mikes_fsd
Banned
yep, like these "simple merges on to the highway"?
https://twitter.com/nitguptaa/status/990683818825736192
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Waymo
- Thread starter Daniel in SD
- Start date
https://twitter.com/nitguptaa/status/990683818825736192
mikes_fsd
Banned
So, you are comparing the software running Autopilot in 2019 to the NN's running on FSD Beta?
If anything, what we've seen with FSD Beta only reinforces the Tesla approach of vision first.
Last edited:
DanCar
Active Member
From 2.5 years ago. I'm willing to bet that has been fixed.
mikes_fsd
Banned
Ah, so, we assume the slow moving company has fixed the software, yet the fact that we have seen Tesla pushout out 6 or 7 releases of FSD Beta showing "2 steps forward one step back" will not solve the outstanding issues?
diplomat33
Average guy who loves autonomous vehicles
Check the time stamp on the video, that's from 2 years ago. Waymo has improved a lot since then.
Of course, I see the improvements that Tesla is making with FSD. That's great. But you want me to just praise Tesla all day for their progress and ignore any progress from Waymo or others? I am not allowed to praise Waymo for their progress and for the FSD that they have?
Tesla is putting advanced driver assist features directly into user hands that nobody else has. That's awesome. I can't wait to get FSD Beta on my Model 3 and try it out. I hope it works great. I wish Tesla the best. I hope they achieve true FSD.
The fact is that Waymo has true FSD. Is it perfect? No of course not. It has limitations. But it is true FSD. I am not going to shy away from praising Waymo for what they have accomplished so far.
I wish Tesla the best. I hope they do succeed in achieving true FSD. But I am not going to ignore the progress of other companies. Waymo, Zoox, Mobileye, Cruise are also making awesome progress on FSD.
I wish all AV companies the best. I want to see as many safe and reliable AVs on public roads as possible. If it is a Waymo car, a Cruise car, a Zoox car, a Mobileye car or a Tesla car, I will applaud them all for their progress.
mikes_fsd
Banned
Last I saw, they still do not allow their cars to even go on a highway without safety driver in the dessert!
Have they changed that now?
EVNow
Well-Known Member
Its actually very arrogant to dismiss others observations (esp. when you didn't even apparently see the conditions I was talking about - night + rain etc).
mikes_fsd
Banned
Do you want an award? or a star sticker maybe?
Or do you think that should make me agree with your POV somehow?
EVNow
Well-Known Member
It has been clear (as day light in AZ) for a long time that Waymo is better in a lot of ways than Tesla. Duh ... they have immense resources compared to Tesla (and they started long before Tesla).
The question has always been their ability to expand beyond ideal conditions in that one city suburb. I'll be impressed when they start expanding fast in multiple cities.
At the moment it looks like Tesla is expanding their feature set faster than Waymo is expanding their geofence.
Last edited:
diplomat33
Average guy who loves autonomous vehicles
Why the snarkiness? No. I am not asking you change or agree with me. I am just letting you know where I am coming from.
You are the one that keeps starting these debates with me.
Yes, that is a key question.
Tesla and Waymo are approaching FSD from two different angles. Tesla is going "wide", expanding first to everywhere while also improving the driver assist features until they become reliable autonomous driving. Waymo is going "tall", focusing on getting to super reliable autonomous driving where they never need a safety driver first and then expanding everywhere later.
And I have said before that if Tesla does improve FSD on the current fleet to where it is driverless everywhere, then Tesla would absolutely "win" and crush everybody, including Waymo. Obviously, Tesla is not there yet.
I think I have also said that Tesla and Waymo are different businesses. Tesla is aiming for FSD on consumer cars. Waymo is aiming for driverless ride-hailing. Both are equally valid IMO.
DanCar
Active Member
There was some news recently that Waymo wanted to expand in California but was denied, because of billing issue. The California government wants to spend a couple of years figuring out what billing should be before giving a permit. Super ridiculous.
mikes_fsd
Banned
powertoold
Active Member
My biggest concern with Waymo's approach (based on current video examples) is that LIDAR is used for localization based on environmental reference points, but when those reference points are obstructed or distorted (by mirror-like reflections), the car seems to be less certain of its / other cars' positions. I've seen this various times during drives where the steering wheel jitters around when:
The car is in between two large cars / trucks
The car is near large reflective surfaces
The car is in narrow roads with parked cars on both sides
The car is in between two large cars / trucks
The car is near large reflective surfaces
The car is in narrow roads with parked cars on both sides
diplomat33
Average guy who loves autonomous vehicles
Yeah, the CA government created this ridiculous 2 step process for getting permission to launch a driverless robotaxi service. First, you have to get a permit from the CA DMV that certifies that you have reliable driverless FSD and then you have to get a permit from the Public Utilities Commission to figure out the pricing etc... And you can't do both simultaneously. You have to wait until you complete the permit process with the CA DMV before you can even start applying for the second permit from the Public Utilities Commission which just drags the process far longer than it needs to be. The process can take months or years. It's why AV companies are leaving CA and deploying their robotaxis in other States like AZ or NV.
I'm sorry, I get a little triggered when people say a driver assist system is better than a human. I feel like that's a dangerous assumption when there are so many things on the road that may need to be perceived and the system is only really optimized for a subset of them.
I've had autopilot shut off in the rain as well. I think what people keep forgetting that Waymo is trying to validate a system that can operate safely without human oversight. They're using extremely conservative criteria for having the system transition to a minimal risk condition. You can't compare what's safe with no driver vs. what's safe with a safety driver.
mikes_fsd
Banned
I am sorry, I get triggered when people keep claiming that human drivers are so great when technology -- from ABS, AEB and ADAS -- has shown a consistent improvement in road safety over humans time and time again!
FSD is putting technology to use where technology has shown its superiority time and time again!
I was not commenting on technology assisting humans. We are in agreement on that. Technology can absolutely make humans better drivers.
This is the Waymo thread so I assumed we were talking about technology replacing humans drivers.
mikes_fsd
Banned
The logical conclusion of seeing technology doing better in portions of the driving tasks is to replace all the driving tasks with technology (aka autonomous driving)!
Of course. That's why about $100 billion has been invested in autonomous driving tech.
Unfortunately it turns out to be an extremely difficult problem to solve.
Similar threads
- Article
