FSD rewrite will go out on Oct 20 to limited beta

[AlanSubie4Life]

new 1 hour and 24 minute video...also did not know this: "Tesla called me directly and explained that the AEB (Automatic Emergency Braking) has been disabled while using the FSD BETA toggle... I haven’t seen anyone bring this up, it’s very important IMO."

Watch Tesla Full Self-Driving Beta Drive for 1 Hour and 24 Minutes [4K VIDEO] - TeslaNorth.com

Why does everyone insist on using the worst possible camera angles for these videos?

Something like this is REALLY easy to set up. If you want, you could move it slightly further back, to allow capture of the mirrors & side views as well. Obviously multiple camera angles would be nice, but doing production to merge the TeslaCam data with the video from inside the car isn't really expected for quick videos. Someone should do it though (assuming TeslaCam is enabled)!

BTW, this was the video that was posted earlier in this thread (guess it could have been a different related thread), showing the man with a bunch of shopping carts, which the car mistook for a truck on the sidewalk, and the one with the last minute navigation reroute which caused the car to try to turn left from the rightmost lane.

It also detected rings painted on the road in the middle of an intersection with traffic lights and a protected left turn lane as a roundabout, a mistake even an experienced driver can make.

 

[Silicon Desert]

new 1 hour and 24 minute video...also did not know this: "Tesla called me directly and explained that the AEB (Automatic Emergency Braking) has been disabled while using the FSD BETA toggle... I haven’t seen anyone bring this up, it’s very important IMO."

Watch Tesla Full Self-Driving Beta Drive for 1 Hour and 24 Minutes [4K VIDEO] - TeslaNorth.com

Interesting video. Would be nice in the future to be viewing lower so that rear view camera is not in the way, thus being about to see farther down the road.

 

[MP3Mike]

Is this the first video with City Streets driving in the rain?

 

[qdeathstar]

Is this the first video with City Streets driving in the rain?

good video.

 

[Nuclear Fusion]

they seem to agree with you, herd immunity and such....

With nearly quarter million dead & 1 in 40 infected, good thing none of them pothole socialists got to lockdown ma freedoms...yeehaw!!!

 

[clydeiii]

I can buy a FSD Tesla for 55-60k USD. It’s impossible for me to buy a Waymo car that drives me to work. Therefore the Tesla is infinitely more useful to me, even if it isn’t as perfect as the imaginary Waymo car I wish I could buy.

 

[Mo City]

Not sure why you are saying "with all due respect" referring to me. I actually agree that a L2 city system with just a couple disengagements per commute could still add tremendous value to customers. I've even said before that as it stands today, Tesla's "L2 that works everywhere" will benefit more people than Waymo's L4.

This is the bottom line for me. I don't discount what Waymo is doing and I wish them the best. But Tesla's are driven every day by (soon to be) multiple millions of people.

This isn't a robotaxis or bust scenario.

 

[GZDongles]

How would a Level 2 robotaxi work?
I struggle to understand the benefit of a beta autonomous vehicle. This is the quick math that I have in my head. Human drivers make errors which result in accidents approximately every 500k miles. Let's say that the current build of FSD makes mistakes which would result in accidents every 100 miles. That means the human monitor needs to catch 99.98% of mistakes for the car to be safer. Let's say that by the end of next year beta FSD makes a mistake that would result in an accident every 10,000 miles, the human monitor still has to correct 98% of mistakes! What level of vigilance does it require to do that?
Watch the beta FSD videos and observe how quickly the human monitor needs to correct the car while driving on city streets. You've got to be hyper vigilant. I just don't see how that will change as the system gets better, catching 98% of mistakes is not much easier than catching 99.98%. Until it gets to human level performance I don't see the value other than as entertainment. Many people certainly do seem to be enjoying using this feature so that benefit cannot be denied.

