If you thought self-driving vehicles were still on the distant horizon, think again.

Last week, Local Motors, in partnership with IBM, announced Olli, a self-driving transportation system that will soon be operational in Washington, D.C. The announcement comes on the heels of a similar initiative revealed by Lyft and General Motors, which recently declared a joint venture to bring new autonomous taxi services to an as-of-yet unnamed city sometime in 2017. A mere three years ago such technology thought to be decades away. Some even predicted that we wouldn’t see autonomous vehicles in our lifetime.

The pessimists got it wrong.

Seemingly every week, attention-grabbing headlines announce new developments in vehicle automation. From Tesla and Volkswagen to Google and Uber, companies in Silicon Valley and Detroit are investing more heavily in this technology Everyone wants to be the first to market with their vision of what the autonomous future will look like. And more and more it’s looking like we’ll be living in an Uber-style, shared autonomous transportation system.

Undoubtedly, shared autonomous vehicles (SAV) like Olli will be the initial entrants in the transportation market. That should come as no surprise. Many of the near-term benefits will initially accrue to urban residents reliant on taxi services. Whereas human-operated taxi rides in New York City average about $7.80 per trip, an SAV-style system — think Uber but without the driver — could reduce the total per-trip cost to around $1. Those are massive savings that could have a profound impact on how we think about and use urban transportation.

First, however, regulators need to get the rules of the road right.

There’s hope that the balkanized regulatory regime surrounding driverless cars could start adapting as early as this summer. Last week, the National Highway Traffic Safety Administration (NHTSA) announced that it would be releasing “deployment guidance and state model policy on autonomous driving technology” this July. In an all-too-rare nod to intellectual humility, NHTSA has recognized that “its existing legal authority is likely insufficient to support mass deployment of autonomous vehicles.” Initiatives like these will be increasingly important if the United States is to catch up to countries like New Zealand and the United Kingdom, where regulatory impediments to testing autonomous vehicles are minimal.

Over the past two years, there’s been significant progress in addressing many of the concerns related to autonomous vehicles. But barriers still remain. Cybersecurity, privacy, and liability are leading concerns among regulators, to say nothing of the public’s persistent worries about the safety of these vehicles. Surmounting such obstacles will be no small task, but, eventually, the value of driverless vehicles will become evident to all.

As Mercatus Senior Fellow Adam Thierer and I noted in a 2015 article, many of the potential problems with this technology will be settled in due course. Although NHTSA and government agencies have roles to play, much of the onus remains on private actors. Resolving these issues lies, ultimately, with the people “developing and testing the operational systems,” rather than “in endless bureaucratic proceedings and labyrinthine layers of regulatory red tape,” we wrote. “The tort system will simultaneously evolve to help remedy harms that develop. Lawmakers should not interfere with that evolutionary process.”

Nor should lawmakers interfere with data collection practices — such as those currently being conducted by Tesla — that allow auto manufacturers and software developers to gain greater insight into how best to optimize self-driving algorithms. Government can be a valuable partner in promoting best practices and standardizing regulations, but it can also be a roadblock to this all-too-important technological development. The costs associated with self-driving cars will continue to diminish as the technology matures, customers become more intimately acquainted with these vehicles, and big data helps better inform operational roll-out.

Just as both pecuniary and social costs will dwindle, the benefits of adoption will continue to swell. Annually, tens of thousands of lives will be saved, and injuries associated with automobile accidents will plummet. We will see hundreds of billions of dollars in fuel savings, billions of travel time hours reoriented to non-motor vehicle operating tasks, and massive reductions in the total number of vehicles on the roadways. The total comprehensive savings in congestion, fuel, and health care expenditure could easily amount to over half a trillion dollars annually, possibly more.

Driverless cars hold the potential to disrupt massively our current way of life — and for the better. The question is not whether we’ll see these changes in our lifetime but when. I’d bet on sooner, not later.

Op-ed by Ryan Hagemann; originally in RealClearPolicy