The Last Driver License Holder

Tesla's Full Self Driving is driver assistance software and is not an autonomous vehicle[1]. There are rules governing autonomous vehicles in California. They must report all accidents involving their self driving vehicles[2], miles driven, disengagements,... Tesla's FSD does not follow these rules. Tesla's marketing may imply that FSD is an autonomous vehicle. However, today, in California, these are driver assistance programs and not self driving cars. [1]https://thelastdriverlicenseholder.com/2023/02/17/2022-disengagement-r… [2]https://www.dmv.ca.gov/portal/vehicle-industry-services/autonomous-veh…

Every 24 months of EV production in China destroys 1 million barrels a day of global oil demand, as of this comment, at current rates of EV production. We should expect these rates of production will continue to increase. Total annual light vehicle sales are ~90M units. Globally, over 1-in-5 (22%) of new cars sold were electric in 2024. This share was 92% in Norway, and in China, it was almost 50%. In 2025, it was 1-in-4 (25%). The world has 20-30M units/year of EV manufacturing capacity. Light vehicles and heavy vehicles are solved for with battery electric. Aviation and petrochemical use cases are unlikely to disappear in our lifetimes, but that is a small slice of today's oil demand (lets call it ~15-20% for the sake of argument), and will be the remainder as renewables and batteries eat global energy demand. ~40% of marine shipping disappears when you're not moving energy around the world, the remaining half can burn renewable energy produced ammonia or methanol for fuel. Like bankruptcy, the change happens slowly, and then all of a sudden. China now has more power than OPEC because of the facts in this comment and their vast strategic petroleum reserves. It is easy to say "this flywheel is spinning slowly" before it is spinning very fast, while you're watching it spool. How China became the world’s great oil power - https://news.ycombinator.com/item?id=49238158 - August 2026 Laos went 100% electric vehicles overnight with a drastic approach -…

Why is that interesting? Do you expect them to get into a crash in a simple 15 mile journey? Humans go around 1,000,000 miles between crashes. You literally need to make such a journey over 50,000 times in a row without a crash to be a actual viable autonomous vehicle solution. If you are still worried about the possibility of a problem in a single 15 mile journey you are so far away from adequate to be driverless it is not even funny. As a example, by end of 2016 Waymo was doing 5,000 miles between interventions [1]. It took them until end of 2020 to reach ~30,000 miles per intervention [2]. They only began driverless operations in 2022. 6 years of improvement from a capability level of 5,000 miles per intervention before they thought it was safe enough to hit the roads with no driver. In contrast, Tesla is averaging 500 miles at best [3] and the incident rate of their pilot program corroborates that their "Tesla Robotaxi" version is not materially better. As such, getting a cherry picked 15 mile journey should be expected, but is not at all indicative of actual adequacy for scalable autonomous operation. And despite that, they still failed to actually follow the law for the entire duration failing on a task as trivial as not parking in a gigantic red fire lane. This is also in spite of starting their autonomous driving development by 2016 at the latest. So, after 9 grueling years they are still too immature to handle simple signage such as fire lanes. After 9 grueling…

Zoox did several years of closed course testing and started on public roads in San Fransisco about 1 year ago with just 10 cars. Last fall they did a press event and took a few dozen journalists around for rides and everyone had good things to say about the performance of their vehicles. Their first set of disengagement reports for 2017 had them at 1 every 430 miles, which is worse than Waymo or Cruise, but way ahead of everyone else, and especially impressive given how few test miles they had racked up at the time. It lends credibility to the claims some have made that Jesse Levinson is the brightest guy in the industry. Ashley Vance for Bloomberg did a big puff piece on Zoox a month ago, the video is pretty interesting, it's the first we've been able to see of their prototypes in action, and I had been waiting years to see if they were actually following through with their original vision: https://www.bloomberg.com/news/videos/2018-07-17/zoox-and-its-800-mill… A couple days ago some pics of an unidentified av test vehicle was spotted, and one of the smart guys in my subreddit called it out as a zooxmobile with an new sensor configuration arranged to match the configuration of their protoypes: https://thelastdriverlicenseholder.com/2018/08/21/unidentified-suv-spo…

Now do Surveillance Impact: https://thelastdriverlicenseholder.com/2024/10/27/waymos-5-6-billion-r… “Waymo is using around four NVIDIA H100 GPUSs at a unit price of 10,000 dollars per vehicle to cover the necessary computing requirements. The number of sensors – five lidars, 29 cameras, 4 radars”

https://thelastdriverlicenseholder.com/2025/02/03/2024-disengagement-r… says it's about 40% as good as a Waymo if you use disengagement as a metric.