I agree, I don't see any way to have a Level 2 robotaxi. In fact, I think the business case for using a fleet of autonomous Teslas as robotaxis is questionable until you are using a car with much lower acquisition costs. I know Tesla promoted the robotaxi idea back in the early days of the Model 3, but I think there's close to zero chance that happens in the next few years, and a decent chance it never happens.

To expand a bit, I don't think the problem is technical feasibility, it's more a question of economics. Autonomy removes the labor component of an Uber or Lyft and replaces it with higher upfront costs for autonomy and electrification. If compare the per mile and per hour operating costs of a $25k Honda Civic versus a $55k Tesla Model 3 with FSD, the Tesla is in a huge hole to start because of the higher purchase price/capital costs. The average Uber driver makes $15/hour when they are active. It's going to be very difficult to use expensive vehicles like the Model 3 to have a competitively priced ride share service when the only economic benefit is the labor reduction from removing someone making $15/hr. Even without labor and without the price of FSD, a Honda Civic has lower ownership costs than a Model 3 over a 5 year period anyway, so the lower operating costs of an EV never quite offset the purchase price.

Anyway, I didn't mean to go on a tangent about robotaxis, but I do think the economics of ride share make the robo taxi discussion a red herring for FSD.

However, I don't agree that Level 2 features are simply "entertainment". They have value in making the driving experience safer, incrementally easier, and more convenient. Cruise control is the same value proposition on a much smaller scale. It makes driving less effort without compromising safety. I think there's a huge area between cruise control and Level 4, and that's where FSD fits in. For me, FSD was worth the $7k I paid for it even without autosteer on city streets, and any additional Level 2+ features I get are a bonus. I understand not every one calculates value in the same way though.

 

[diplomat33]

This is the bottom line for me. I don't discount what Waymo is doing and I wish them the best. But Tesla's are driven every day by (soon to be) multiple millions of people.

This isn't a robotaxis or bust scenario.

It's two sides of the same coin. Tesla cars are not true FSD yet but they do offer some safety and convenience to a lot of people. Waymo is not yet available to a lot of people but they have true FSD. Which side you prefer depends on what you are looking for.

 

[Daniel in SD]

I agree, I don't see any way to have a Level 2 robotaxi. In fact, I think the business case for using a fleet of autonomous Teslas as robotaxis is questionable until you are using a car with much lower acquisition costs. I know Tesla promoted the robotaxi idea back in the early days of the Model 3, but I think there's close to zero chance that happens in the next few years, and a decent chance it never happens.

To expand a bit, I don't think the problem is technical feasibility, it's more a question of economics. Autonomy removes the labor component of an Uber or Lyft and replaces it with higher upfront costs for autonomy and electrification. If compare the per mile and per hour operating costs of a $25k Honda Civic versus a $55k Tesla Model 3 with FSD, the Tesla is in a huge hole to start because of the higher purchase price/capital costs. The average Uber driver makes $15/hour when they are active. It's going to be very difficult to use expensive vehicles like the Model 3 to have a competitively priced ride share service when the only economic benefit is the labor reduction from removing someone making $15/hr. Even without labor and without the price of FSD, a Honda Civic has lower ownership costs than a Model 3 over a 5 year period anyway, so the lower operating costs of an EV never quite offset the purchase price.

Well, Tesla has been promising robotaxis way more recently than 2018 so promises of driverless operation are not something in the past.
The way I look at the economics is that Uber costs about $1.50 a mile here. I think the total cost of ownership of a Model 3 is way lower than that. Besides, the FSD hardware is dirt cheap so Tesla could license it to a manufacturer capable of making a vehicle with a lower cost to operate.

However, I don't agree that Level 2 features are simply "entertainment". They have value in making the driving experience safer, incrementally easier, and more convenient. Cruise control is the same value proposition on a much smaller scale. It makes driving less effort without compromising safety. I think there's a huge area between cruise control and Level 4, and that's where FSD fits in. For me, FSD was worth the $7k I paid for it even without autosteer on city streets, and any additional Level 2+ features I get are a bonus. I understand not every one calculates value in the same way though.

