Editor’s Note: Welcome back to Jeff Brown’s prediction series. Over the next several days, Jeff will share his biggest predictions for the new year.
Today, we talk self-driving cars and autonomous technology. What is the state of self-driving cars? And where will they go next year? Read on…
Van Bryan (VB): Jeff, let’s talk about self-driving technology. And to start, we have a prediction of yours from last year. You predicted that Waymo – Google’s self-driving unit – would launch in more cities in 2022. Did that happen?
Jeff Brown (JB): It did. When we spoke last year, Waymo’s self-driving taxis were operational in Phoenix, Arizona. I actually had the chance to test out the technology for myself.
Jeff Rides in Waymo’s Self-Driving Taxi
It was fantastic. What Waymo has done is “geo-fenced” parts of the city. So, it’s extensively mapped a large swath of the Phoenix metropolitan area. That’s how the car was able to drive as well as it did with nobody in the driver seat.
But unless Waymo launched in another major city, its tech would appear to be nothing more than a limited pilot. And that’s not material enough for Google/Waymo’s end goal.
So, I knew it would be only a matter of time before it launched in other cities. And in November, Waymo went live in San Francisco, California. And in October, Waymo announced it would be launching in Los Angeles in the coming months.
VB: And will we see more cities go live with these self-driving taxis next year?
JB: Absolutely, and not just with Waymo. I’ll even argue that the rollout of self-driving technology will be nationwide, thanks to Tesla, but I’m sure we’ll get to that topic in a moment.
These early cities by players like Waymo, Cruise, and Motional are really just a precursor to rolling out the service in major metropolitan areas across the country.
It won’t happen all at once (except for Tesla’s release to its customers). Each ride-hailing company will have to launch on a city-by-city basis. They have to deal with the different regulations for each place they roll out the service.
So yes, I predict that sometime next year, many of our subscribers in these areas will have access to this kind of ride-hailing service.
But we should remember that this isn’t Google’s real ambition for launching a self-driving ride-hailing service.
VB: How do you mean?
JB: We should always remember that Google – first and foremost – is a data collection and advertising company. About 92.5% of its revenues come from advertising. I’d argue it’s more of a data surveillance and collection company. And Google’s lifeblood is our data, which it monetizes through targeted ad sales.
Many don’t see it, but I view Google’s Waymo initiative as another way for the company to collect data on all of us throughout the day or night. And it makes perfect sense.
If we were riding in a car with Waymo’s self-driving software, Google would be able to know our favorite places to visit and how often we visit. Just as one example, is a person visiting an upscale boutique once a week? That’s valuable data for a company like Google.
It will also have the ability to serve up ads to a screen in the Waymo car that are targeted based on what it already knows about us. After all, sitting in the back of a robo-taxi, we are a captive audience.
I even envision that Waymo will license its technology to major automotive manufacturers, not necessarily for ride-hailing services. This model will look more like what Tesla is doing for its own cars.
It will be an autonomous driving package that the automotive manufacturer can sell to its customers as an upgrade. Waymo will benefit from these deals financially, and also through the ability to collect additional data from any car where its technology is employed.
So, the long game here isn’t necessarily to launch a nation-wide, self-driving taxi service in its own name. I continue to believe that Waymo will ultimately license its self-driving software to other automotive makers and ride-hailing services.
After all, why would Google build and maintain its own cars when it can license the software and let somebody else build the vehicles?
This is exactly what Google did with its Android operating system. Its Google Pixel phones are essentially a showcase of its software technology. We can think of it like a reference architecture.
Google’s goal all along with regard to the smartphone industry was to license the software and make it easy for others to build and sell Android phones. And the Android OS software is now the dominant software platform for smartphones worldwide.
This has always been Google’s master plan. And I did predicted we would see an announcement along these lines this year. And it just happened.
Last month, Waymo announced it had licensed its self-driving software to China-based Zeekr. Most subscribers likely don’t recognize that name.
Zeekr is a subsidiary of Geely, one of the fastest-growing car manufacturers in the world. It is actually the seventh largest automaker in China, and the owner of Swedish auto brand Volvo. It’s a major player.
The plan is for Zeekr to launch an autonomous ride-hailing service in China, powered by Waymo’s software. This is the first of what I predict will be many more deals to follow. I predict that we’ll see at least two more major deals next year from Waymo in the industry.
VB: I know you’ll keep tabs on Waymo’s progress next year. But I’d also like to ask you about autonomous technology for logistics. Last year, you predicted we’d see more adoption of this technology for delivery purposes.
JB: And we have seen that. One of the more surprising names adopting self-driving technology for delivery purposes is Walmart. Last December, the company announced a partnership with an early stage company called Gatik.
The idea is that the autonomous trucks can commute between Walmart’s stores and the ghost logistics centers. This isn’t happening at every store yet, but it’s been expanding coverage. These autonomous trucks are now in use daily.
The Gatik trucks were already operational in parts of Arkansas. And in May, Gatik rolled out operations in Kansas. I suspect this will be a city-by-city rollout for Walmart and Gatik. Like Waymo, it’ll have to work through the regulations in each state.
And we have to mention Tesla.
Tesla finally launched its own self-driving semi-trucks just in the last few weeks. The company announced its plans for the truck all the way back in 2017. It feels like decades ago.
But this year, Tesla confirmed it had started production on the Tesla Semi. And it already has its first customer.
Pepsi ordered 100 of the semi-trucks for routes between Modesto and Sacramento, California. And the plan is to deliver the first truck by the end of the year. This is happening right now.