it is obvious to the user when this happens; the car pauses, the screen shows a message saying it is asking for help. I've seen it happen twice across dozens of rides, and one of those times was because I broke the rules and touched the controls (turned on window wipers when it was raining). They also report disengagements in California periodically; here's data: https://www.dmv.ca.gov/portal/vehicle-industry-services/autonomous-veh… and an article about it: https://thelastdriverlicenseholder.com/2025/02/03/2024-disengagement-r…

Choosing between apples and oranges would be “pursuing the goal of eating”, but you’re still comparing apples with oranges. People are doing 2-3hour drives with FSD 12 with no disengagement. You should check out the recent progress. Meanwhile, Waymo is somehow going to map out in full detail entire countries in order to ever scale up. Btw, the definition of disengagement is up to the companies, not the DMV https://thelastdriverlicenseholder.com/2024/02/03/2023-disengagement-r…

> I wonder, for every one gallon of gasoline I pump into my vehicle, how many were used to extract, refine, transport, store and pump it. Well, almost 40% of all ship cargoes are fossil fuels (oil, coal, methane gas): * https://thelastdriverlicenseholder.com/2022/01/12/almost-40-of-all-shi…

For starters, Waymo reported 1 disengagement per 29,944 miles driven in 2020 to the California DMV [1], while in the talk, Karpathy implies that a Tesla being able to drive around the SF area for 2 hours without a disengagement is unusual. Note that Tesla didn't file a disengagement report because they didn't do any autonomous testing on public roads in California in 2020. There are issues with reading too much into disengagements, but there certainly seems to be a large difference here. [1] https://thelastdriverlicenseholder.com/2021/02/09/2020-disengagement-r…

Note both Waymo and Tesla have sensor fusion systems. I believe both are using CV and (ultrasonic?) radar, Waymo is additionally including Lidar. https://thelastdriverlicenseholder.com/2020/03/04/waymo-reveals-more-d…

Just a small example: the small Tesla has more market capitalization than all German car manufactures combined (bmw, daimler, volkswagen): https://thelastdriverlicenseholder.com/2020/06/01/teslas-market-cap-ex…

> https://thelastdriverlicenseholder.com/2019/08/06/apple-car-... I have some pics of my own if anyone's interested :) https://twitter.com/suryadsuryad/status/1184912385624010752?s=20 https://twitter.com/suryadsuryad/status/1215836960603852800?s=20

I found a picture of the car, the "mattress" approach is in regards to the sensors on the top, not the car design: https://thelastdriverlicenseholder.com/2019/08/06/apple-car-with-new-s…

By disengagement reports cruise is second behind Waymo and these two are way ahead of everyone else. They are testing 160 or so Chevy bolts in California it seems like. https://thelastdriverlicenseholder.com/2019/02/13/update-disengagement…

You're right; 2016 to 2017, not much improvement. Then, twice as good in 2018.[1] (CA DMV stats run on a year from December to November, so the most recent data set is for 2018.) [1] https://thelastdriverlicenseholder.com/2019/02/12/disengagement-report…

> Waymo, widely seen as leading the self-driving vanguard, says it has made progress teaching its software to better filter out “noise” from precipitation. The last Waymo leak claims that it is more than just 'progress', but that 'rain is solved', incidentally: https://thelastdriverlicenseholder.com/2018/09/14/waymo-plans-autonomo…

Recent Waymo leak suggests rain is solved and snow looking good to be solved next year. https://thelastdriverlicenseholder.com/2018/09/14/waymo-plans-autonomo…

One piece of evidence is disengagement rates. In 2017, Waymo led with ~6,000 miles per disengagement. Cruise was in second with ~1,000 miles per disengagement. Also, if you look only at the last few months of 2017, Cruise had improved to around ~3,000 or so as I recall. The other players testing in CA are far behind Waymo and Cruise. Of course, there are many caveats to comparing disengagement rates. These are self-reported by companies, and each company may have different policies about which disengagements are serious enough to report. Also, not all miles are equal. Cruise likes to emphasize that its SF miles are more difficult than Waymo's more suburban Mountain View miles. Here's a nice summary of the 2017 data I found for you: https://thelastdriverlicenseholder.com/2018/02/01/disengagement-report…

mercedes has about 1 disengagement per mile. zoox is a respectable third, actually, behind cruise/gm. https://thelastdriverlicenseholder.com/2018/02/01/disengagement-report…

