- mainstreet memos
- Posts
- Ghostride the Whip
Ghostride the Whip
autonomous
On Friday, Uber announced that riders in Austin and Atlanta will be able to hail a Waymo robotaxi through the app in early 2025 as part of an expanded partnership between the two companies.
If you’ve been to Phoenix recently, you understand how common Waymo cars are in the city
100,000 weekly rides take place between Phoenix, SF, and LA
700 Waymo vehicles operate in Phoenix, SF, and LA
Operations have resulted in an estimated $2B loss for Google… for the first half of 2024
Waymo’s expansion into other markets feels like the beginning of a dystopian fever dream
Would you ride in a driverless Waymo car? |
For the time being, I will suspend judgement and cruise through the controversial history behind Waymo, the market for autonomous vehicles, and what’s next
back to the future
The race to autonomous cars starts in 2009 with Google incubating driverless technology in its X Lab before spinning Waymo out to stand on its own as Alphabet subsidiary.
Check out Waymo’s 2015 model:
A few years later, the company decided to modify commercially available models rather than re-create the wheel. Today, a white Jaguar F-Pace is the car of choice
Now around this time, other autonomous projects began to pop up in droves. Incumbent car companies and tech companies alike saw the potential for hundreds of billions of dollars in revenue down the road.
GM Cruise: launched in 2013
Amazon Zoox: launched in 2014
Apple Car: launched self-driving project in 2014
Uber: launched self-driving project in 2015
At the time, Uber just raised $1.2B at a $17B valuation.
But Uber founder Travis Kalanick was not happy.
One of his most notable investors, Google, was planning to crash into Uber’s plans for autonomous ride-hailing. When he contacted Google founder Larry Page, Kalanick was met with what he considered to be disrespect. Before the incident, Kalanick considered Page to be a mentor
Internal Google documents showed plans to “consume” all of Uber’s profits by 2025.
fly home travis
Kalanick enters his villain era and goes full throttle in developing autonomous tech to compete with Google’s Waymo project. His first move is hiring engineers from Carnegie Mellon University in early 2015. After realizing the CMU grads weren’t going to cut it, he starts talking to a man by the name of Anthony Levandowski. This is where things take a wild turn
Levandowski was a co-founder of the Waymo project at Google who lost faith in the initiative circa 2015. So him and Travis Kalanick scheme up a way to work together to transfer Google’s Waymo trade secrets and tech to Uber
They settle on the following arrangement:
Levandowski would launch a startup named Otto (Jan 2016)
Otto was founded by Levandowski, Lior Ron, Claire Delaunay, and Don Burnette, + 11 other Google employees
Otto retrofitted big rig trucks with self-driving kits
Acquired by Uber in July 2016 for $680M
levandowski recently founded the first “ai church”
Safe to say the attempt to sabotage Waymo did not work well
Uber wound up firing Levandowski in 2017 after he refused to relinquish his constitutional rights against self-incrimination when questioned about whether he stole trade secrets before leaving Google.
Kalanick was pressured by investors to step down as CEO a month after Levandowski’s firing, partly because of concerns about Waymo’s lawsuit.
In 2019, Levandowski was charged by the DOJ for the alleged theft of trade secrets from Google's self-driving unit Waymo. Charges alleged Levandowski downloaded thousands of files from Waymo's predecessor, Project Chauffeur.
In mid-2020, Levandowski was ultimately sentenced to 18 months in prison for “the largest trade secrets crime” that Judge William Alsup had ever seen.
Pardoned by former President Trump on his last day in office, Jan 2021
crashing out
So how does Uber go from legal disputes with Google to working alongside them?
Post-Kalanick, the company winds down their internal autonomous project in 2020. A big reason was because one of their cars fatally struck a pedestrian.
Self driving vehicle company Aurora acquired the employees and technology in an all stock transaction. Uber also invests $400M into Aurora and CEO Dana Khosrowshani joins the board of directors
After the transaction, Aurora was worth $10B with Uber controlling 26% of the company.
It’s currently worth $8B on the stock market.
With past tension resolved, Uber and Google were free to collaborate, and the partnership materialized in mid-2023, bringing driverless ride-sharing to the Phoenix area.
On one hand, it seems like mass adoption is just around the corner
But there is an element of fragility baked into the autonomous race. Driverless vehicle crashes are high profile events that bring scrutiny and sometimes permanent closure for projects
GM Cruise spent most of this year recovering from an incident in which a driverless Cruise vehicle hurt a pedestrian. The company’s permit to operate in California was suspended and it was probed by the DOJ, SEC, and NHTSA. The event led to internal stock price cuts for the Cruise project (-50%). Just last month, Cruise announced that their robotaxis would be available on Uber in 2025
Amazon Zoox was under investigation (May 2024) by the NHTSA for two rear end crashes by its vehicles, one in Nevada and one in San Francisco. A motorcyclist suffered minor injuries in one of the crashes
Waymo recalled software in all of its vehicles after two cars in its fleet crahed into a pickup being towed in Phoenix (February 2024). A similar recall took place after one of its cars crashed into a pole while trying to pick up a driver
From a technology standpoint, experts seem to believe we’re barely halfway to full potential.
For reference, Tesla’s self driving capacity is Level 2 autonomy, relying only on cameras and requiring human drivers to be attentive.
Waymo is Level 4 autonomy, with cars taking in information from cameras, radar sensors, and lidar sensors. An AI component is being developed by most Level 4 producers, including Waymo. No human attention is required.
Lidar stands for light detection and ranging. It uses laser lights to measure distances and construct maps of the environment. By emitting lasers and measuring the time it takes to reflect back to the sensor, it can create 3D maps and keep autonomous vehicles aware of surroundings more effectively than just a camera.
Level 5 is optimal autonomy.
Reply