And most importantly, Tesla Semis have all the same autonomous driving technology and sensors as Tesla’s passenger vehicles.
VB: Speaking of Tesla, you also predicted that Tesla’s self-driving software would continue to improve when we spoke last year.
JB: Actually, I predicted that Tesla would release its full self-driving (FSD) commercial release within 2023. And I’m happy to say that just happened this month with Version 11.
It was a fun year watching Tesla scrap and rebuild its FSD artificial intelligence (AI) using neural networks and then to see the technology fly in the weeks that followed. It was really extraordinary.
Every two or three weeks, Tesla would release an updated beta version for testing and we could see the material improvements in performance.
It was really the last two months where the technology was clearly maturing and getting very impressive. By November, Tesla users were demonstrating point-to-point drives that lasted for hours without a single intervention. The AI was in control the entire time.
With perhaps the exception of Musk and his team, I don’t know anyone who believed Tesla would deliver FSD this year.
VB: I remember seeing a video earlier this year of a Tesla driving from downtown San Francisco to the airport without any human assistance.
JB: Right. That’s a great proof point.
Driving down a highway is one thing. But operating through a dense urban area like San Francisco and managing the on and off ramps and all the turns is what makes self-driving technology different than something like an advanced driver assistance technology.
And what’s really impressive is that these cars don’t need extensive mapping and geo-fencing like the Waymo, Cruise, or Motional cars. What Tesla is doing is building a general-purpose artificial intelligence that can “think” and “decide” the best way to operate the vehicle.
It accomplished this so quickly by employing advanced neural network technology, a form of AI, and building one of the worlds most powerful supercomputers – the Dojo – to train the AI quickly.
And one final point about Tesla’s competitive advantage.
Because it has so many cars in the field collecting data on self-driving mode (AutoPilot), Tesla has the world’s largest real-world data set for self-driving. Combine that with the Dojo and Tesla’s AI and we get Version 11 of FSD.
It has been a remarkable year.
VB: Before we wrap up, I have to ask you about one of Tesla’s more surprising announcements this year. What do you think of Tesla’s autonomous robot, Optimus?
JB: Optimus is Tesla’s bipedal – or humanoid – robot. The plan is for Optimus to be a general-purpose robot. So, it can walk, carry objects, and generally do work that humans would typically do.
Optimus comes at an interesting time because of what has happened in the labor markets.
During the pandemic, we saw this massive cultural shift in the workplace which has been persistent. Sadly, this shift happened for the wrong reasons. It was politically driven.
The terribly misguided and ineffective pandemic policies forced many of us to stay home. But in many cases, we were paid to stay home. That has negatively impacted the labor market.
Most don’t know this, but even though the pandemic has long been over, Biden has extended the pandemic emergency state formally until February of 2023.
That means that we continue to see distortions in the labor market, resulting in major labor shortages in many areas.
When people are paid to not work, of course they won’t work. And for those who are working in a discipline with a labor shortage, we’re seeing a lot of short-term employment.
People are joining for six months and quitting, taking time off, and then repeating the process. I know some have referred to this as the “Great Resignation.”
Lower-paying jobs have been most negatively impacted by this trend whether it has been in healthcare, food services, logistics and distribution, or similar areas.
But necessity is the mother of invention, right? And robotics companies really saw this as a chance to accelerate their development. Tesla was no different.
VB: A lot of people were skeptical that Tesla could pull it off.
JB: Actually, they still are. I think I’m probably one of the few that is very bullish on what Tesla is doing with Optimus.
But I assure you that I’m not enamored by Tesla and Musk. The basis for my predictions around Tesla and what Musk is doing is purely my analysis from my understanding of the technological approach and underlying tech.
Musk announced plans for Optimus in September of last year. And he promised to have a prototype ready before the end of 2022. And there were plenty in the media who were laughing, saying it was impossible. The same was true of experts in the robotics industry.
But again, here is that fundamental misunderstanding about what type of company Tesla is.
The reality is that Tesla’s self-driving software can be used to navigate a car around city streets. My underlying thesis was that Tesla’s full self-driving software could be the foundation of the AI for Optimus.
That meant that Tesla wasn’t starting from scratch in its development. Quite the opposite.
I would also argue that navigating a warehouse or distribution facilitiy, a house, or a healthcare facility is a lot easier than navigating busy city streets at 45 miles an hour.
So, Tesla’s task is to take those advancements from FSD and apply it to a machine (Optimus) that is much smaller and needs to perform a different set of tasks.
It was just a matter of integrating it with the hardware. And sure enough, Tesla unveiled a prototype of Optimus in October.
Tesla’s Optimus Prototype
Source: Tesla
VB: And do you think we’ll see mass production of these robots next year?
JB: It’s not impossible. But I predict what we’ll see next year is a pre-commercial product. So, we can think of it as a version 0.9. The robot will be something that can be tested in the field and put through real-world applications.
I predict that’s what we’ll see before the end of next year. And then by 2024, we’ll see the beginning of commercial production for Optimus.
The implications of having humanoid robots working alongside us to perform important, or menial, tasks is going to have a profound impact on society and the workplace.
And it is going to happen in a time frame that very few are expecting.
VB: Thanks for your time, Jeff.
JB: Of course.
Editor’s Note: Check back tomorrow for our next prediction from Jeff Brown. We’ll sit down to discuss the electrification trend, why EVs face a major obstacle going forward, and the companies set to succeed in the new year.