Autonomous Car Progress

[diplomat33]

https://twitter.com/x/status/1677341698572312578

 

[spacecoin]

Looks like people in SF have found a new way to make their voices heard ahead of the CPUC meeting on AVs: crippling the fleet by putting traffic cones on their hood.

Fed Up Citizens Disable San Francisco's Self-Driving Cars With Traffic Cones

An activist group called Safe Street Rebel are promoting the tactic on TikTok.

jalopnik.com

This like saying the people of NYC grew tired of the city planners on 9/11.

This clearly organized sabotage.

Edit: Criticizing the reporting in the article if that’s not obvious.

 

[Dewg]

This is so dumb. And it will not make SF safer. On the contrary, it will only cause more stalls and cause more traffic issues. If they really cared about resolving stalls, they should work with Waymo, not try to pull stupid tiktok challenges.

I'm not convinced they are trying to improve safety. I think they want to create enough chaos that regulators pull the permits and remove AVs from the road completely. They just don't want the technology on the roads. The more chaos, the more likely something will have to be done.

 

[diplomat33]

I'm not convinced they are trying to improve safety. I think they want to create enough chaos that regulators pull the permits and remove AVs from the road completely. They just don't want the technology on the roads. The more chaos, the more likely something will have to be done.

Yes. They have an anti-AV agenda. They simply don't want the technology on public roads and are trying to get it banned. It is despicable IMO. And their methods are deplorable. To purposely create chaos and vandalism to try to get something banned is not the right approach.

 

[Bladerskb]

This is so dumb. And it will not make SF safer. On the contrary, it will only cause more stalls and cause more traffic issues. If they really cared about resolving stalls, they should work with Waymo, not try to pull stupid tiktok challenges.

If this was happening to Tesla, the amount of screeching Tesla fans will be doing and what @stopcrazypp @Knightshade and others will saying would be apocalyptic.

They would be saying Big Oil and big ICE companies are behind it and how they are paying people because Tesla is so good and 10 years ahead. Blah blah blah.

But since it’s not Tesla “it’s just people trying to get their voice heard bro!”

It’s astonishing how they don’t see themselves.

 

[Dewg]

If this was happening to Tesla, the amount of screeching Tesla fans will be doing and what @stopcrazypp @Knightshade and others will saying would be apocalyptic.

They would be saying Big Oil and big ICE companies are behind it and how they are paying people because Tesla is so good and 10 years ahead. Blah blah blah.

But since it’s not Tesla “it’s just people trying to get their voice heard bro!”

It’s astonishing how they don’t see themselves.

Cool story, bro. But we were actually talking about Waymo, the public perception of AVs in SF, and this new grass-roots movement to hinder Waymo cars by placing traffic cones on their hoods, causing more chaos and attempting to have robotaxis removed from the road. I get your frustration with Tesla fans, but you're starting to sound like a politician with a stump speech, always pivoting the question/discussion back to their stump speech, no matter what the topic is.

 

[Bladerskb]

Man you write a ton without ever actually saying anything.... probably for the best given when you DO try to say something it's easily debunked with a link or two

Protip- if you don't want people quoting your own words being self-contradicting, try to maintain a more consistent set of FUD statements.

ME: The narrative that billion of miles of data makes Tesla 10 years ahead is a complete sham, the only help billions of miles of data contributes to is validation and mapping. For one, like 0.001% of data are actuallly sent back to HQ so not billions, two; even 10 seconds of data are Gigabytes of data and there is a small data storage on the car making it impossible to hold let alone for Tesla to store data that massive, three; Even with the campaigns a small minority of a few thousand cars get them out of the millions of Tesla vehicles. Four; Every single product that Tesla have which is claimed to be the only result of having billions of miles of data has been disproven because there are other products just as good or better that were developed with 100 or less development cars (Mobileye, Huawei, etc). Due to all of the above and a dozen more, the idea and narrative that is pushed that Tesla is 10 years ahead due to billions of miles of data is not just illogical but also nonsensical.

