Apple's failed 'Project Titan' was a Full Self Driving gamble
The abandoned Apple Car shifted gears several times over the past decade, but ultimately the challenge of full self driving proved insurmountable -- and continues to elude the industry as a whole.
Apple had hoped to produce a car design that would revolutionize the industry, but finally pulled the plug on its electric car dreams. While companies such as Tesla have also hoped to create a vehicle with a Level 5 self-driving system, they have sensibly opted to iterate on the idea with limited self-driving today, and the promise of Full Self Driving (FSD) in the future.
Bloomberg's Mark Gurman reports that Apple's decade-long "Project Titan" ultimately failed because faith in Apple's engineering team to solve the riddle of FSD technology turned into hubris over time.
Apple intended to leapfrog the industry with a car design so capable of full self-driving that it would have no steering wheel and no pedals. It would also look like nothing else on the market at the time.
According to Gurman, the company and its engineering team in 2014 felt invincible. Apple had revolutionized smartphones, tablet computers, and the entire music industry in the recent past -- and was preparing to upend the watch industry.
In hindsight, Gurman noted that a car design that would have "still reflected Apple's design chops and be fully integrated into the company's ecosystem of products," likely made in partnership with an existing manufacturer, would have appealed to the company's fan base.
Instead, all the design work was focused around the advanced concept of a truly "driverless" car that didn't require human control. Gurman cites a source who was involved in the decision-making that compares the plan to attempting to "skip all the early iPhone models and jump right to the iPhone X."
Apple executives understood the folly of their "big bet" a few years ago, but had already spent billions chasing the original vision.
Even if the engineering department had come up with a full self-driving breakthrough, the project's cost, profit potential, and cost to the consumer -- rumored to be over $100,000 -- became unsustainable.
Read on AppleInsider
Comments
If you are running to hit a wall or the fall in a risk… and stop before that happening… Would you call that a ‘failure’ or a ‘wise decision’?
Oh… I forgot! Blaming Apple or showing that Apple fails… brings page views to your site… a.k.a. ad revenue.
“Failure” is many times the best path to success. Not every project that Apple embarks upon results in a new product that is successful. We know for a fact that there have been many projects at Apple that have never seen the light of day. There are also products that had very limited markets, and were discontinued.
Thomas Edison said that invention is 1% inspiration and 99% perspiration. It takes a lot of work to make a product come from nothing to reality. It turns out that Project Titan may have had great inspiration but still needed a lot of work before it could ever become viable product for Apple.
My sister-in-law has been on a 100% electric car in Spain since 2017 and hasn't seen infrastructure problems in spite of living in the country.
China is still pumping out EVs at a crazy pace and, as subsidies (direct and indirect) are phased out there and elsewhere, the market will see some convergence for sure.
FSD will come at some point but requires advances. 2014 definitely wasn't a year for those advances. Were there projected advances of any kind that could lead to FSD in ten years from 2014? I'm not aware of any. If Apple thought it was possible then they were simply being unrealistic.
FSD has of course existed for a while but in restricted scenarios, not the open road. Those scenarios present the perfect opportunity to accumulate data.
In open road scenarios there has always been a line of advance through the different levels. If Apple thought it could leapfrog those levels then, again, it was not being realistic.
FSD isn't an encapsulated affair and isn't only about understanding data from sensors. It requires communicating with other vehicles (for example V2X and ICT) and with road infrastructure. It requires compliance with regulatory bodies too.
All this was clear even back in 2014.
I find it hard to believe that Apple wasn't aware of all of this all the while which leads me to believe that management decided to axe the product for a bunch of possible reasons ranging from economic to technical or more probably a mix of things.
If they really thought they could leapfrog to the forefront of FSD in a few short years then, at best, it was wishful thinking and at worst crazy.
And so we try to pick things that are in their springs. And if you choose wisely you can save yourself an enormous amount of work versus trying to do everything.
... Sometimes you just have to pick the things that look like they're going to be the right horses to ride going forward."
-- Steve Jobs, June 2010 (Has Apple, Inc. Been Betting on the Wrong Technology All These Years? | The Motley Fool)
Apple, in the last 10 years bet on the wrong horse. They bet on autonomous systems while the world was heading towards generative AI. That said, I still think Apple should have made a go in the EV market even if it only had Level 2 autonomy. Like the rest of their product lineup, they could iterate over time. I would be more interested in seeing an Apple Car than a Vision Pro.
Not really. Even the first iPhone leap-frogged everything else that was out there at the time--but instead of having Blackberry as the dominant player to be overtaken, there was Tesla, which was significantly much more of a challenge, since Tesla was already manufacturing what could be described as a very Apple-esque car. So yes, Apple gambled on a moon shot and lost. I'd rather they took the big swing than not. But even if Apple had succeeded, it still faced the issue of low margins in the auto business, which is counter to Apple's whole business strategy of high-margin products.
My question: since full self-driving could be described as one giant AI puzzle, how much of the AI research done for Project Titan might be applicable in some way to the rest of Apple's product line? Or has all the research for Titan simply reached a dead end.
