Previously I considered 2 independent single-direction corkscrews. No particular reason. On reflection, I have a hard time seeing a justification for 2 separate corkscrews when you could combine them into 1 double-corkscrew. For human drivers, this doesn’t make as much sense. People have to be able to look at and understand the direction of traffic intuitively. The automated car dispatch algorithm has no such constraint, so we should consider how that affects the potential designs.

This tit-for-tat argument for/against could go on quite a bit, but I will cut it off here because it’s too abstract. Better to start out by painting a picture.

Accepting the above reason for why we don’t combine vertical transportation between multiple buildings today, we can consider what is different for self-driving car ramps. The cars can eliminate the inconvenience of additional horizontal walking distance by simply driving it. More importantly, the ramps can operate in a continuous fashion without the logistical constraint of a two-way track with no passing allowed between cars. That means that one ramp can have huge capacity, pushing toward consolidation.

The strong reason — Elevators have a significant capacity constraint in addition to a logistical constraint. The logistical element is why a significant building might have 6 elevators all right next to each other (instead of building 1 shaft with an enormous car). Since the logistical constraint demands multiple elevator cars in order to speedily serve one building, it makes no sense to consolidate shafts from multiple buildings with walkways going between them.

We rarely (never?) see this method employed in the world today. One building does not avoid building an elevator simply because it can make a walkway to the next building over which does have an elevator. We need to dive deep into the logic of the world that we live in.

While true, these points don’t help buildings that have a fairly small volume of foot traffic. Those places also don’t have much space to spare. Thinking further into it, I realized there’s another component that comes into play.

A ramp suitable for a 2-ton car to drive up is consumes much more floor space than an elevator made strictly for humans. My answer (incomplete at the time) was that two characteristics makes this (admittedly larger) structure able to accommodate vastly more traffic.

I should qualify that what I proposed wasn’t a “car elevator” but a car ramp. I prefer to assume the redditor knew this but I do not know. Otherwise it’s a good point, and there was separately some thought-out discussion on it as well.

Most importantly, I need to revise the original design in order to address space efficiency issues. Here is a comment from Reddit that hit on the one logical flaw of the argument that I made:

This is an ad-hoc followup to a previous story that I wrote. Primary thesis: dumb car ramps will replace elevators because of self-driving cars. This was a pretty good futurist thesis in my book, because it contained this mix of ingredients:

I drew the first “single corkscrew” as a design where there is two-way traffic. The other two designs could use one spiral as the “up” road the other other as the “down”. The dispatch algorithm might make two-way more viable that it seems, because it is aware of the locations of all the cars and can orchestrate a change in direction seamlessly.

How about scaling? A typical car park ramp slope is 1-to-6 in terms of rise-to-run. We could argue some margin to use steeper slopes because the dispatch algorithm has better control of conditions. As the number of floors per corkscrew turn increases, the diameter of the construction does so as well.

Building Connections

In practice, we need a marriage of the ramp designs with the physical location of related building. This isn’t totally trivial, consider a situation where would want to use a low-throughput single corkscrew to serve 2 buildings right next to each other. Just imagine it in-place and start counting the floors… and you run into a problem.

single corkscrew, two building — incorrect configuration

It’s difficult to predict an exactly perfect configuration for the basic parameters of single-corkscrew two-buildings. Maybe it would involve extra scaffolding outside of the ramp itself. The below top-down view would work, although it certainly feels expensive.

Revision that includes ramps and walkways

Here, we resolve the problems by simply adding additional car ramp area and walkways. If that’s going to almost double the size of the construction, then one might reasonably argue that just going with a double corkscrew might be easier. Let’s say we use the “double corkscrew” and connect 2 buildings.

double corkscrew servicing 2 buildings

There are some interesting logistical implications of making both criss-crossing ramps one-way. Let’s say that the two buildings are A and B. In practical terms, you may need to get to floor 3 in building B. If there are 7 floors in total, it may be the case that the car needs to drive the full 7 floors up, and then back down 4. This would be a uniquely annoying experience for people in building A floor 2, who would have to wait through many floors of driving.

Still, this design doesn’t fully answer the question of “exactly where do the cars stop?” If they stop directly on the ramp (in the flow of traffic), then we’re going to have a new throughput bottleneck. If there is additional drivable space to the side to act as a platform, then that’s more expensive.

