Self Driving Car - Tesla Autonomy Investor Day
Depending on who you talk to, you’ll get a lot of conflicting information on autonomous driving. On one side you’ve got people saying, “autonomous driving is closer than you think.” On the other side you have people saying, “autonomous driving is years away at the earliest.”
Tesla just held an autonomy investor day to show off their latest tech and software, as well as to show how the company plans to take advantage of it. So why did Tesla hold this event now and how far off are we?
Before we dive in, take a moment and hit the subscribe button, so you don’t miss out on future videos like this one. I’m Matt Ferrell ... welcome to Undecided. Now, let’s see if we can make a decision on this.
So let’s kick this off by jumping into what autonomous driving even is. There are actually some guidelines for the different levels of automation from the NHTSA:1
- Level 0 - No Automation.
- This is the majority of cars on the road.
- Level 1 - Driver Assistance
- This vehicle can assist with steering or braking, but not at the same time. Think adaptive cruise control.
- Level 2 - Partial automation
- This vehicle can assist with steering and braking at the same time, but still requires the driver’s full attention. This is where most modern cars that have some kind of “automation” fall today. It’s basically lane assist and adaptive cruise control.
- Level 3 - Conditional automation
- This is where a driver is still required, but they don’t have to keep their eyes on the road. The car handles almost everything.
- Level 4 - High automation
- This is where companies like Waymo currently operate. A driver is only required in certain circumstances, so if the conditions are right, then the car can completely drive itself.
- Level 5 - Full automation
- Exactly what you’d expect. This is when no human driver is required at any point.
As of today, cars like Teslas are all falling into the level 2 bucket of automation. The car up until this point has basically been a combination of lane assist and adaptive cruise control, but Tesla has really upped the ante recently with the seamless Navigate on Autopilot firmware update. The car can now pass slow cars on it’s own, take you through highway interchanges and ramps, and all without user interaction. From today’s event they stated that over 100,000 automatic lane changes are happening everyday. All without an accident. The driver still needs to be alert and ready to take over, but the car is doing all of the driving on the highway.
There’s also the underlying technology to consider with self-driving cars. Right now there are two basic approaches: LIDAR2 vs. vision. Most companies, like Waymo, have taken the path of LIDAR, which is shooting a spinning laser into the environment and getting back very accurate data for objects and distances. One downside is that this technology is very expensive, breaks down easily, and struggles with certain conditions like fog and rain.
Vision, which is what Tesla is working on, uses affordable cameras paired with ultrasonic sensors and radar to see the environment. One of the benefits of the vision approach is that the car sees much of the environment like we do, and if trained properly, could learn to navigate our existing roadways. See and understand stop lights and street signs. The downside is that this method requires immense amounts of data for deep learning to train the AI to drive the roads.
Which method is the better path? That’s still up in the air as nobody has fully cracked a self-driving car in all conditions yet. Elon stated very strongly that LIDAR is a crutch and he expects all other companies to back away from that path. I’d strongly recommend checking out Lex Fridman, who is a research scientist at MIT, and his YouTube channel for some great lectures on the subject. I’ll include a link in the description.3
Why did Tesla hold this event?
So why did Tesla hold this Autonomous Investor Day event? It comes down to the fact that Tesla considers themselves a leader in self-driving technology, but they’re often overlooked by the marketplace. Companies like Waymo and Uber tend to capture mindshare around self-driving capabilities and research. Many also consider LIDAR the better technology for self-driving cars as well.
Navigant Research created an autonomous driving leaderboard based on a company's autonomous vision and execution, which puts Tesla near the back of the pack.4 It’s a really strange, and in my opinion deeply flawed, leaderboard they’ve created because most companies are still working behind closed doors. There’s no real product on the market, and for those that do it’s very limited in scope at this point. How do you compare companies autonomous driving tech, strategy, sales, and product when most of this is still a work in progress? It’s really strange.
Add to that the recent IPO of Lyft and Uber’s upcoming IPO, which some have valued at between $76 and $120 billion.5 That would give Uber a much larger market cap than Tesla, which is currently under $50 billion, even though Uber doesn’t manufacture their own cars or have a clear path to profitability.6 Ride sharing and autonomy is seen as one of the next big things.
Tesla is trying to combat the perception that they aren’t a leader in autonomous driving. They don’t see themselves as just a car company, they also see themselves as software company. They’re continually hammered by critics around delivery numbers and company profitability, rightfully so, but that's only one part of the story. Not only is Tesla successfully selling one of the most popular luxury sedans right now, but they’re also one of the companies that’s closest to having a full self-driving car on the market.
Unlike their autonomous car rivals, Tesla has over 400,000 cars on the road (nearing 500,000) with the full sensor suite, which includes 8 cameras, 12 ultra sonic sensors, GPS, and radar. Their competitors are in the thousands. And when it comes to winning the machine learning and autonomous race, data is the key. The amount of data Tesla is collecting is immense and dwarfs their competition, we’re talking about over 1 billion miles logged with Autopilot, which is something that often gets glossed over when it comes to Tesla in comparison to those other companies. This quest for data and doubling down on autonomy is also the driving force in Tesla’s decision to include autopilot on all cars they manufacture now ... it’s no longer an add-on option.
And as soon as Tesla has full self-driving available, they’ll be able to activate their own ride sharing service, the Tesla Network. Hundreds of thousands of cars on the road, that can start earning their owners money by shuttling passengers around ... on their own. It’s not to say that every Tesla owner will do that (I most likely won’t), but the fact that Tesla could become a major player in the ride sharing world overnight is something Lyft and Uber should be worried about.
