Apple Cancels Work on Electric Car, Ending Decade long Effort
The most recent approach discussed internally was delaying a car release until 2028 and reducing self-driving specifications from Level 4 to Level 2+ technology.
Many employees on the car team — known as the Special Projects Group, or SPG — will be shifted to the artificial intelligence division
@cstross I can't say I'm surprised. All the Apple Car rumors made me wonder "where's the value for Apple here and why should they bother?" and I could never figure out an answer that seemed worth the hassle.
Apparently that's true of people inside Apple too.
@FeralRobots @wordshaper Per Statista (not the best site) Apple has sold about 54 million watches from 2015 through 2022. So maybe 60 million by now, at an average price of say $500 (assuming most folks buy the cheap model), that'd be $30Bn, or maybe $4-5Bn/year.
Yes, that's a SMALL product by Apple standards.
@FeralRobots @wordshaper ... And MacWorld estimate that Apple Watch passed the 100 million mark in December 2020; every third iPhone purchaser in the US also buys an Apple Watch, Watch is in fourth place in sales behind iPhone, iPad, and Mac.
https://www.macworld.com/article/676271/how-many-apple-watches-has-apple-sold.html
@cstross @FeralRobots ...it's entirely possible the Apple Watch exists because Tim Cook runs and was annoyed by how bad fitness trackers were. A petty reason but maybe the case. (And if so I'm good with it, I rather like mine as a fitness gadget)
Possibly they're also taking a long view and just waiting for some new health sensors -- if it ever gets blood pressure or glucose monitoring those sales numbers will get a nice boost.
@wordshaper @FeralRobots Wrist-mounted blood pressure is a non-starter: wrist cuff monitors exist, the accuracy is piss-poor (they depend on height relative to your heart). Non-invasive glucose monitoring would be the killer app, but the FDA just had to post a warning that no such devices exist yet and are licensed ...
@SpeakerToManagers @cstross @FeralRobots right, the tech for wrist mounted BP exists. Unfortunately the tech for *accurate* standalone wrist mounted BP tech doesn’t exist. Aside from a number of issues with calibration and skin tone, the existing tech is… not great. It needs to be externally calibrated regularly, has a significant set of error bars, and doesn’t work for lots of folks.
Samsung, for example, tried this. The results were unfortunately awful.
@SpeakerToManagers @cstross @FeralRobots it depends on the reasons for the inaccuracy. I suspect it’s not a calibration or tuning issue but rather that trying to measure blood pressure entirely optically is just inherently inaccurate.
If you have reliable measurements then machine learning algorithms can translate them into meaningful numbers, but if the underlying measurements have large error bars then no amount of statistical cleverness is going to help.
@cstross These are the people I want to see break out because they are the ones doing something interesting: https://aptera.us/
(Warning: I did kick investment money their way)
@cstross One thing I found out today is that running enough computers for level 4 self-driving *halves* the endurance of an electric car's battery.
Ah, it wasn't quite two:
"Cruise, GM's disgraced "robot taxi" company, had 1.5 remote operators for every one of the cars on the road. They used AI to replace a single, low-waged driver with 1.5 high-waged, specialized technicians. Truly, it was a marvel."
https://pluralistic.net/2024/01/31/neural-interface-beta-tester/#more-7904
@MyLittleMetroid @cstross
Or they got pipped by Tesla
In an alternate world, there might have been an Apple car instead of Tesla, likely with fewer corners cut
@cstross Tesla always I think understood that. Hence staying a tech bro luxury brand - even Elon could work out that the Chinese would eat the non-luxury market. Ford and GM always claimed it would be easier for them to learn tech than Tesla learn to make cars but they face a similar problem with China.
I am guessing cheap Chinese cars in Europe first, then lots of armwaving about tariffs then China building car factories in Europe to keep everyone happy, just like the Japanese in the 1980s
"setting its sights on a fully autonomous electric vehicle with a limousine-like interior and voice-guided navigation"
So they wasted a gazillion dollars on a bunch of stuff that wouldn't work any better in an ICE, aimed at the already thoroughly saturated luxury market segment.
It made good sense for Apple to explore the self-driving aspect. That problem turned out to be as hard as most people thought it was. If it were doable, I think they'd potentially be very good at it.
Take that away, and the EV market is basically a battery market, which is not Apple's thing.
Anyway the space is invaded by hucksters at the moment (not just Elon) so I'd suppose that makes it even less appetizing.
But come back in 10 years.
@cstross Signs continue to indicate that self driving is really freakin hard, and probably not going to happen without some kind of breakthrough. Companies have thrown a LOT of money at this problem, without success.
Neural nets are a very enticing technology - often throwing out results that look great. You think you can tweak the parameters a bit and get great results every time, but alas it ain't so easy even for the big players.
@heavyboots Apple's AI strategy predates the current fad by some years: it runs on the Apple Neural Engine coprocessors embedded in M-series and A-series CPUs, doing stuff like realtime OCR and face recognition on your phone. Some of the LLM work will be rolled into improving Siri, which they first rolled out in 2011.
Hopefully they're not buying into the OpenAI/NVidia driven GPT fad …
@cstross I will say that I would have more faith if Siri wasn’t so bad at its job compared to the Google Assistant.
Also, I think “generative AI” was specifically mentioned in some articles I’ve seen which increases my level of concern. All we need is every iOS/macOS device on the planet added to the pool of things that can easily generate incorrect data to help swamp the internet’s trustworthiness…
Charlie Stross
Dan Sugalski
FeralRobots
SpeakerToManagers
❓ucblockhead
TonyC
Que
Nemo
Very Human Robot
Jon
ronbrinkmann
Angus McIntyre
Merc
Óscar Morales Vivó
sabik
The Penguin of Evil
Jon Hancock
Dave J
josh buermann
Pastor of Muppets
GLC
pr06lefs
heavyboots