KnightShade: ZOMG YOU CONTRADICTED YOURSELF!!!!!

 

[stopcrazypp]

This is so dumb. And it will not make SF safer. On the contrary, it will only cause more stalls and cause more traffic issues. If they really cared about resolving stalls, they should work with Waymo, not try to pull stupid tiktok challenges.

Waymo/Cruise has refused to work to address the concerns of SF, saying they are misinformed and invalid. Instead they go to CPUC for what seems to people like rubber stamping. Waymo/Cruise under CA law is not obligated to get public approval in the cities they are testing, so there is very little local citizens (or even the local government) can do when they oppose the testing.

The idea behind these stunts is to get media and the company's attention, to make it so it's impossible to ignore concerns. It may force Waymo/Cruise to employ safety drivers again, which would address the concerns of the people. Basically people were not up in arms when they had safety drivers that could get the vehicles moving when they stalled, but they removed them. This protest method is pretty easy to defeat by simply having a safety driver in the car.

Currently there are on average about 2-3 reported incidents per day with only a fleet of around 200 vehicles per company and limited operating hours. People are worried what will happen when 24/7 operation is allowed and the fleet size expands.

 

[Dewg]

Waymo/Cruise has refused to work to address the concerns of SF, saying they are misinformed and invalid. Instead they go to CPUC for what seems to people like rubber stamping. Waymo/Cruise under CA law is not obligated to get public approval in the cities they are testing, so there is very little local citizens (or even the local government) can do when they oppose the testing.

The idea behind these stunts is to get media and the company's attention, to make it so it's impossible to ignore concerns. It may force Waymo/Cruise to employ safety drivers again, which would address the concerns of the people. Basically people were not up in arms when they had safety drivers that could get the vehicles moving when they stalled, but they removed them. Currently there are on average about 2-3 reported incidents per day with only a fleet of around 200 vehicles per company and limited operating hours. People are worried what will happen when 24/7 operation is allowed and the fleet size expands.

The operative word here is "stunt". I completely agree with this. However, it's not a static situation. Waymo is getting better over time, and will continue to improve. A little forward thinking would stop many of the complaints, but most people can't see past the end of their nose and only care about whether something is inconveniencing them right then and there.

 

[spacecoin]

Waymo/Cruise has refused to work to address the concerns of SF, saying they are misinformed and invalid.

I don't think it's resonable to put Cruise and Waymo in the same basket. 90% of the issues are with Cruise.

 

[stopcrazypp]

The operative word here is "stunt". I completely agree with this. However, it's not a static situation. Waymo is getting better over time, and will continue to improve. A little forward thinking would stop many of the complaints, but most people can't see past the end of their nose and only care about whether something is inconveniencing them right then and there.

Well Waymo/Cruise has done almost nothing to assure the public they care or that they are making steps to minimize the impact. It's hard to get people to just trust them when the incidents have only increased and there is no indication anything is being done to improve it. That's part of why people have resorted to these means.

I also need to point out that the general public don't think like we do. They don't feel like self driving cars (especially fleet based) necessarily are a desired or inevitable end goal. As such any inconvenience brought on by them is purely unnecessary and there is no value to accepting that inconvenience. The fact of the matter is for the foreseeable future they are worthless to the commuters (that commute to/from East or South Bay) that make up a bulk of the traffic and also inconvenience people that ride and support public mass transit.

You really have to get into the mindset of the average person. A similar thing is true of EVs in case you need a reference point, I'm sure a lot of people here are familiar with the challenges of getting people to mentally accept EVs over ICE vehicles.

 

[stopcrazypp]

I don't think it's resonable to put Cruise and Waymo in the same basket. 90% of the issues are with Cruise.

I mention Waymo only because @diplomat33 mentioned them (and not Cruise). Yes Waymo is a ton better than Cruise, but it's not zero, and similarly they haven't really done steps that show there is desire to address stalling.

