To see what will happen with self-driving cars …

… Just look at what it happening in the world of photography.

Digital camera sales have fallen off a cliff and are unlikely to rebound to the highs of 2008 to 2011.

Why?

Smartphones are a simple answer. Smartphone photography has gotten better and better in recent years, with advances in lens and sensor technology allowing them to almost compete with quality DSLR cameras.

It’s the next wave that has already started that shows us what to expect.

Computational photography is here, and it is getting better results out of technically limited hardware. With computational photography, you no longer need to understand the photography triangle to produce stunning images. You no longer need a fast lens (read: expensive) to produce sharpness and bokeh that make your photos pop.

And very soon, computational photography will produce compositions that are not far off those of all but the most talented. All in a device that were buying anyway because it does so much more than take photos.

The computational in the sentence is the most essential part.

Self-serving cars are the computational-driving of driving, and it will be a better driver than you, me and everyone else other than the most highly skilled and trained drivers. Manual car sales will fall off a cliff when it happens.

But as with cameras, boutique or specialised models will become valuable and highly sought after, allowing the development of niche markets for these products. There will probably be driving leisure centres springing up in or around large cities that, for a fee, will allow you to drive a “real” car for fun.

The industrialisation of the combustion engine did this for horses, digitalisation is already having an effect on photography and computation will have this effect on cars.

7 November 2019 — French West Indies

Matthew Cowen @matthewcowen