Full Design Incorporation

I would argue that all of the above configurations look like the ramps are afterthoughts. Here, I want to talk about a distant-future case of heavy use of the ramps serving a very large building network. This starts with the “double corkscrew with spacing”, and it is my favorite. For each turn of the corkscrew, this design rises 4 floors. First consider the following space utilization concept from a single-floor perspective.

Space availability for a double corkscrew with spacing

Because there is space between the car-ramp corkscrews, there is empty space in a horizontal cross-section. This would become available space for… stuff. I have a hard time picturing the ramps fully wrapping around buildings, but it’s not hard to imagine a central area that is something like a transportation (or even commercial) hub for the surrounding buildings. Here is some detail for that vision:

One envisioned breakdown between office space and intermediary space

I’ve placed extra road space in the center of the corkscrew to act as the platform for humans getting on and getting off. From there, there’s a direct walking line on level ground to where people might be going in the surrounding buildings.

To me, this doesn’t “feel” much different than modern building design. There’s a certain cadence to the disembarking area that invites office plants and vending machines. After you get in the car to leave, you disappear into a hole in the ground. Kind of feels like the future.

Horizontal Travel in the Sky

Will we have increasing sizes of Skyscraper clusters as a byproduct of a construction automation boom? I’m inclined to think so. Will self-driving cars hasten the trend toward more massive urban centers or slow it down? That’s not settled, but my personal opinion is that it will hasten further mega-urbanization. When I think of a cluster of skyscrapers, I think of the city depicted in Ghost in the Shell.

The artwork triggered all kinds of assumptions in my head, about how all those buildings in the city center must tower even higher than the record-holders of today. This fictional place must hold so much that any transportation systems of today would collapse under the load.

I can picture the above double corkscrew pattern occurring in many different forms inside such a futuristic downtown. There may be a non-trivial amount of trips that use horizontal roadways on the 40th floor, without ever going back down to ground level. In this system, capacity is king. Corkscrew ramps need not exist for every block, and major up-and-down throughways can be shared among large parts of the entire city, with some similarities with suburban thoroughfares of today. Keep in mind that the complexity of trip planning isn’t really a problem because of the coordinating system, so the result could entail a dizzying level of complexity.

The counter-argument also needs to be made. If robo-Uber is so ubiquitous and so effective, then that may encourage more development of shorter, but vastly wider buildings — a pattern of relentless sprawl. Even so, autonomous cars may still be an enabling factor just by shuttling people back and fourth from different sides of industrial complexes/campuses within this sprawl. Elevators would also be quite costly to maintain in this vision, so I’d imagine long roads inside 2nd-4th floors for people-transport, and if you accept that, some ramps in some form are going to be necessary somewhere.

Context Shifts in Commutes

My original impetus for this concept had absolutely nothing to do with the growing logistical arguments laid out here. My original thinking was that we would implement this system because at some point in the future, we’re already using the robo-Uber cars for tremendously long commutes and single-floor renters would simply prefer to move the drop-off point to their front door as opposed to the building’s front door.

Pushback Against Homogenization

Comments on my previous post (several, in fact) brought up an extremely valid point that this future is depressing, undesirable, or outright objectionable. My first reaction is “I’m not advocating, I’m predicting”. It’s true that most futurology is combined with advocacy, while some of it is combined with warnings. For this particular idea, the thing that intrigues me the most is that I don’t really grasp what the human-level impacts are. I don’t know exactly how worried or enthusiastic I should be.

Elective and Non-Elective Modes of Transit

I feel relatively well-justified in my logic for why people would move toward door-to-door (or even living-room to living-room) transportation, once self-driving cars are ubiquitous. No one wants a longer commute, and given the choice, we will shorten our commute at every opportunity.

I want to make the point that the gross time involved in commuting might actually be less important to people than context switching. Self-driving future taxis should offer the lowest degree of context switches. Taxis today would off this, but the human interaction to deal with arranging the trip could, itself, be considered a context switch. With a flat, constant, reliable context that doesn’t require active attention, people might opt to spend half of their workday in commute. I don’t have much doubts that self-driving cars will lead to exactly this for some fraction of the population.

Others have encouraged diversifying the modes of transit we use. However, I think there are 2 dimensions to this — diversity in the modes we have to use within a single trip, versus the modes available to us for a trip. History is full of activities that we used to think of as a chore, but then became a past-time once it technology made it non-necessary. The walking legs of our commute might see the same transformation into dedicated, intentional, walking time as a replacement for walking from the parking lot, or the bus stop.