Living downtown you see a lot of this, people just in the way.
I have to admit, #TeslaFSD did a pretty good job of navigating around these obstacles.
The parked truck and the parked car, with a third vehicle moving around the other two. This doesn't prove it's safe, but it is impressive.
Another recent instance where #TeslaFSD just can't pick a lane. After spending time getting out of the 7th St turn lane, it decides it wants to exit anyway. Then once off the exit, it can't seem to decide which direction it should go.
Spoiler Alert: Either way would get me where I wanted to go.
I know it’s me, but I think things like this are very funny. There are very few cars in the parking lot, the spaces are wide open, and it still takes #TeslaFSD three corrections to park.
When it started the third attempt, I had to end it.
Why is it always one steps forward and two steps back?
I appreciate that #TeslaFSD actively watches out for pedestrians. Even if they are not sure of what I am going to do.
It's hard to tell if the system's hesitation is because of the pedestrians not sure if I am letting them cross or getting out of the way of the road and vehicles behind me.
I like the extra help #TeslaFSD provides while navigating the bike lanes in downtown. The bike lanes are intertwined with car lanes so the additional "eyes" are useful.
It's also interesting that, just like human drivers, crossing the double yellow line to go around a stopped vehicle is common.
It's amazing #TeslaFSD knew to line up this early for the turn.
The light and the turn are on the other side of the overpass up ahead.
It's the law in most states to let pedestrians cross when they are waiting at a crosswalk. Living downtown, it's a common occurrence but I have to admit that #TeslaFSD tends to be less forgetful and tries to be more polite.
Thoughtful video of NVIDIA founder and CEO Jensen Huang showing off their autonomous driving solution.
It's hard for me to see, with all this competition, how Tesla actually pulls this off. #TeslaFSD is good but doesn't rise to the level of seriousness that these guys do.
youtu.be/ZMpDiGoU4Wg?...
As much as I criticize #TeslaFSD, it does do amazing things. Here at night, navigating through relatively complex construction lanes on McDowell in Phoenix, it can determine the turn lane.
In addition, distinguishing vehicles from construction in oncoming traffic for the turn is really impressive.
Coming up to an intersection, it appears #TeslaFSD can hear/see the firetruck ahead at the light as it pulls over short of the truck in front.
Of course, it's not smart enough to know that pulling over here makes no sense since there are cars all around and would block entrance into the driveway.
I've got to give #TeslaFSD credit that it's more polite (and dare I say more lawful) than I would ever be.
In Phoenix, drivers must yield the right-of-way to a pedestrian within a crosswalk (marked or unmarked) on your half of the road, the moment they step into the roadway or appear to.
Missed pedestrians "by that much."
Something #TeslaFSD would never had done, attempted that turn with those pedestrians in the crosswalk (which they had the right of way).
Also surprised FSD didn't react more to the oncoming car clearly over the double yellow line. Speed too slow?
This was a neat trick.
It's hard to tell what this guy is doing or what his intentions were but I think #TeslaFSD reacted in an appropriate way by slowing down slightly and, since there were no other cars, moving over in the lane.
Making a left turn. Map shows it's making a left turn. And yet #TeslaFSD fails to get into the left turn lane. I had to manually move into that lane as it was clear.
This is not a dangerous situation but again, totally inconsistent and you never know what it's going to do wrong. Can't rely on it.
As it contemplates a left merge before the second light ahead, it does take its time to move over. There is plenty of space but #TeslaFSD waits to the last minute, having to cross 2 lanes of traffic at once.
It's not necessarily unsafe, but it does seem aggressive and drivers need to be ready.
What a jerk.
We're almost in downtown Phoenix and this guy's doing 70 mph in heavy traffic.
It's hard to see in the video but #TeslaFSD does react and moves over slightly to let this guy by. Scared the crap out of me because I didn't know what the car was reacting to until he flew by.
I don't like the back way #TeslaFSD uses to Costco, through a dense neighborhood and loading area. But it does avoid a lot of congestion on 44th St. And while there are tons of people walking the parking lot, it does a nice job.
First time it successfully drove the entire way and parked itself.
I love how #TeslaFSD waits for these guys. Even when one backtracks into the street for something he dropped.
Although, it never really stops which is why I kept my foot near the brake the entire time.
Of all the pot holes along this street in Phoenix, #TeslaFSD decides to avoid just this one for some reason. Again, it's the inconsistency that's frustrating and concerning.
This is a great write-up on why less than 20% pay for #TeslaFSD. It's not delivering an experience that provides value. Maybe in the future, but it's not there now and doesn't look like it will be anytime soon.
Hopefully, Rivian won't make the same mistakes.
youtu.be/VH5EF_AtaRw?...
One thing they can't seem to fix is picking the correct lane.
In addition, it misses an opportunity to merge quickly by getting in front of the blue Chevy. I attribute this to an incorrect "visualization" of distance by #TeslaFSD with the white car in front instead of knowing (i.e., with LiDAR).
This is interesting but like a lot about #TeslaFSD, it's unreliable. Sometimes it works and other times it ignores the input. Some of this might be because it determines conditions are unsafe (who knows) but it gives the impression that it's not fully baked.
www.notateslaapp.com/news/3631/ho...
I continue to be impressed with these later versions of #TeslaFSD as it has been consistently very good at navigating around construction. This shows a closed left lane ahead of a double left, turning south from Shea Blvd onto State Route 51 and it finds its way very smoothly.
#TeslaFSD shows itself to be amazing in some situations with pedestrians who are hidden by cars, trees and bushes, who might not be paying attention as much as they should. Glad I am.
Although, I am not wild about how it starts to move before they have cleared the crosswalk. Too much like a human.
Coming out of Costco, you can see #TeslaFSD sees the stopped traffic and makes the decision to wait for vehicles that are wanting to turn.
Hard to tell if it's the black Nissan turning left in front (probably) or the white Toyota wanting to turn right. But it definitely decides to let them turn.
This was a complicated left turn. There is already a car waiting in the turn lane, the truck comes up from behind and also wants to make a left turn. And there is a car coming in the opposite direction.
#TeslaFSD does a good job of navigating through this and even finds an empty space and parks.
It's a little hard to tell in the video, but I almost get rear-ended here (you can see the driver pick something off the floor because he stopped too short).
While I might not have stopped at this yellow, I often wonder if #TeslaFSD would inch forward if it could, to avoid a rear-end collision.
One noteworthy aspect of #TeslaFSD is how reliable it is when watching out for pedestrians, especially in slow moving, urban environments. Partly, because it is ridiculously cautious about coming to a full stop. But also waiting to see what nearby pedestrians, who can be unpredictable, will do.
Another HAWK beacon test. This time there is no lead car to clue #TeslaFSD in on when and where to stop. I was curious as I think this issue is finally resolved.
