works and how hard it is for the average person to think about it correctly.

https://singularityhub.com/2018/07/15/why-most-of-us-fail-to-grasp-coming-exponential-gains-in-ai/#sm.0000gd3m7uhzrd45seh1949ynbo4v

The reason I’ve chosen that article is that most of the really interesting stories and advancements I talk about here are in areas that are experiencing exponential progress. The simplest example is the growth in computing power, where there has been a steady trend of doubling computing capability every 2 years or so. Think about that for a moment – roughly every two years we will have doubled the entire capability that we developed over the history of computing. That is rather ridiculous and is also a very well established trend.

Many of the exponential trends I cover (e.g. robotics) are fueled by underlying computing power increases, plus advancements in the way things are done (like different computing architectures), plus advancements in related fields (like AI). These different influences can layer up to drive progress at truly breakneck speed.

Thinking exponentially, the early days of VR are simply early prototypes on a clear exponential journey to fully immersive, better-than-real, virtual worlds full of boundless potential. The early steps in genetic engineering are simply early prototype steps towards fully customised genomes, full control over the physical coding of nature, and synthetic biology. The early steps in AI are simply early prototype steps towards artificial intelligence that will vastly out-perform our natural human minds. These things are inevitable in exponential systems and a couple of decades in time delivers stunning, seemingly magical progress.

The article talks about exponential progress always looking the same at any point in history (i.e. it will always look like ‘hockey stick’ progress will be gained in the next few years). Whilst technically true, the huge difference around our own time (now) is that we have already come so far and the baseline for growth has become so substantial. For example, the next great leap in supercomputers will deliver exascale computing (probably around 2021); that’s a billion billion calculations per second. To put that into context, every person on earth would need to do a calculation per second for over 4 years in order to match a second of calculations on this machine. Then, in a couple of years after that, we’ll probably double it again, then again, then again, then we will find a whole new way of doing it, then double again; rinse, repeat. That is truly meaningful progress that will utterly re-shape the world in which we live … exponential progress will propel us somewhere completely new, and it’s not going to take a particularly long time to do it.