Nearly two years ago, while celebrating the seemingly imminent arrival of General Motors' self-driving cars in Lower Manhattan, Governor Andrew Cuomo declared New York "the hub of autonomous vehicle innovation in the nation." That title may have been a bit premature. Soon after, Mayor Bill de Blasio expressed "real safety concerns" about the technology, claiming the city's transportation department was never consulted. The GM pilot was quietly mothballed within a year, and the robo-taxis have steered clear of New York ever since.

Into this fraught and AV-barren environment cruises Optimus Ride, a Boston-based driverless tech company that has set its sights and sensors on city streets. On Wednesday morning, the company launched a free shuttle service in the Brooklyn Navy Yard, which will carry passengers between the office park's new ferry dock and the complex's entrance on Flushing Avenue.

Each of the half dozen vehicles seat four passengers, with the front two spaces reserved for non-autonomous operators (humans). A software engineer sitting shotgun collects data, while a “safety driver” keeps his hands inches from the steering wheel. The company says it has spent the last few weeks mapping the one mile route, which is entirely on private streets, and thus does not need DMV permission.

The next step, according to Optimus CEO Ryan Chin, is to make the case to state regulators and the DOT that the technology should be expanded onto city streets.

"The end point is that autonomous should be everywhere," Chin told Gothamist. "Our view is we have to start here and then move over." For now, the service bills itself as "New York state's first self-driving vehicle program."

So what should New Yorkers expect now that the long-promised robo-chauffeurs have finally arrived in Brooklyn? Judging from our test ride yesterday, driverless utopia feels more driver's ed in a stretched out Kia Soul.

In one dicey moment, an oncoming car straddling the yellow line came inches away from the vehicle, forcing our human minder to abruptly jerk the wheel, lurching the car to the right. The vehicle was already making this maneuver on its own, the human explained. Sitting shotgun, the company's director of system engineering and testing, John Sgueglia, described the interaction as a useful data point, adding, "Next time, I might give it a bit more space."

The company boasts that its proprietary software is Level 4 automation, interpreted by some automakers to mean the driver's seat occupant can literally fall asleep at the wheel, which does not seem wise.

Other aspects of the ride did little to inspire confidence. Both parking and three-point turns—required at the end of every trip—seem to be an intricate two-person operation, somehow involving more effort than the human-powered alternative. The turn signals are manual as well. The cars do not have air conditioning, though this is apparently on the "growth path." Despite calling itself an "accessible form of mobility," the vehicles are not yet wheelchair accessible.

According to Bryant Walker Smith, a law professor at the University of South Carolina whose research focuses on autonomous vehicles, a wide gulf between self-driving service promised and service offered is commonplace in the autonomous vehicle industry.

"So often the headline is 'Driverless Cars Now in X,' but the reality tends to much more limited," he said. "If someone is in the vehicle supervising, it's not a driverless vehicle—it's an expensive bus, or a really expensive golf cart."

The service is funded by the Brooklyn Navy Yard Corporation, the non-profit developer and manager of the city-owned Navy Yard, which in recent years has become a sort of innovation hub of its own. Both BNYC and Optimus Ride declined to share the project's cost.

Not that a bumpy ride on an early version of a new transportation mode is so unexpected. But Smith notes that being upfront about the limitations of the technology is crucial for public buy-in, in order to avoid "unrealistic expectations about the present, or even the immediate future."

The immediate future for self-driving tech, says Smith, will probably mean "a lot more announcements like this," combined with niche applications that bear little resemblance to the widespread deployment that Silicon Valley had previously anticipated.

Indeed, industry observers note that driverless boosters have experienced a series of reality checks in the last eighteen months, blowing past major deadlines (see: Elon Musk's promised cross-country Tesla trip) and gaining little traction with a public that remains overwhelmingly terrified of driverless technology.

Meanwhile, some of the most AV-friendly states have also begun to take a less welcoming approach, spurred in part by a woman's death in Arizona by an SUV outfitted with Uber's self-driving technology. Investigators found that the system had sensed the pedestrian but failed to act, while the driver was reportedly watching television on his phone. Once a deregulated haven for autonomous vehicles, Arizona Governor Doug Ducey banned them.

Governor Cuomo did not herald Optimus Ride's arrival with a press release, and the Governor's Office did not respond to a request for comment.

A mayoral spokesperson said the city was fine with the project as long as it remained in the Navy Yard, but reiterated de Blasio's "strong opposition" to testing the technology on city streets. In testimony before the Senate last summer, DOT Transportation Commissioner Polly Trottenberg outlined those concerns—namely, the weakness of federal safety standards, and the fact that worsening congestion could be exacerbated by "single-occupancy, or even zero-occupancy vehicles flooding our city streets." There are few signs that those fears have been assuaged.

For many of the industry's top leaders, the realization that they may have overestimated the ease of seeding a complex transportation environment with robots is beginning to crystallize. Deploying not-quite-driverless vehicles in heavily constrained locations might be seen as a sensible reaction to that reality. But for those still convinced that the autonomous revolution is just around the bend, recent setbacks seem to have pushed their thinking to the absurd.

Take for example this story published in the Times last week, which examines the "societal" problem of jaywalking through the lens of autonomous developers. According to an unnamed industry official, the solution is installing "gates at each corner, which would periodically open to allow pedestrians to cross."

That may sound drastic, but experts say that such incursions into shared street space are well within the realm of possible outcomes once driverless technology goes mainstream.

"That's probably not going to take the form of cages," Smith, the law professor, told Gothamist. "It's much more likely going to be a large information technology company deploying these systems, recording anyone who steps in front of the vehicle, identifying them with facial recognition technology, and seeking public enforcement or blocking that person from using their AV service as a penalty. I think that's even more dystopian than cages."

While noting that there are environmental and safety benefits to a well-designed autonomous program, Smith also says there is legitimate concern that "enthusiasm for automated driving is going to turn us back to the 1950s, where pedestrians are again shoved out of the way for cars."

Asked on Tuesday if cracking down on jay-walking and other unruly pedestrians would be a necessary element of driverless adoption in New York City, Optimus Ride Co-founder Ramiro Almeida didn't deny it was a possibility.

"I think overall we'll need to interact with different elements of the environment in a much more responsible way," he said. "Jaywalking might be a part of that. If we respect the rules of the environment, if we are comfortable with making sure that we go by those specific guidelines or rules, I think it will be beneficial to everyone."