One thing Tesla and Comma.ai overlooked in self-driving
Disclaimer: I'm not a AI Engineer, but only an enthusiast, everything that you read here should be taken with a very big heap of salt
Remember back in 2019 when Autopilot was all the rage, Tesla stock was “too high” and Elon famously proclaimed that “Lidars are stupid, and Waymo is going to die”. I’m paraphrasing of course, but you can watch the 2019 Tesla’s Autonomy Day here, it’s basically what was implied.
Now, six years later you wake up in the morning and get hit with a different narrative:
”It’s Waymo’s World. We’re All Just Riding in It. Google’s driverless-taxi company just cracked 10 million rides. If you haven’t taken one, you will soon.”
Most importantly let’s notice the trajectory:
”Waymo was doing 10,000 paid rides a week in August 2023. By May 2024, that number of trips in cars without a driver was up to 50,000. In August, it hit 100,000. Now it’s already more than 250,000.”
Holy-moly, Waymo said they have doubled the trips in five months!
Now let’s look at the other camp:
”Tesla’s Austin, Texas, robotaxi service will kick off with 10 vehicles and expand to thousands, moving into more cities if the launch goes well.”
Ten cars and “if the launch will go well”. Of course we don’t know all the details, and judging by very scarce amount of information available, it’s unclear what exactly we will see. Hopefully not a robotaxi in a A → B tunnel!
I’d say Musk is probably rushing it, Waymo is scaling and Tesla is launching 10 cars service. I’d also argue that it will probably not be very safe, otherwise nothing would stop Tesla from rolling out “the safe” autopilot to car owners today, unless you think that something will change in one month and we are getting some kind of breakthrough in software, so it seems like taxi will run today’s software.
I mentioned comma in the title, while they are not launching any taxis, they are effectively in the same basket with Tesla regarding the software, George said himself that they are always slightly behind Tesla.
So what was overlooked you ask?
Waymo was frequently mocked that they are heavily geofenced, and “it cannot drive anywhere”-mentality. I have to say that I was also a subscriber to this camp and it seemed like universal “no map”-approach was to go. I was cheering for Tesla Autopilot team.
But then I noticed something:
83% of the U.S. population lives in urban areas, up from 64% in 1950. By 2050, 89% of the U.S. population and 68% of the world population is projected to live in urban areas.1
https://css.umich.edu/publications/factsheets/built-environment/us-cities-factsheet
Yep, nothing less than 80/20 rule! God damn it, we got fooled. Sorry, but why do we need an Autopilot to drive everywhere? Yet solving the last 20% of self-driving might be 80% of effort.
Waymo/Google are happily chasing densely populated areas.
Expensive LIDARs?
1. Waymo is a taxi, not a private car, therefore cost of LIDAR and other fancy tech is spread across better, because unlike your car that sits in a garage, their taxis really can run 24/7 with only stops for tanking
2. LIDARs are getting cheaper too. As one person famously said “If you are good at manufacturing, you can produce at cost of raw materials plus IP”
Hint: Elon said this
Another point is that city driving seems to be a safer in general, there is much less chance if things go wrong when you are in a taxi driving 30-40mph, than cruising with Autopilot 60+mph and hitting a truck on a freeway.
I don’t really remember reading about bad Waymo accidents, but I definitely remember couple of people getting killed in Teslas.
Then there is a “stack” argument, while I’m not really strong here, but my understanding is you can always re-train, re-equip cars as needed. Whatever Waymo is doing, it seems to be working, even if it’s not “end-to-end” model.
If Waymo will be Google of robotaxis taking the 80% of the market, Tesla might be Bing?
As far as the money goes, yeah Waymo is burning cash, but what’s money to you when you got Google’s ad money printing press anyways? I guess scaling fast now will show if you can run this business profitably.
Good luck to all companies, excited to see what’s waiting for us in Austin from Tesla!
I feel like once you have urban areas down, the rest are relatively easy. There's so much more negotiations happening with pedestrians, cyclists, and other motorists, all contributing to the chaos and ambiguity.
I'm exited for the future, I just hope cities remember that urban areas should be built for people, not cars. I'm also hoping this tech will eventually come to buses/street cars. That'll really alleviate the driver/operator shortage.
Elon is nowhere near as smart as he and his lackeys claim. He depends on the skills of his employees, and demands extraordinary sacrifice from them, often expecting 100+ hour workweeks with little to no additional compensation or recognition beyond the original agreement.
He treats his workers as though they should feel privileged just to be part of his vision, regardless of how exploitative or unsustainable the conditions become. Loyalty, burnout, and personal cost are rarely rewarded, and gratitude is nearly absent from the culture he cultivates.
In recent years, he has burned countless bridges, alienated former allies, and revealed a management style so toxic and erratic that much of the once Elon-enamored tech workforce now wants nothing to do with him.