It's obvious they don't want to go back to having all safety drivers, so that is last resort, but even some smaller gesture like adding the law enforcement override mode SFMTA is asking for would at least be a step. Or maybe show the public regularly updated statistics on stalling and the work they are doing to increase their response team to such stalls and decrease reaction times to stalls. There's a lot of stuff they can do to at least show they care and are making progress, but they instead just choose to ignore.

 

[diplomat33]

Waymo/Cruise has refused to work to address the concerns of SF, saying they are misinformed and invalid. Instead they go to CPUC for what seems to people like rubber stamping. Waymo/Cruise under CA law is not obligated to get public approval in the cities they are testing, so there is very little local citizens (or even the local government) can do when they oppose the testing.

I mention Waymo only because @diplomat33 mentioned them (and not Cruise). Yes Waymo is a ton better than Cruise, but it's not zero, and similarly they haven't really done steps that show there is desire to address stalling.

It's obvious they don't want to go back to having all safety drivers, so that is last resort, but even some smaller gesture like adding the law enforcement override mode SFMTA is asking for would at least be a step. Or maybe show the public regularly updated statistics on stalling and the work they are doing to increase their response team to such stalls and decrease reaction times to stalls. There's a lot of stuff they can do to at least show they care and are making progress, but they instead just choose to ignore.

Well Waymo/Cruise has done almost nothing to assure the public they care or that they are making steps to minimize the impact. It's hard to get people to just trust them when the incidents have only increased and there is no indication anything is being done to improve it. That's part of why people have resorted to these means.

Cruise is not addressing the concerns. They are ignoring SFMTA concerns. But Waymo is not ignoring them. They are working to address these concerns. Waymo works closely with First Responders and Law Enforcement. Waymo has also improved their software to better handle construction zones and road closures which has reduced stalls already. Could Waymo do more? Sure. But Waymo is not ignoring their concerns.

That is why I don't think it is fair to lump the two together. They are different tech. The stalls caused by each are different. Their levels of safety are different. Their levels of cooperation with SF are different. Each company should be looked at separately. And if they need safety drivers put back in part-time or full-time, it should be decided on a company by company basis.

 

[Doggydogworld]

Waymo/Cruise has refused to work to address the concerns of SF, saying they are misinformed and invalid.

You're just making crap up. The SF protest letter filed this January says:

The San Francisco Municipal Transportation Agency (SFMTA) and the San Francisco
County Transportation Authority (SFCTA), and the Mayor’s Office on Disability (MOD)
(collectively San Francisco) do not oppose the Commission authorizing Waymo to deploy
commercial service in San Francisco. For several years, Waymo has actively sought City input
about its AV testing and deployment, has demonstrated intent to address several city concerns,
and appears to have apparently invested significant resources in doing so.

Of course they still wanted to delay, restrict and hobble Waymo's ability to operate and collect fares, plus force Waymo to provide all manner of proprietary data. Then in the May letter to CPUC they distorted and lied about some of the data Waymo does provide to promote a false narrative about safety. CPUC rightly called out those lies and distortions.

Getting back to the coneheads, they're a sub-cult of car haters. They don't just want to stop Waymo, they want to ban all human-driven cars, too. Including Teslas. So be careful who you hump into bed with.

 

[Knightshade]

ME: The narrative that billion of miles of data makes Tesla 10 years ahead is a complete sham

You're gonna hurt your back moving the goalposts that fast my dude.

You've made 3 entirely different claims about what the "sham" is now as you keep getting called out on not having your story straight.

This time you made up a "narrative" nobody appears to have even said to rail against. The kids call that a strawman.

For one, like 0.001% of data are actuallly sent back to HQ so not billions

You're conflating "percent of data" with "miles driven" for...some reason?

, two; even 10 seconds of data are Gigabytes of data

This is objectively false.

You're getting worse and worse at making it clear you've no idea WTF you're talking about on a technical front.

https://twitter.com/x/status/1096322821096124416

Green discussing the AP trip logs and the fixed disengagement reports of autopilot trip logs which they collect billions of miles of data from (routes, speeds, how well AP recognized the road, how/when you disengaged, etc). This is all text/numeric data, so they're quite small. The disengagement ones are under 1kb for example.

