Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Cancels Self-Driving Vehicle Testing Permit in California After Abandoning Car Project
- Thread starter MacRumors
- Start date
- Sort by reaction score
also, you really need to learn more about the tariffs in the first place. If Apple made a car the tariff wouldn't really apply to it just like it doesn't really apply to Tesla
www.theverge.com
US to raise tariffs on EVs, batteries, solar cells, and computer chips from China
Soaring tariffs lock Chinese EV manufacturers out of the US.
www.theverge.com
You're right! This is something that people miss about AI-anything - it all lacks human intuition. That's not something that as far as we know can be calculated or simulated.
Apple cancelled their car in February 2024, upon learning about the upcoming 100% Chinese import tariff. It was officially signed into law in May 2024.
It was a very smart move on Apple's part. The only other option was to manufacture their car in the US. That was likely not a viable option Apple wanted to take on as it would cost $$$$$$$$ to set up, and then there would be recurring costs for US-level wages/benefits/etc to keep it going.
What does that have to do with anything?
SURELY even Apple wouldn't be stupid enough to build cars in China.
That was their best option in order to make a viable business case.
You may not know China produces 2/3rds of the world's EVs, and is a world leader. Tesla has a huge manufacturing operation in China producing almost 1 million cars per year.
"Human intuition" kills 40k+ Americans a year. It's not really an advantage. And it's not really that hard to detect "that car over there is behaving somewhat erratically, maybe give it a bit of extra space". I don't know for a fact that's part of the algorithms, but it would be insane for it not to be.
The real problem is when you're dealing with Turdla's system that doesn't know what the side of a tractor-trailer or a motorcycle looks like, and doesn't use radar or lidar as an extra check for "something is directly in front of me".
Last edited:
Building in China was never going to be a viable business case for cars.
Smart would have been US, Canada, or Mexico. Not stupid would have been Japan, Korea, or Germany. China would have been idiotic.
It was very viable before the tariff was imposed. Tesla has cars made in China, for example.
Only that they sell in China. And they're known to be even lower quality than the scrap they build in the US.
I agree sociopaths make great CEO’s not so good for the country or consumers. Musk is following the path of another great innovator Howard Hughes, won’t turn out well for him either.
Except he's not an innovator, he just had enough money to make some lucky buys. And with an attitude more like Henry Ford.
You want to talk about Waymo as if there haven’t been instances of Teslas running into things, including parked tractor trailer, something much larger than a pole and resulting in loss of life. As I said before, fusion of data from multiple sensors is hard. It won’t be satisfactorily solved for some time but unlike Musk, no one seriously working this problem has claimed it is ready for prime time through deliberately misleading naming of features (Autopilot, Full Self Driving) and actually selling/charging for them.
I didn’t realize having you or I take a few rides is enough to certify an autonomous driving system is safe. That’s awesome! I’m sure auto manufacturers will be thrilled to know that they can just stop spending R&D money as soon as we get a few rides without abnormalities and call it done. Not to mention all the money I could make hiring myself out to them.
Dude, the tariffs have nothing to do with it as they would not apply here. I've posted why, did you not check it?
Tariffs are for Chinese companies, US. Tesla still makes cars in China and you don't see tariffs there either.
Tariffs are for Chinese companies, US. Tesla still makes cars in China and you don't see tariffs there either.
I predict, when the time is right, the Apple Car project will be back on again… 😉
The reason Waymo is in trouble in the long term is not that their technology is behind. It might be the best. The problem is the price. Each Wamo taxi costs about $300K to build. Tesla plans a $30K taxi.
LIDAR really does work well and is easier to process than video. This is why Wamo uses LIDAR. But those lidar units cost $75K each while the typical cellphone-like camera costs about $15. Cameras are so cheap that Telsa can put like 8 or 10 of them on one car.
Price really does matter
That said when you remove the steering wheel from a car you REALLY have to have very excellent FSD. Only a full Level-5 system will do and Tesla is not close to this. I'm thinking it might be available in the 2030s, not the 2020s.
Last edited:
I like how you look at the bright side of life!
For that to happen the market will need to change. Apple will never be a price leader. Apple sells at the higher end of the price range. The market for over $100K cars is too small for Apple.
