2016 marked incredible technological leaps in driverless vehicles. From Telsa’s autopilot to Uber’s self-driving cars in Pittsburgh, to Google covering over two million miles in their vehicles. However, recent setbacks from players like Apple, Uber, and even Tesla, may indicate that the market adoption of these vehicles may be further than we think.
So, as a thought experiment, I wanted to apply some of the tech strategy patterns / frameworks, I’ve demonstrated on this blog to understand what may propel this trend into a “Slope of Enlightenment” in 2017.
Barriers to adoption
Removing the technology risk from this analysis (there are far smarter people working on those problems), let’s apply the Diffusion of Innovations framework to understand what may unlock these technologies for greater adoption. Of the seven factors listed out in the model, two stand out the most:
- Regulatory – clearly this will be a long, tedious, process, that is likely segmented by state or region.
- Compatibility – with so much infrastructure developed for drivers, how can we develop a model that fits with those existing systems?
Let’s explore each of these, and some potential solutions.
Overcoming regulatory issues by unbundling
One approach to tackling a thorny regulatory environment, is to take small, simple, incremental steps that are easy to say “yes” to, or to even turn a blind eye to. For instance, what if manufacturers instead of building “fully autonomous” vehicles, “unbundled” key use cases? Doing so, would take the scary idea of a self-driving car, and turn it into a series of benign, bite-sized functions, such as parking, changing lanes, highway cruise control, etc. In this way the driver would still be in full control, but move more into orchestrating how and when to enable each routine.
Following this approach, as Tesla has, lowers or eliminates the regulatory barriers. This enables regulators to “catch-up” to the technological shift, and provides them specific, bite-sized, non-controversial steps.
Compatibility with existing infrastructure
It’s been 60 years since President Eisenhower helped establish the Interstate Highway system, and since then we’ve accrued many systems, processes, and regulations firmly establishing them into society. So, modifying our current systems will be challenging. Once again, rather than taking regulation head-on the way Uber & Google have, asking for minor modifications to existing infrastructure is a much easier battle. For instance, HOV lanes can be enhanced to accommodate automobiles adhering to compliance measures. Similarly toll-highway sensors could be augmented to monitor consumer compliance. The net result could be cars using a new, “enhanced” cruise control whipping by at 100mph, in their own dedicated lanes.
Realizing the whole product
Often when new technology platforms arise, early players take a “full stack”, or vertical, approach to development by providing much of the whole product. We’ve seen this play out in this industry with the approach Google, Tesla, Uber, Mercedes, and others have taken.
That said, as industries mature, they move to more of a networked, or horizontal approach. The classic example is the Apple (vertical) vs. Microsoft (horizontal) in the PC business of the late-80s, and early-90s. Could we see similar approaches taken in the autonomous vehicle space, where a common operating system layer emerges, with standardized networking & communication protocols? Doing so would enable more specialist vendors to tackle various pieces of the technology stack, where common interfaces enable them to integrate into the OS layer. Similarly, vehicles themselves could communicate with one another to enable efficient traffic flows. So perhaps we could see the next Novell or Microsoft emerge for this platform?
Looking back and then ahead
It’s impressive to think how far and how fast the technology has evolved here to accomplish such an incredibly complex task. We are clearly at (or perhaps even past) the “knee” of the exponential curve, with a massively transformative shift ahead. However, while we can build something, we still need to apply these kinds of frameworks to better understand how these technologies may be adopted en masse.