Since you clearly need help on the tech stuff- that's smaller than "gigs"

Also this data is all sent over cell, so local storage is no issue at all since the car doesn't need to wait to upload them till you're on wifi.

Oh, and they DO sometimes capture video of course... but even then your math is hilariously wrong

Each camera captures at 30 megs of data per minute

That means 10 seconds of video, per camera, is only 5 MB. Times 8 cameras that's 40 MB. Even if the interior cam is included that's 45MB for 10 seconds of video.

Spoiler: 45MB is much smaller than "Gigabyes of data"

; Even with the campaigns a small minority of a few thousand cars get them out of the millions of Tesla vehicles

This is also outright false.

https://twitter.com/x/status/1096322827177914368

almost all cars get these "campaigns", but not every car gets all of them

It'd be great if you could stop posting about things you clearly do not understand at all.

 

[stopcrazypp]

You're just making crap up. The SF protest letter filed this January says:

The San Francisco Municipal Transportation Agency (SFMTA) and the San Francisco
County Transportation Authority (SFCTA), and the Mayor’s Office on Disability (MOD)
(collectively San Francisco) do not oppose the Commission authorizing Waymo to deploy
commercial service in San Francisco. For several years, Waymo has actively sought City input
about its AV testing and deployment, has demonstrated intent to address several city concerns,
and appears to have apparently invested significant resources in doing so.

Of course they still wanted to delay, restrict and hobble Waymo's ability to operate and collect fares, plus force Waymo to provide all manner of proprietary data. Then in the May letter to CPUC they distorted and lied about some of the data Waymo does provide to promote a false narrative about safety. CPUC rightly called out those lies and distortions.

Getting back to the coneheads, they're a sub-cult of car haters. They don't just want to stop Waymo, they want to ban all human-driven cars, too. Including Teslas. So be careful who you hump into bed with.

Putting aside the letter does not mention Cruise (which was doing considerably worse at the time), it was a letter written before Waymo driverless expanded and Waymo had way less driverless miles than Cruise. Waymo's reported cumulative incidents were in the single digits (it has now reached triple digits). SF commendations are in regards to safety driver based operation (as below SF contended in the same letter, testing without safety driver have not been proven). As I mentioned, there is no objection by the city to safety driver testing, as that has been proven to be safe/effective, I'm talking solely about the dispute over driverless testing, which both companies have been far less cooperative on.

Here's the Waymo response to SF's concerns about driverless operation:
"San Francisco asses that Waymo’s request for authorization is “unreasonable in light of the fo lowing circumstances”: (1) incrementalism; (2) transparency; (3) insucient driverless testing; (4) inadequate repoing and monitoring. None of these issues serve as a valid basis 5 for protesting Waymo’s Advice Leer."

"San Francisco acknowledges that “Waymo has tested both automated driving and passenger service extensively” but objects that “as of August 31, 2022, the latest date for which information is available to the public, Waymo had not tested any passenger service in San Francisco in vehicles that have no safety driver.” 22 San Francisco contends that “[t]his increases concern about the potential for increased AV street obstructions.”23 San Francisco’s concern is unfounded. As noted above, the Deployment Decision does not mandate that an AV-TCP rst conduct a passenger service pilot before being authorized to charge for trips provided to the public- whether in a drivered or driverless conguration. What the Deployment Decision does require is 30 days of operations folowing receipt of the foundational DMV permit. Waymo’s Advice Leer includes this required aestation. The Deployment Decision should not be read to include any additional pre-deployment application conditions of CPUC pilot permit use."

https://www.cpuc.ca.gov/-/media/cpuc-website/files/uploadedfiles/cpucwebsite/content/licensing/autovehicle/reply-of-waymo-llc--waymo-advice-letter-0001-tier-3-january-30-2023--final-pdfa.pdf

The letter in general says SFMTA's objections to driverless operation is invalid and request for more data is also invalid.