Some L2 features are useful. I just don't think that city FSD is useful. Even if it gets 100x better than what it is now it will still require the same level of vigilance. I don't want to drive with my hands hovering over the wheel ready to take over in a fraction of a second, that sounds exhausting. I would use it occasionally just because I'm an engineer and it does seem fascinating.
Obviously there's also the concern that people won't maintain the proper level of vigilance when using the system. I think Tesla should be required to do detailed reporting to make sure the system isn't actually making us less safe.

 

[JulienW]

Is this the first video with City Streets driving in the rain?

Nice video, but I wouldn't call that city street driving. More like a town with some outskirts driving.

 

[verygreen]

Hey @verygreen does that mean that Tesla has figured out how to HD map all these states and cities and roads faster than the "industry leader in autonomous driving" Waymo?

you are comparing incomparable things.

In no way Tesla FSD is an AV. did you read the release notes disclosures?

 

[verygreen]

Hey @verygreen Tesla mapped the entire state of NC in "centimeter level accuracy" since you posted your tweet revealing the "truth" about Tesla using HD maps!

"entire state"? How did you determine that?

 

[AlanSubie4Life]

I don't agree that Level 2 features are simply "entertainment". They have value in making the driving experience safer, incrementally easier, and more convenient.

I agree with that, as long as most of the features operate in the background, especially in the more complex environments. There’s currently no safer substitute for an alert driver performing nearly 100% of the driving task. But if you multiply that by a background capability that can avoid accidents, you start seeing some real safety improvements!

You just don’t want the accident rate, which you’re reducing by the FSD accident reduction factor, to go up by an order of magnitude (or whatever)! Especially if the baseline rate goes up in areas where the accident rate was already very low...

That’s why I believe never using FSD is the best way to maximize the expectation value of my return on my investment.

 

[mikes_fsd]

"entire state"? How did you determine that?

The videos I saw, he was not restricted by any addresses.
In fact he did not have to have an address at all.

Do you have any evidence to show that the releases are geographically limited (geo-fenced)?

 

[mikes_fsd]

Waymo is doing autonomous driving in 25 metro areas in 7 States across the US! I

No, Waymo is testing their prototypes and mapping the roads in 25 metro areas.
That is not autonomous driving.
There is only one place they are doing autonomous driving without driver supervision, and that is 50 sq miles in the desert.

 

[verygreen]

The videos I saw, he was not restricted by any addresses.
In fact he did not have to have an address at all.

Do you have any evidence to show that the releases are geographically limited (geo-fenced)?

I never said Tesla testing in the current "FSD beta" were geofenced, this is something you came up with and want me to prove it for you it appears?

 

[mikes_fsd]

I never said Tesla testing in the current "FSD beta" were geofenced, this is something you came up with and want me to prove it for you it appears?

I gotta say, you really love your audience to be obedient and only ask shallow questions, but when someone calls you out on your bs all you have is diversion to a different subject.

What evidence do you have that Tesla is using HD maps - "centimeter level accuracy" like you said in your tweet??

 

[mhan00]

Another vid from Brandon in Sacramento. He will for sure be my baseline from now on about the safety of the system. The new update doesn’t seem to have much for him and the car is still pretty damn dangerous. If/when the beta starts working well for him, I’ll start feeling good about the beta being released to the general public. As is, it can’t be trusted at all.

 

[mikes_fsd]

Another vid from Brandon in Sacramento. He will for sure be my baseline from now on about the safety of the system. The new update doesn’t seem to have much for him and the car is still pretty damn dangerous. If/when the beta starts working well for him, I’ll start feeling good about the beta being released to the general public. As is, it can’t be trusted at all.

I will have to agree that his have been by far the "roughest" drives I've seen.
But I love how he posts the videos without any gimmicks or trying to hide anything (or at least it comes across authentic).

And I second your feelings his experience being the baseline for me as well!