FSD Beta Videos (and questions for FSD Beta drivers)

[Daniel in SD]

180 Degree forced turn - 2020.40.8.13 - 9 Nov 2020 - 1:19 - James Locke

I wish he had waited to see what would have happened the first time, for science. He used a lower speed for the successful attempts.
Obviously you'd want to benchmark the stopping distance before really pushing it.

 

[n.one.one]

Handling several different scenarios - 1:53 - Tesla Owners Silicon Valley
v.13

 

[n.one.one]

Day 2 180 Degree forced turn - 2020.40.8.13 - 10 Nov 2020 - 6:57 - James Locke

 

[n.one.one]

4.5 minute continuous drive - 4:35 - Tesla Owners Silicon Valley
v.13

 

[n.one.one]

Climbs Up to Coit Tower - 2:02 - Whole Mars Catalog
Version unknown

 

[n.one.one]

Heads to Painted Ladies (houses) in San Francisco - 2:18 - Whole Mars Catalog
Version unknown

 

[mhan00]

No longer watching Mars’ videos because he never holds the damn wheel. Tugging on the wheel once every few minutes is not responsibly using the beta system.

 

[turnem]

No longer watching Mars’ videos because he never holds the damn wheel. Tugging on the wheel once every few minutes is not responsibly using the beta system.

Your boycott will SURELY teach him a lesson!!

 

[mhan00]

Your boycott will SURELY teach him a lesson!!

I don’t give a crap if it affects him. I just have no interest in watching someone use the system irresponsibly.

 

[n.one.one]

Static object detection - 10:15 - AIDRIVR
v.13

 

[n.one.one]

Static object detection - 10:15 - AIDRIVR
v.13

This is a well organized and interesting series of tests. It has good video quality with excellent editing.

 

[mhan00]

Static object detection - 10:15 - AIDRIVR
v.13

damn impressive.

 

[turnem]

I don’t give a crap if it affects him. I just have no interest in watching someone use the system irresponsibly.

Understand.

I personally think he has shown that he's fully ready to take over when necessary. It's not like he's drinking coffee and eating pastries while driving.

I'd personally have my hands on the wheel (similar to how James Locke does it) but I wouldn't go so far as to say that this other guy is being wildly irresponsible.

 

[n.one.one]

Handling mountain roads at night - 3:42 - Tesla Owners Silicon Valley
Version unknown

 

[n.one.one]

Handling mountain roads at night - 3:42 - Tesla Owners Silicon Valley
Version unknown

Very challenging mountain road with no lane markings or curbs and multiple hairpin curves.

 

[uujjj2]

On another topic, not that there was any debate, but that video sure does prove without a doubt that the car uses quite detailed map data to generate the base visualizations. (The steep approaches to the intersections which still show full detail including crosswalks, even when there would be no way for the high-mounted camera to see the intersection, make it very clear.) I guess they blend in the perception onto that base data or something. Presumably if the perception is deemed good enough and it conflicts, then that takes priority?

Knowing the number of lanes doesn't require HD maps. Even my little old Garmin with SD maps knows how many lanes wide each road is. Yes, clearly they are blending map data, but there's no evidence it is LIDAR mapped super-duper resolution.

 

[EVNow]

Your boycott will SURELY teach him a lesson!!

Problem is - nobody is complaining about it in the comments (to the video).

 

[turnem]

Problem is - nobody is complaining about it in the comments (to the video).

So you think that the watchers of the videos have a duty to chastise the driver for not having his hands on the wheel?

 

[n.one.one]

Gravel roads and more - Video #6 - 16:25 - Zeb Hallock
v.13

 

[Mardak]

I personally think he has shown that he's fully ready to take over when necessary.

It seems like his threshold of "necessary" is to avoid an accident. I realize he's testing this case at 1am with nobody nearby, but FSD beta suddenly changed lanes in an intersection then switched back after realizing it followed the wrong lines, and he let the car do its thing:

Unclear if Autopilot team is relying on active disengagements and feedback to notice this issue or if there's triggers to detect patterns of "this sudden maneuver wasn't necessary / turned out to be wrong."