Here's the actual recent one that has SF's response based on updated data of SF had to gather themselves (since Waymo refused to supply as per above):
"As of January 25, 2023, when San Francisco filed its protest of the Waymo Advice Letter, the record of incidents reported to the City by members of the public, city employees, media and social media about Waymo operations was quite limited; the large majority of Reported Incidents involved Cruise vehicles. Unfortunately, that picture has changed. In 2023, monthly Reported Incidents involving Waymo driverless operations have increased six-fold."

The CPEC part of CPUC also shares SFMTA's concerns:
"The Draft Waymo Resolution shows that the Consumer Protection and Enforcement Division (CPED) shares some of San Francisco’s concerns about impacts of driverless AV operations on first responder safety and performance, transit safety and performance, road safety and travel delay, the effects of these driverless readiness performance problems as driverless operations scale up, and the failure of existing reporting requirements to adequately assess and address these concerns."

"The Draft Waymo Resolution proposes approval of the Waymo Advice Letter based on the fact that Waymo has submitted a Passenger Safety Plan that is complete and seems reasonable to the CPUC under the existing Deployment Decision—a decision that the Commission adopted long before there were any driverless AV operations on San Francisco streets and before the driverless readiness problems the City has documented were apparent.5 Since that time, Waymo driverless AVs have committed numerous violations that would preclude any teenager from getting a California Driver’s License.6 "

https://sfstandard.com/wp-content/uploads/2023/06/SF-Comments-on-Waymo.pdf

This haven't gotten into discussing Cruise yet, which is on a whole other level.

 

[GSP]

George Hotz answers the question: How close are we to self-driving? Seven minute video with Lex Fridman.

TLDW: George’s thinking is similar to Elon’s. They both have figured out how to do it, and just need a year or two to get more data and compute power to execute their plans.

GSP

 

[Bladerskb]

For anyone who are actually interested in facts, truth and not complete nonsense.

The existence of Mobileye, Huawei, Xpeng literally disproves the "billion of miles data" none-sense.
No single Tesla fan in history will ever address this. EVER.

Secondly Tesla collects RAW images not post processed reduced images.

1.18 MB per image * 8 cameras = 9.44 MB per 8-camera. If tesla were to upload 10 seconds of all cameras at 30fps it would be 2.8 GB.

Years ago tesla fans were averaging 100-300 mb of uploads per day. With the new 5 MB camera, the data requirements are 5x larger.

How much data does your HW2+ Tesla upload?

Ask people with wifi monitoring to contribute their upload data? I’m trying to get a rough ballpark estimate of how much data (in MB or GB) Tesla is collecting from each HW2+ car (HW2 or HW2.5). If you monitor this on your car, please let us know here. If we can get a rough monthly average...

teslamotorsclub.com

1. 2.8 GB for 10 seconds of 8 cameras video at 30fps.
2. With new 5MB camera, the data requirements are 5x larger.
2. 10-20 campaign triggers per week which only select cars received some.
3. Triggers lasted 24 hours, data deleted when they expire.
4. Wifi needed to upload
5. Tesla owners upload traffic shows avg of 100-200 MB/day
6. Disengagement triggers that lead to upload are rare

All of the above are 100% FACT!

What Tesla used to do was take 10 seconds video from one camera (forward camera) and that comes out to 300-500mb just by it-self.

https://twitter.com/x/status/1321243695782502401

https://twitter.com/x/status/1134604857858580480

 

[stopcrazypp]

For anyone who are actually interested in facts, truth and not complete nonsense.

The existence of Mobileye, Huawei, Xpeng literally disproves the "billion of miles data" none-sense.
No single Tesla fan in history will ever address this. EVER.

Secondly Tesla collects RAW images not post processed reduced images.

1.18 MB per image * 8 cameras = 9.44 MB per 8-camera. If tesla were to upload 10 seconds of all cameras at 30fps it would be 2.8 GB.

Years ago tesla fans were averaging 100-300 mb of uploads per day. With the new 5 MB camera, the data requirements are 5x larger.

How much data does your HW2+ Tesla upload?

