I think that, rather than the self-driving automation itself, the bigger disruption to the nature of the auto industry is likely to be the distributed impacts of electric power and the reliance on sophisticated electronics. The simplification, increased reliability, and modularity that imposes. Consider the workers at your typical auto repair shop. Are these people likely to maintain the sophisticated electronics, opto-electronics, and mechatronics of tomorrow's robotic vehicles? Apple requires its service staff to actually travel to California periodically and take series of classes to be certified to repair their products, and there’s generally very little left on them that can be readily repaired. Often —in large part because of the ubiquity of surface-mount electronics parts— they have no choice but to ship things to centralized facilities with specialized tools for lower level repairs —which they now discourage in favor of whole replacement with notoriously scary repair prices. (which the media is now starting to give them heat over) That’s common with all the mobile devices now; phones, tablets, laptops. With most electronics now bought through general retailers or Amazon, most of the manufacturers have no dealerships of their own and tell their customers to ship broken items to distant centralized service centers for any diagnostics and repairs.
Can you imagine that working with cars? Doesn’t seem likely to me. So I think companies would be compelled to compensate for the increasing sophistication of cars by compartmentalizing and containing that complexity through modularity. The car will have to physically evolve into something akin to the old desktop PC to remove the local need for skills and tools to make repairs. Imagine a platform or space frame chassis into which everything functional is plugged-in —like a motherboard— over which is a bolted-on 3D printed plastic body shell. Controls are all drive-by-wire. Electric motors integrate into the wheels eliminating any mechanical drive train and making all-wheel steering the norm. This could make cars so physically simple that it makes more sense to build them to order at dealerships than to ship them whole. At their size, they are far more cost-efficiently shipped as parts, its easier to support more options to customers (and state or national ‘localization’), and the more generic core components are, the less losses incurred by the market failures of individual models. The more generic your inventory, the longer its value holds.
But this has the potential for the same kind of effects as the design of the IBM PC had. The potential for parts development to disperse among international communities of competing ‘OEMs’ (original equipment manufacturers), accelerating their improvement, and the core architecture of the most common cars to then be appropriated by cartels of those OEMs and made universal like the PC. Thus we could see brandless cars much like brandless PCs. People repair, customize, upgrade, or even build them at home to save money, ordering their parts online. Luxury car companies position themselves like Apple. Use the same tactics. Engineer around proprietary software, exclusive tech, and design IP created by celebrity designers. Create an upper-class of dealership/service centers and specially trained service people in lab coats instead of overalls. But most of the cars in the world become internally generic, wildly varied and personalized in design, and near-disposable cheap.
With the home, an ability for movement depends largely on the necessity and frequency of adaptation of structure. This is more useful in a nomadic context where such physical change of a structure is very frequent —the frequent deployment and packing-up of a dwelling as it is frequently moved from place to place. The less frequently a structure is changed or moved, the less the cost of the mechatronics and mechanics for automating that movement can be justified. How much might a house need to change, day to day, season to season? There are a lot of self-deploying structure concepts, such as shape-changing Hoberman structures or self-mobile interlocking panel, frame, or block robots that traverse over each other to find their positions.
In another context, a fixed home might be made a robot in order to physically assist its occupants. This is the general idea of contemporary home automation, though currently it is limited to the control of stationary systems; heating and cooling, lights, security systems, media systems some ’smart’ appliances. But it can be taken further. We typically think of the home robots as a self-mobile assistant device. But this suffers many limitations. Self-mobile robots are very limited in their on-board intelligence and power. Thinking about this in a context of home care for the elderly, the Japanese long ago took a very different approach. Making a self-mobile robot with the mass and strength to physically assist a disabled person —help them into and out of beds, up and down stairs, help them put on clothes, feed them, etc,— is really tough. And the typical decor of homes is a problem. Robots still get stuck in the carpet. So instead of big, heavy, self-contained self-mobile robots they considered the idea of robot manipulators suspended from an overhead monorail. This would give them unlimited electric power, unlimited off-board intelligence, great physical leverage, the ability for special-purpose manipulators to team-up when necessary, and easier access to all the surfaces of a home. This would be combined with machine vision and voice systems distributed about the house. In effect, the whole environment of the house would be made into a personal robot. Unfortunately, working on this in the late ‘80s, such large elaborate installations were extremely costly and required considerable modification to the conventional wood structure of the home. The idea did, however, inspire the design of the very practical robot Gerty in the movie Moon.
The one logical inconsistency with the Gerty design was that it featured a mobile primary user interface unit that contained the core of the robot’s intelligence. This makes no sense, except that it allowed for a physical character presence to the robot cinematically and dodged the HAL trope. In practice, if such a user interface unit was used at all (as opposed to just a voice), it would likely only need an expressive touch/gesture display with integral machine vision and depth sensing —a ’smartscreen’— to host an animated graphic character. (something I anticipate today’s Google Assistant and Amazon Alexa to evolve, once they figure out that personality is the real application...)
Another take on this idea is that of the micro-cellular robot, where cellular robots can be made small enough (high enough in resolution) that they can self-assemble to form not only structure but a variety of furnishings and machines within that structure. And thus a home —along with just about anything else a person might need— could be a continually self-transforming robot swarm. We’re far off from that sort of technology, however.
I tend to see the structure of a home more as a ‘backplane’ to facilitate the ‘nesting impulse’. We’re apes. Apes make nests. Homemaking is a very fundamental form of human expression, which has been stolen from us in modern times by financial and property systems compelling our relinquishing of this basic human act to a series of experts and professionals creating heavy, permanent, non-sustainable structures for the sake of imposing life-long debt on us for a very basic human need. So I’m less concerned about automating the creation and adaptation of the home than I am about simply facilitating it at a solitary unskilled human scale. I often say that, in the 21st century, there’s no excuse for anything in the home still needing more than one person to move around —including the house itself. Automation is certainly very fun and overcomes many limitations of the solitary human in the building of large heavy structures, but I question the need for those large heavy structures in the first place in an age of high-performance structures and materials. This is why I’m sort of obsessed with the pavilion; the minimal shelter that was the basic human shelter of most primary cultures once we started making things and now can perform just as well as tons of rock, wood, or adobe. Of course, in the urban environment the compulsion to density compels a reliance on heavy materials again, but even there we don’t need to sacrifice the nesting impulse. So my urban visions tend to be based on parametric superstructures as urban landscapes designed for free individual retrofit. Big backplanes.