Tesla's FSD has different approach / tradeoffs compared to dedicated robotaxi services. FSD has to be cheap and energy efficient, run completely on-board, and it must work everywhere. They're trying to do more with less, which has so far been impossible. Their cybercab and robotaxi service will probably work more like Waymo, with a slightly relaxed set of limitations. Some differences compared to Waymo: - Waymo has / can use more on-board compute, from [0] "It has also been revealed that Waymo is using around four NVIDIA H100 GPUSs at a unit price of 10,000 dollars per vehicle to cover the necessary computing requirements." - Waymo uses remote operators. This includes humans but can also have remote compute. - Waymo's neural network model can be trained / overfit on specific route or area. FSD uses the same model everywhere. - Waymo's on-board hardware can use more energy, because it's possible to charge the battery between trips. - Robotaxi services charge customers per mile, so it makes sense to run longer routes which are also easier to drive, i.e. the routing algorithm can be tuned to avoid challenging routes. This would be possible to implement on FSD too, but it seems that FSD drives fastest route. [0] https://thelastdriverlicenseholder.com/2024/10/27/waymos-5-6-billion-r…

> If you've ever tried to squeeze yourself under the kitchen sink, work one hand behind the 5" gap around the garbage disposal, work the other hand around the tiny gap between the garbage disposal and the main sink, try to coordinate the 2 hands to tighten the faucet, then I'm sure you'll agree with OP. What's hard for me doesn't tell you anything about what's hard for a differently shaped body, and robots don't need to look like us — two left hands is easy, extra joints is easy. Most robots don't look like us, one way or another. Famously Spot, but there are as many sizes and shapes as there is commercial demand for them. Humanoid ones are popular right now, but I grew up hearing about how that was "actually a bad shape for a robot" (was the past just sour grapes, or are the current humanoid ones CEOs watching too much SciFi?) > We don't even have a commercial robot to unload and store dishes from the dishwasher nor unload and store clothes from dryer. Both dishwashers and clothes driers are special purpose robots designed to solve problems without having to first solve human-level sensation or dexterity. > What makes you think robotics will be solved in a century, let alone ten years? Have you seen demo videos from the last 12 months? The videos are sped up, indicating they're compute-bound; the accuracy of joint positioning, the strength, etc. is already superhuman whenever we want it to be. Quality of vision, that's the hard part. Cars can only lose their steering…

I genuinely don't think this is true. FSD is better than a lot of people make it out to be, but I'm pretty sure they're still leagues behind Cruise who is behind Waymo. I don't think Tesla publishes their data, but from the 2021 report [1] Cruise is at 42,022 miles/disengagement and Waymo is at 7,800. In fairness, Waymo switched their car base car, which seems to be causing issues. Waymo was at 29,944 miles/disengagement in 2020. The videos I've seen of Teslas appear to have a much lower miles/disengagement number. Even if we presume they're going 60mph the whole time (they're not), they would need to have 1 disengagement per 130 hours of video just to keep up with Waymo's new, much lower stats. They'd need 1 disengagement per 700 hours of video to keep up with Cruise. With a more realistic estimate of an average 15mph, you'd have to quadruple those numbers. There's some wiggle room, because disengagements and whether they're reportable is somewhat subjective. The delta seems large enough to say that Tesla isn't particularly close, though. 1. https://thelastdriverlicenseholder.com/2022/02/09/2021-disengagement-r… (the dataset is available from the CA DMV, but it's in CSV so I used this because graphs)

Will they be the current cars, or cars with the new next-gen hardware that are substantially safer? The cars that end up driving could signal a lot to the industry. New lidar was spotted a few months ago and more recently on the Jaguar cars and the CEO said it’s an order of magnitude better: https://thelastdriverlicenseholder.com/2019/04/21/new-waymo-lidar-spot… https://medium.com/waymo/waymo-iaa-frankfurt-2019-b3cca36d8479 One would imagine this early announcement comes in reaction to their recent “valuation” cut: https://www.cnbc.com/2019/09/27/waymo-valuation-cut-40percent-by-morga… If Waymo starts driving without safety drivers en-masse, even on small fixed routes, it will signal certain takeover metrics have been hit. It will be interesting to see if Waymo uses existing cars (the ML and planners just got better) or if they use new hardware and signal a certain lidar improvement is necessary for achieving a viable takeover rate. The signal won’t be clear but should be there. And then it’s up to the rest of the industry to show how fast they can react to the change.