Ask people with wifi monitoring to contribute their upload data? I’m trying to get a rough ballpark estimate of how much data (in MB or GB) Tesla is collecting from each HW2+ car (HW2 or HW2.5). If you monitor this on your car, please let us know here. If we can get a rough monthly average...

teslamotorsclub.com

1. 2.8 GB for 10 seconds of 8 cameras video at 30fps.
2. With new 5MB camera, the data requirements are 5x larger.
2. 10-20 campaign triggers per week which only select cars received some.
3. Triggers lasted 24 hours, data deleted when they expire.
4. Wifi needed to upload
5. Tesla owners upload traffic shows avg of 100-200 MB/day
6. Disengagement triggers that lead to upload are rare

All of the above are 100% FACT!

What Tesla used to do was take 10 seconds video from one camera (forward camera) and that comes out to 300-500mb just by it-self.

https://twitter.com/x/status/1321243695782502401

https://twitter.com/x/status/1134604857858580480

Wow you are clearly moving goal posts and somehow claiming you are still right, when you made an incorrect statement and people are giving you thumbs up for it? @GSP @DanCar @diplomat33, I question what is the logic you guys see?

From your own post:

two; even 10 seconds of data are Gigabytes of data and there is a small data storage on the car making it impossible to hold let alone for Tesla to store data that massive,

From this post, the facts by green:

there's a special trigger that listens for it. In general they take everything-everything including 10 seconds of video and the decision logs and all that stuff.

so how many gigs does it upload daily? keep in mind one full snapshot is 300+Mb.

Doesn't that directly disprove your own claim that 10 seconds of data is Gigabytes of data? And note no-one is claiming Tesla captures every byte of data available on the car in a snapshot, that's clearly a strawman, given lots of data would be irrelevant.

 

[stopcrazypp]

Secondly Tesla collects RAW images not post processed reduced images.

1.18 MB per image * 8 cameras = 9.44 MB per 8-camera. If tesla were to upload 10 seconds of all cameras at 30fps it would be 2.8 GB.

Years ago tesla fans were averaging 100-300 mb of uploads per day. With the new 5 MB camera, the data requirements are 5x larger.

How much data does your HW2+ Tesla upload?

Ask people with wifi monitoring to contribute their upload data? I’m trying to get a rough ballpark estimate of how much data (in MB or GB) Tesla is collecting from each HW2+ car (HW2 or HW2.5). If you monitor this on your car, please let us know here. If we can get a rough monthly average...

teslamotorsclub.com

1. 2.8 GB for 10 seconds of 8 cameras video at 30fps.
2. With new 5MB camera, the data requirements are 5x larger.
2. 10-20 campaign triggers per week which only select cars received some.
3. Triggers lasted 24 hours, data deleted when they expire.
4. Wifi needed to upload
5. Tesla owners upload traffic shows avg of 100-200 MB/day
6. Disengagement triggers that lead to upload are rare

All of the above are 100% FACT!

What Tesla used to do was take 10 seconds video from one camera (forward camera) and that comes out to 300-500mb just by it-self.

PS what you posted above is also incorrect. Tesla's snapshots do not include captured RAW at 30fps, they also don't only capture from one camera. This is what the 10 second 300+ish MB snapshot captures (along with other non-video/image data not mentioned but included in the data), digging up some older posts when AP2 was discussed.

As I linked, we already saw footage and snapshots from AP2 gathered by @verygreen. 10 seconds .h265 at 30fps from main and narrow. And 10 seconds RAW at 1 fps from all 7 B&W cameras.
AP2.0 Cameras: Capabilities and Limitations?

green actually gave you some examples back then, but you may not have opened it and analyzed them (the rest of us did in the thread):

HW2.5 capabilities

That's literally 5 cameras rendered useless. @verygreen can confirm this The night footage is here, feel free to process it (there's also a dashcam video for comparison): https://app.box.com/s/jah6ovf6y5eoq9faejeylt7fm2j2od4i

teslamotorsclub.com

The data link is still valid, so anyone curious can still download them and look to verify what's actually in a snapshot.