Dev Notes 20: Draft Concepts on AV Policy

Dev Notes (DN) document progress towards some larger article. All discussion is preliminary

These notes start putting together thoughts for a future article on autonomous vehicle policy. It is in some regards a filler article on Thanksgiving.

I admittedly have a lot of fires burning right now. There's the Wikidata project, which is a bit on hold while I work on quicker completable projects this month. There's part 2 and 3 of the IMDb movie analysis series. I'm allowing time in between those, so can revisit that again next week. There's another project which should finish up very soon. I probably shouldn't open new thread, but I want to write at least one rhetorical article before November ends (so far my blogging this month has all been exploratory/researchy and none have a rhetorical argument. That seems unfortunate if that was all I did).

There are a few points I aim to discuss in this upcoming article.

General Thesis

The general argument is that autonomous vehicle (AV) transit is closer than many people realize. This can provide everything we want out of transit: safe, convenient, and cheap in a way that no other transit system can. It's also a blank slate for a new set of public policies. Autonomous vehicles can't achieve their full potential if they just sit in traffic. While technical solutions can help here (in theory AVs can have closer following distance at faster speeds) this is challenging. Thus in order to reach full potential of this new source of transit we need to figure out a way to shift away from single occupancy vehicles. I plan to discuss several options here, but one of the simplest might be the best. That is policy can essentially pay people (through small per-mile subsidy) when they take a non-single occupancy vehicle. There is some precedent for carpool subsidies nationally that provide a template, and one of the most successful transit systems within California is CalVans, pointing to the potential of such a carpooling system.

Waymo Shows AVs are Real

This month the autonomous vehicle (AV) company Waymo began rolling out highway access in the San Francisco Bay Area. Previously it has been successfully giving rides within San Francisco, hundreds of thousands of rides per month. Highway access allows them to connect all the way from San Francisco to San Jose, opening up a whole new set of options for transportation.

Traffic as Externality

I hope to discuss how taking a single-occupancy vehicle causes externalities. The traffic created is a public cost. There is benefit in making this part of the cost.

Transit Spending is Already Long Term Minded

I hope to discuss some ongoing projects nationally, and specifically in the Bay Area. Many transit projects are built over a decade, and the value equation factors in decades of time. Even if you still believe AVs are future tech, it is well within the range of the planning window of typical transit projects.

Some example projects include a new transbay crossing tunnel, which is intended to be standard gauge (in the bay area there is the BART subway system, which does use the standard gauge, ie track width, as is in typical rail around the country). Additionally, there is a multibillion-dollar, years-long plan to dig a connection between a station in San Francisco to connect with the more centrally located Salesforce Tower Transit center.

I am going to have to thread a thin needle here to have a reasonable approach. But working hypothesis is these kinds of local rail investments simply don't make sense with coming AVs. While I love taking trains, and regularly use BART/CalTrain/Muni, almost all systems' most destinations still take multiple inconvenient transfers. We can achieve better cost and better convenience with AVs. I worry these decades long local rail projects are going to be obsoleted by the time they are built.

As painful as it is for the urbanism dream, we should invest in buses and road improvements now, as these will have direct transfer as AVs roll out.

BART is Insanely Expensive to Operate

I hope to pull some exact numbers later. But currently there is a per-trip deficit of over $10/passenger trip. This goes to show that we already subsidize transit a lot. We can possibly be more effective.

Conclusions

Again this was only a partial filler on the holiday with the argument and evidence not really fleshed out. I hope to make some progress on this topic over the weekend.