Apple Reports Self-Driving Car Disengagements to DMV, Earns Worst Rank

[apolloa]

I see it’s going well for Apple then.....

A few years in, countless billions spent I’d imagine, and the worst performing self driving tech.
They really should ditch this whole idea, stick to computers and consumer electronics, expand in that health sector like they want to..

This will push its share price down I would think?
[doublepost=1550042883][/doublepost]

Dang kinda skewed to not have Tesla on there.

They are not self driving despite what people think. In the US alone you are not legally allowed to let the car drive itself I believe, because it’s crashed too many times doing so, Tesla will advise you to not let the car drive itself.

 

[otternonsense]

Are they letting Siri navigate?

 

[Norwegian_Imposter]

These statistics will be more or less depending on how many cars and trips they make right? could it not be that Google simply have more cars?

 

[mejsric]

Should Apple should got the highest rank now?

 

[cwerdna]

These statistics will be more or less depending on how many cars and trips they make right? could it not be that Google simply have more cars?

Last year, from https://www.dmv.ca.gov/portal/wcm/c...-97f6f24b23cc/Waymofull.pdf?MOD=AJPERES&CVID= via https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/disengagement_report_2017, unless I miscounted, Waymo had 75 vehicles for 2017 in CA doing autonomous mode testing on public roads covering 352.5K miles w/63 disengagements.

 

[juanm]

Yes, my question too: possibly Tesla is not is attempting the level are the others are testing for??
....
OK, someone answered it: NVIDIA equals Tesla

This is autonomous driving, different from what Tesla offers at the moment.

Afaik Tesla and Nvidia have entirely separate programs (even though Tesla uses Nvidia hardware).

Apples to oranges, at this point.

 

[WatchFromAfar]

Fascinating! Looks like it's a long road ahead for Apple's auto ambitions without a major acquisition.

Whatever happened to "People who are really serious about software should make their own hardware"?

 

[Michael Goff]

Cut arms and legs of google assistant and it won't be much better.

I'm not following you here. What are you trying to say?

 

[JoshBoy]

You’re driving it wrong

 

[michelepri]

Maybe they were using data in Icloud

 

[twistedpixel8]

Next week: Apple purchases autonomous driving company Waymo. MacRumors reached out to Apple for a statement. Apple responded, “Apple buys technology companies from time to time, and we generally do not discuss our purpose or plans.“

 

[aylk]

The short answer is Tesla doesn't have a self driving car of that capacity.

If this answer confuses you, its because you fell for a businesses advertising/marketing.

nVidia is Tesla.

 

[dilbert99]

Just wait until Apple Car Series 2. 50% fewer disengagements and 50% safer!

Wait for Apple Car series X, its so thin it looks like a skateboard
[doublepost=1550049126][/doublepost]

Obviously this list is only for those in the Cupertino area, but worth noting that Apple hasn't (at this stage) suffered any fatalities from their self-driving efforts, compared to the Uber incident 11 months ago, and the 3 Tesla incidents. I don't even use a reverse camera or blindspot monitoring, nothing bets having the actual skills to drive properly yourself!

But how many miles have Apple cars covered?

 

[aylk]

Such a weirdly specific news item. Where exactly is Waymo driving those cars? Actual highways or testing parks of some sort. 11000 miles per disengagement is stupendous if it is in real life situations. Didn't realize humanity had reached there.

I see Waymo vans driving around the Bay Area very frequently.

 

[theluggage]

There is something missing about this report

Yes - what’s missing is any data that you can actually form any conclusions from - one way or another.

Want to game the “disengagement per mile” score? Have a bunch of cars driving up and down the freeway, notching up the miles in the sort of uncomplicated driving conditions that current lane-assist and auto breaking systems can pretty much deal with already - in what are already the safest driving conditions (certainly per-mile) - so you don’t need to worry too much if your safety drivers are engrossed in Facebook.

Not worried about crude statistics? Focus your testing on “problem” areas like town centres with wall-to-wall parked cars, pedestrians, and wired road layouts - and have your safety drivers on a hair-trigger, reporting every minor imperfection.

Of course, Apple’s system could just be rubbish - but you can’t tell from these statistics.

 

[Mildredop]

This isn't too surprising. Apple's probably going to be making missteps and all. It's a fairly new field they're entering in.

Whereas Waymo have been at it for the last forty years...

 

[TheFluffyDuck]

Recently Apple have made garbage computers and software with more bugs than new features, all at a premium price, with a release window for some machines rapidly approaching the criticised gap between XP and Vista. So having a car that loses its way and runs off the road, seems pretty poetic for a 2019 Apple. Why break the habit of a decade?

 

[Reason077]

Accidents/incidents per million miles might be a better gauge of autonomous-driving safety than disengagements.

 

[w_aldo]

For me the shocker was NVIDIA was testing a car lol

Why? Tesla uses NVIDIA hardware to power its autopilot feature.

 

[gnasher729]

Literally a 10,000x spread between Waymo and Apple, that's insane. I always wondered how Apple could hope to catch up to Waymo, and for now, the answer is they don't.

Disengagements doesn't mean forced disengagements. Car is driven by a driver, turns "engagement" on to see how the car manages a lane change in a left curve, and disengages after that curve. Because that's what he intended to do. You get lots of disengagements that way.

 

[LooZpl]

I see a misunderstanding about Tesla.

Tesla did not test any vehicles on public roads in California in autonomous mode, as defined by California law

Tesla has several hundred thousand vehicles on the streets with autonomous driving systems that operate in shadow mode.

Their tests run in a completely different way and Tesla has a giant achievement in the field of autonomous driving and this will be said by everyone who uses it for everyday driving.

This cannot be compared to a test drive that complies with California law.

What's more, Tesla is the only one (?) using only cameras and radars, not LIDAR, and you can turn it on everywhere - from California through U.S. towns to European villages. Most other systems, including WAYMO, work only on previously scanned and implemented roads.

 

[Smartass]

Apple should just enter Tesla with big share and then work from there on...

 

[Fancuku]

It doesn't surprise me. This is the company of Siri and Apple maps.

 

[G5isAlive]

Why? Tesla uses NVIDIA hardware to power its autopilot feature.

Well aren’t you the well informed? Because it’s mostly about the software. Besides rendering what other major software project is NVIDIA involved in?

But mostly I was joking around.

 

[metallicpail2]

I think that without knowing what they tested this graph cannot be used for any type of judgement. I am sure any of those self driving cars could go on the highway and drive thousands of miles without much interference. So unless all systems took the same route, you can't judge anything.