Tesla recently announced a lease option for the Model 3, but with a strange twist. At the end of your two year lease, you won't have the option to buy the car. Instead Tesla wants to hold onto these cars and use them as part of the Tesla Network, which means Tesla itself is going to own and operate a fleet of ride sharing vehicles.
This changes the whole dynamic of cost of ownership when it comes to a Tesla. As Elon put it in a recent interview on Lex Fridman’s Artifical Intelligence podcast:
“Buying a car today is an investment into the future. I think the most profound thing is that if you buy a Tesla today, I believe you are buying an appreciating asset — not a depreciating asset.” - Elon Musk7
Someone who spent $45,000 dollars on their new Tesla Model 3 can start having the car earn it’s own keep by acting like a taxi in its downtime. It’s conceivable, much like Elon is suggesting, that your car could actually turn into a profit maker for you as an owner. That is just ... bananas.
Is self-driving really that close?
Is self-driving really that close? Well, if Tesla keeps to their word then, yes ... sort of. I do believe Elon and Tesla that they’ll have the new full self-driving computer and software being pushed out to their fleet by the end of this year. But there’s a big difference between having a car capable of full self driving and actually being able to do it. There are major regulation hurdles that Tesla will have to jump through and prove that their system is actually safer than a human driver before anyone can take a nap during their commute to work.
I conducted my own survey last week because I was curious how others felt about self-driving cars and how quickly they’d be willing to jump on board. I asked three questions:
- How likely are you to use a self-driving car as soon as it’s available?
- Would you buy a self-driving car?
- What safety standards should a self-driven car be held to vs. a human-driven car?
I had a hunch where the results would land, but I was surprised by some of the responses. On the first question I was expecting the majority of people to fall in the middle somewhere. In fact, I thought most people would be in the, “yes, but in a very limited way” group. Instead, half of respondents actually said, “very likely! I’ll use it all the time.” 42% were split between “in a limited way” and “I’ll wait until the bugs get worked out. But that still means 92% are open to self-driven cars at some point and in some form.
On the second question, 83% said they’d buy a self-driving car. Again, that really surprised me. I expected it to the be the majority, but not by that margin.
On the last question, this fell in line with what I was basically expecting. 55% said a self-driven car needs to be dramatically safer than a human-driven car. 40% said it only needs to be as good or better than a human driver. Removing all emotion from the argument, it makes logical sense that a self-driving car only needs to be as good or better than a human driver. But we’re not a group of Vulcans here.
Emotion plays a big role in acceptance. So that’s where my optimism gets a healthy dose of cold reality. It’s going to take a lot to gain the confidence and approvals from regulators and the public in order for drivers to be able to take a backseat ... literally. Elon made that exact point in the same Lex Fridman interview:
”How much safer than a person does Autopilot need to be, for it to be okay to not monitor the car? And this is a debate that one can have, and then, but you need a large amount of data, so that you can prove, with high confidence, statistically speaking, that the car is dramatically safer than a person. And that adding in the person monitoring does not materially affect the safety. So it might need to be 200 or 300% safer than a person.” - Elon Musk
He says that would be measured by incidents per mile, which includes all forms of crashes and injuries. Even after they release the full self-driving capabilities, autopilot will still require a driver to be attentive, hand on the wheel, like they do today. Elon’s suggestion that it might take 6 months of that testing before they could go to a driver kicking back and tuning out seems a little optimistic to me. It’s Elon time, so it might be better to double or triple that estimate at the earliest, but it’s still impressive that Tesla is that close to having full self-driving capable features out in the wild for people to start using. Again ... bananas.
The final piece that’s probably the biggest factor is all of us. How will people react when they see a car driving itself through a parking lot? How will people react if they’re driving behind a autonomous car that they think is to slow or cautious? People that are angry and fear losing their job driving a car for a living? We already see people lashing out against EVs with coal rolling or ICING charging spots. We’re probably going to see more aggressive behavior or vandalism of self-driving cars ... until it starts to become the norm.8
I do think autonomous cars are closer than many people think from a technology standpoint, and that Tesla will launch full self-driving capabilities this year or early next year ... but that we won’t see regulatory approvals for true self-driving for quite some time. It’s going to vary wildly state by state, country by country, and I’d guess that we’re looking at quite a few years before it’s rolled out to most places where current Tesla owners reside.
What do you think? Do you think that Tesla is going to hit those milestones and have full self driving available sometime in 2020? Or do you think that autonomous driving is just much further off? Add a comment below. I’d love to hear what you think.
And if you liked this video, be sure to give it a thumbs up and share with your friends because it really helps the channel.
If you’d like to support the channel, there are some ways you can do so.
Want some cool t-shirts, like the one I’m wearing? Check out my SFSF Shop for some cool Tesla, Space X, science, and Undecided shirts. Every purchase helps to support the channel.
Get some great Tesla accessories, like console wraps, LED lights, and more from Abstract Ocean and save 15% off your first order with my referral code.
And a big thank you to all of my Patreon supporters. You’re helping to make these videos possible. If you’re interested in early access to videos, behind the scenes posts and polls, check out my Patreon page for additional details and I hope to see some of you over there.
And if you haven’t already, consider subscribing and hitting the notification bell to get alerts when I post a new video. And as always, thanks so much for watching, and I’ll see you in the next one.