State Of Autonomy: July Recap

4 min readAug 2, 2018

Daimler, Bosch and NVIDIA are in talks to deploy a robotaxi service “somewhere in Northern California.”

Every month, I recap the news articles I’ve consumed around autonomous vehicles, calling out the highlights and keeping track of market projections. This is also your chance, dear readers, to nominate a topic for discussion in the following month.

July exposed the speed at which automakers are streamlining and disrupting their operations, as a continuation of the ground-breaking move Delphi made last year when it cleaved its business into two distinct entities. It’s notable, perhaps mostly because it’s the way things should go rather than the way one would traditionally expect them to go. Behemoth blue chip companies in transportation are increasingly recognizing that their large, lumbering shadows don’t scare tech companies — and that trying to absorb revolutionary tech into automakers’ portfolios without any real strategy simply results in some David Cronenberg-esque blob of waste.

Ford announced a clean break between its traditional operations and autonomous vehicles, including a $3B investment (plus the previous $1B injected into startup Argo AI.) This fancy, futuristic division will be called “Ford Autonomous Vehicles.” Ok.

Daimler followed suit by proposing to shareholders a divvying up of their operations into three “largely independent units”: cars, commercial trucks, and whatever crazy stuff falls under the umbrella of mobility.

In China, a paper partnership has emerged between FAW Group, Dongfeng, and Chongqing Changan — three of the nation’s largest automakers (and domestic stakeholders in foreign automotive partnerships.) These three claim they’re going to be developing a state-owned ridesharing service, likely utilizing autonomous vehicles as soon as feasible, and rivaling Didi Chuxing’s behemoth presence as the incumbent local rideshare platform.

This Month’s Highlights:

Commercializations:

Drive.ai is the brainchild of AI demigod Andrew Ng, who previously ran deep learning at Google and Baidu. After several months of testing on Texas roads, the company has now deployed four Nissan NV200s in a very limited public trial, geofencing trips inside a two square-mile area of Frisco, TX:

Two callouts worth making here:

These are the first public-facing AVs to communicate with pedestrians through displays. The vans have four small screens that display what Drive.ai belives to be the most vital statements: Waiting (for you to cross), Going (i.e. driving has begun), Entering/Exiting (passengers), and Human Driver (i.e. please don’t publish some bullshit article about how our self-driving vehicle crashed when it wasn’t even in self-driving mode.)
Prior to this program, Drive.ai ran a few tests in Frisco without a safety driver present — rare air thus far, with Waymo being the only other company claiming such a feat. Drive.ai’s team speculates that they’ll start removing drivers from this public trial by end of year; it’s worth noting that they’ve built in a remote safety operator redundancy to solve for whatever sub-catastrophic problems might arise.