A better fit might be robotics, real robotics, not an iPad on a motorized base. A machine that can bus tables and clean bathrooms and fetch an Amazon package from the front porch and pick up trash along a roadside. There are MANY robot mechanisms. Tesla showed off their last night but these robots were just puppets that have human remote control or they play back recorded dance motions. They are as dumb as a toaster oven. Apple needs to work on the AI to make real-robots "work" for consumers. It would be a long-term project.
Lidar prices have dropped significantly, 3d sensors are now in the hundreds, not tens of thousands. And visible light has SEVERE limitations that make cameras utterly unviable for self-driving as the only data acquisition option. We use it as human drivers because it's all we've got, there's no reason to handicap computers like that.
And Turdla isn't even using cheap radar sensors. That's a big part of why they're murdering motorcyclists. I know when I'm riding I now have to stay as far away from Turdlas as possible.
Bingo. Apple aim is to get into the car without the user getting an entire car that is made by them. I also think Apple plans were to integrate hardware into other manufacturers cars as well, kind of how some cars have tech from NVIDIA as the brains for autonomous driving. Building the Apple car was simply a means to an end.
Yep.
There was NEVER an Apple Car as an actual plan for a shipping product. I think they did at one point talk to Hyundai as a potential partnership, but it never got anywhere. And Apple is not and never was a car company, they don't have a clue how to do that. I mean, somebody in this thread seriously suggested that they'd just outsource it to a manufacturer in China, which was NEVER a viable option for a car, no company is going to be able to destroy that many American jobs when both major political parties are ready to shut that down.
No, waymo is way ahead of Tesla. Elon doesn't understand metaphysics and thinks of the self in terms of cause and effect which is self-refuting since the self assumes free will. He believes given enough data, something like human consciousness and intelligence can arise, eliminating the need for all these sensors since humans just need eyes and a mind to drive. He's never heard about the hard problem of consciousness and is basically a materialist and interprets data through this paradigm. Extrapolating from that he believes consciousness arose after organisms after 'evolving' gain enough data that gave rise to consciousness. This is the problem with that, meaning is not inherent in matter. It is a mental category we attach to the external world, similar to how cause is something we attach to ball a hitting ball b and 'causing' ball b to move, but cause isn't in the sense data, we apply cause that event in the mind. So data in of itself is meaningless without consciousness already there. So essentially he thinks meaning and consciousness arose from non-meaning. This is a contradiction. AI will always be something that is deterministic and moves according to an algorithm and its code. It cannot replicate the human mind as humans are essentially spiritual. If you deny humans are spiritual then by definition we are material objects that move according to cause and effect, which would make knowledge impossible, since justified true belief assumes our ability to freely weigh evidence. What elon presented will not at all be ready in 2 years. This is just a show and he is still holding out hope just given enough data his AI system will have human level cognition. Even Waymo knows they are no where near level 5 autonomy and need background remote control in case, and this is with all their sensors and everything.
Last edited:
FSD 12.5 is observable.
Legacy Automakers have nothing that currently competes against FSD 12.5 - observable
Waymo loses billions and can't even scale fast enough - observable
Waymo is driverless. Waymo SHOULD NOT be making these dumb mistakes if they are running it driverless. Tesla's current FSD requires supervision because Tesla thinks it will occasionally make dumb mistakes until they improve the software.
That makes no sense. Then why is Waymo running driverless service today if it's not ready for prime time?
It's not misleading. Autopilot refers to pilots in planes using a system to control the path of vehicle without constant manual control, however human pilots are still required.
Tesla has always called it Full Self Driving (Beta) or Full Self Driving (Supervised). Not misleading.
No one is claiming to certify in any official capacity. That's the government's job.
However, a person using Waymo and FSD personally has more weight in the argument compared to someone who has never used Waymo or FSD. This is just a fact.
I don't believe so. "Best" needs to be useful to as many people as possible. Their technology will never scale to cover the globe so it can't be the best. FSD will work in places that Tesla has never traveled before.
Waymo actually processes video along with LIDAR. They need to fuse the sensors together which is far more taxing and troublesome.
You're correct in lidar being very expensive.
Level 5 system has no say on how much better a system should perform compared to a human. It just says it should be better. Statistically we all know that all level 5 systems will always encounter some form of crash for the foreseeable future.
If your target is it just needs to be 2x safer than a human (which means it'll still crash, but half the time compared a human driver) easily 2025 is achievable. If your target is 5x safer, that's 2026-2027. etc...