On the one hand, there are signs that innovation is accelerating. General Motors figures it will be first on the road next year with an in-cabin assistance system that “watches” where drivers are looking and nags them if they’re not alert enough. Meanwhile, Tesla is updating its Autopilot software with a revision that might have prevented that fatal accident in Florida—the technology will now use radar instead of cameras as the primary way of gathering data about a car’s surroundings.

But other companies are approaching apparent detours. Apple’s project may be taking an alternate route after dozens of layoffs, according to a report by The New York Times. The company is already testing fully autonomous vehicles, but the whole initiative has been so mysterious that it’s tough to tell if the effort is stalling or just switching lanes. Google’s project suffered a similar shakeup in August. It’s been experimenting since 2009 but still hasn’t managed to establish a demonstrable lead.

Just about the only thing anyone can say with certainty about self-driving cars is that it will completely upend the business model for auto insurance—a new projection from Aon suggests that premiums could be reduced by up to 40%. So, yes, insurers certainly seem to think robots will be more reliable drivers than humans.

Get Data Sheet, Fortune’s technology newsletter.

But it’s important to remember that the car “upgrade” cycle is far longer than for smartphones or computers. U.S. car sales reached an important mile marker in 2015—growing to around $570 billion, a 5.7% increase over the previous year. Considering many Americans keep their cars at least 10 years when they buy them outright, maybe Apple and Google aren’t as behind as you might think.