"Here's to the crazy ones. The misfits. Rebels. Troublemakers. The ones who see things differently. Whole some may see them as the crazy ones, we see genius. Because the ones who are crazy enough to think they can change the world are the ones that do."
It was never about making an EV. Apple could do that tomorrow. There are plenty of them on the market already, it's mostly the lack of charging infrastructure (especially in the USA) and a handful of miscellaneous factors that is causing the current decline in consumer interest. Government mandates are pushing more infrastructure but it takes time. There will be better battery technology in the future (this is key).
Apple was really shooting for FSD which requires a conditional use permit from the California Department of Motor Vehicles for public street testing which is really necessary. You can't simulate all driving situations on some dry lake out in the middle of Nowhere, USA. Apple -- like all FSD organizations -- is required by law to disclose basic driving data: miles logged by date, VINs, disengagements (human control overrides), accidents, etc.
In all of its years of testing Apple has never exceeded program mileage of the primary players. The Waymo (Alphabet) FSD vehicle testing program logged about 8x more miles than Apple in 2023. There were long spells of zero activity, sometimes like 6-8 month absences of road testing by Apple, likely brought on by program management changes.
Never once during Apple's time did FSD public street testing approach the point where it was clear they were really, really seriously approaching any sort of threshold where operational vehicles would become a reality.
Nothing prevents Apple from resuming research into FSD at a later time. But for sure, they learned something from programming AI models for FSD. That knowledge can be applied elsewhere.
Most of the FSD test vehicles are modified gas-powered automobiles right now. They don't really need to test in EVs. The most interesting autonomous driving vehicles were Waymo/Google Drive's retired "bubble cars". Those had a stick instead of a steering wheel. It's pretty much a given that Apple learned a lot about user interfaces and controls from any design work they did.
Today we are seeing many conventional vehicle manufacturers back away from total touchscreen control. Things like entertainment system volume controls, climate controls are returning back to the dashboard or steering wheel because they're quicker and safer to operate than tapping through a bunch of touchscreen menus. My gloved hands can operate the fan speed in my 2005 Toyota very easily.
A lot of this research is about how far one can go, sometimes going too far, then understanding that a middle ground solution might be the best right now.
Hell, I routinely see Tesla drivers struggle to parallel park in my city's downtown zone. You'd think the first thing FSD could do would be to park the damned car. And machine-assisted parking isn't new technology by a long shot.
And regardless of where their technology was, an Apple Car never made much sense from a gross margin perspective, just like Gene Munster's beloved Apple Television set. Apple would have to change the fundamental mindset of shareholders on what expected profitability would be if they jumped into the auto market. And the auto industry is heavily regulated like the consumer banking industry. This is why Apple doesn't run their own banks, they don't own cellular networks and they back away at calling the Apple Watch a true medical device.
I'm still quite unsure of the rideshare thing, from both ends: would people ride in an FSD rideshare and would people let their cars be part of the program? A lot of uncertainty on the whole idea. The whole ridesharing utopia idea seems to be one where less cars are sold. How Apple makes profits, revenue or makes a difference in this market is quite uncertain.
The auto industry is in the midst of transition from gas to battery. Lots of two steps forward, one step back type of things are happening with such transitions, but it is all but inevitable. So, it would be great to see another OEM enter the market with their own takes on how to do things. So, I looked forward to Apple entering the fray. It's been boring and slow waiting on EV OEMs to do some cool things, with crazy designs and features. Too bad.
I'm essentially waiting on an EV with V2G and solar surfaces. Like a Lightyear One, but with V2G. All the car makers with solar surfaces have fallen by the wayside, or can't even make it to production. It's been a sad tale so far. I'd love it if I could get a car with 1+ KW capacity of solar PV and V2G on the car. Even better, 2 or 3 of them for my household of 7.
Unfortunately Apple is the absolute opposite of three nouns today. But then again the whole campaign was (great!) brand marketing!
They are now making videos of 'presentations' and projecting them to in house audiences.
'Crazy' would be getting up on stage and putting some real life into it!
A live event for that live audience.
Maybe this year they'll ship Vision Pros out to everyone for WWDC and let them see into the Matrix right from their own homes!
What better way to make that AI love fest take us to the next level.
Generative Adversarial Networks, or GAN's, were developed in 2014, and are at the heart of self-driving vehicle "intelligence". The creation of a GAN back then is also what made the more advanced Generative AI that everyone and his brother is chasing a reality.
So taking that further, yeah, it is possible that Apple could use what they learned developing a GAN (I'll assume they built their own) for their planned self-driving vehicle to expand into Generative AI.
I guess we'll start to see how much of Titan work is applicable, or not, when Apple puts some meat on the bones of its AI strategy at WWDC in June.
Sir: "My 10:07am rideshare customers had three-way sex in the back seat." "My 1:17pm rideshare customers passed around a joint." "Would you like me to search for a custom detailer?"
Some business prospect.