Why don't we have self-driving cars yet?

Driving during rush hour isn't one of my favorite pastimes, and let's face it, we've all had the same thought: it'd be wonderful if the car could just drive itself. We would put in the address, sit back, and let the vehicle do all the work. A couple of years back, investors were piling cash on these self-driving companies. However, now, many autonomous vehicle startups are disappearing, and cars still aren’t driving themselves. In reality, self-driving cars are getting farther away from us, and we'll have to be patient.  

The idea of a self-driving car has been around for ages. Still, in the last twenty years, we have started to consider it a viable alternative to driving. (There's something to be said about replacing cars with more efficient public transportation, but that's another topic). Throughout the 90s, the US Government invested an equal of $1.5 billion in today's money into optimizing and researching the most acceptable way to automate vehicles. Other efforts followed, but none actually panned out. They were hampered by lagging technology, extremely high costs, and a general lack of interest. That's not to say some big names haven't tried. 

Back in 2009, Google wasn't the giant that it is today. Still, even then, the company had big ambitions regarding self-driving cars and started its secret autonomous vehicle division. The company leveraged its ability to scan maps, using the 100 Toyota Priuses it had bought to optimize Street View, to further understand how a self-driving car should work. Eventually, through the years and many iterations, Waymo was born, and it's through this name that we've come to recognize it as one of the biggest autonomous vehicle companies around. Moreover, lately, there's news that more Waymo vehicles will act as taxis in San Francisco, but that's not to say it's been an easy ride.  

Waymo has been at the forefront of autonomous technology, clocking millions of miles on public roads and refining its self-driving systems. If you live in major cities, chances are you have seen one of their vehicles (with 25,000 of them it should be easy). Still, for years, the company struggled to gain ground on certain aspects, such as safety. For example, in June 2015, 12 Waymo robot vehicles were involved in accidents; the next month, the number increased to 23. 

What's interesting about this aspect is that Google has gone through all the crashes and justified some while defending itself in others, and that's what I find utterly ironic. Humans have to explain what their robots do, and this goes to show one of the biggest flaws we're currently having with artificial intelligence. Nevertheless, Waymo's progress serves as a beacon of hope, inspiring others to pursue self-driving cars. It's one example of a successful endeavor, and the same cannot be said for hundreds of other startups in this field.

Let's place a philosophical dilemma on the table. You're boarding a plane, and right before you sit down, you realize there's no pilot. Instead, there's one computer and nothing else. Would you fly? Right now, an aircraft can, theoretically, fly itself, but that doesn't mean that passengers want to because we still feel that technology hasn't advanced enough to replace a pilot. 

So, the problem stems from AI itself. Yes, we've made massive advances in artificial intelligence, but driving, ironically, doesn't only require a series of algorithms but the ability to change scenarios as soon as possible. So, as journalist Christopher Mims put it, autonomous vehicles have taken a step back because they require an ideal set of conditions, which rarely happens. 

When scenarios change, autonomous cars could end up in a crash, as has happened before, and therein lies another problem. Since it’s new technology, the world will be looking for any accident that happens and judging it harshly. After all, there have been around 30 fatalities involving these vehicles, which sounds tragic, but let's put this in perspective. Every year, there are 40,000 highway deaths in the US alone. So, if we want AI to take over driving, we need to make it safer, and we don't have the conditions yet. 

Many hyped autonomous vehicles, including Elon Musk. So, everyone from small startups to big-investors were in on it the self-driving car trend, but there's also the other side of the story. Let's not forget that even Waymo's CEO at the time, John Krafcik, said in 2018 that a self-driving car was a long way from happening. 

That's the realization that many investors have come to lately, even those who seemed on the right track. Let's take Cruise, for example. Seven years ago, this startup was one of the most promising ideas for the near future, and it had the right backers; GM and Honda. Both companies poured billions into the company and hoped to release a self-driving car by 2021, which still hasn't happened. 

Argo AI was another startup that had promised the world, and itself, that it would have driverless cars by 2019, then by 2020, and then by 2021. Volkswagen, Ford, and other big investors had poured $3.6 billion into it, and to no avail, as the company shut its doors in 2022. 

Other startups, such as Nuro, still use safety drivers in every vehicle, just as pilots on a plane, to avoid accidents. It's obvious, but worth highlighting, that even in airplanes, accidents still happen, which shows how fragile AI can be, especially when dealing with people's safety. That's why even a company as big as Waymo is still taking baby steps towards self-driving cars, and that's where this article might come off as Debbie Downers, but what if autonomous vehicles are decades away? We will have robot taxis soon in restrained areas that might not continuously operate efficiently, and that's something we'll have to contend with for years. 

Self-driving cars are currently in bumper-to-bumper traffic; slowed down by their technology, plus a massive series of legal and technological challenges. Still, it's an idea that's moving forward. My bet for the closest thing we'll have in the short run is a series of very advanced driver-aid technologies, which will undoubtedly make the experience much safer for everyone, but we're in the cradle stage; we're in what could be defined as the 90s of the internet but for cars. Eventually, someone will make it big, and to get to that point, many will have to pull to the side of the road and leave their cars to die. 

‍

Want more content like this? Subscribe to our free weekly newsletter! It will quickly become your one-stop shop for startup news.

Subscribe here

‍