-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Autonomous Car Progress
- Thread starter stopcrazypp
- Start date
-
- Tags
- Autonomous Vehicles
boonedocks
MS LR Blk/Blk 19”
We have a model X with the exact same PPF and CeramicPro and it has never once had the same happen \i/ Only difference is the S is HW 2.0 and our X is 2.5
AP 2.5 is a Continental Radar and 2.0 is a Bosch radar.
strangecosmos
Non-Member
I didn’t learn too much new from this podcast, but I still found it interesting. It’s about Waymo, Cruise, and Tesla’s plans to launch autonomous ride-hailing services, and how they are each taking different approaches.
Tasha Keeney (the ARK Invest analyst being interviewed) made an interesting point that Waymo’s approach is more incremental than Cruise’s, since the Phoenix suburbs are an easier place to launch than downtown San Francisco. Starting with an easier problem and working your way up can be an effective way to make good progress over time. (Timothy B. Lee points out that Voyage’s approach might be
even more incremental than Waymo’s, since they are testing in a gated retirement community at low driving speeds.)
Tasha Keeney also points out (or maybe I made that connection in my head; I can’t remember now) that Tesla’s approach is even more incremental than Waymo’s, since Tesla is releasing driver assistance features bit by bit. Maybe small but rapid iterations is the way to go. If so, Tesla is doing it right.
Understanding the Autonomous Vehicle Landscape and the Opportunity it Creates for Investors - FYI - For Your Innovation
Tasha Keeney (the ARK Invest analyst being interviewed) made an interesting point that Waymo’s approach is more incremental than Cruise’s, since the Phoenix suburbs are an easier place to launch than downtown San Francisco. Starting with an easier problem and working your way up can be an effective way to make good progress over time. (Timothy B. Lee points out that Voyage’s approach might be
even more incremental than Waymo’s, since they are testing in a gated retirement community at low driving speeds.)
Tasha Keeney also points out (or maybe I made that connection in my head; I can’t remember now) that Tesla’s approach is even more incremental than Waymo’s, since Tesla is releasing driver assistance features bit by bit. Maybe small but rapid iterations is the way to go. If so, Tesla is doing it right.
Understanding the Autonomous Vehicle Landscape and the Opportunity it Creates for Investors - FYI - For Your Innovation
Many words come to mind when thinking back on the Autopilot 2 progress and its iterations but rapid isn’t one of them.
Same could be said about the first 20.000 years of human technological progress, so I'm happy to live on the steep part of the exponential curve. We would only be able to recognize if Teslas progress was actually an exponential curve once it really takes off.
(Here's to hoping...)
True. Belief in Tesla’s autonomous progress turning the Autopilot 2 cars from 2016 into actual autonomous vehicles requires belief in a lot of theoreticals at this stage. There is not much factual evidence in the public to suggest they can ”make it” faster than the competition so it seems a massive leap of faith.
Unfortunately unlike with Tesla’s BEV roadmap which was ahead of its time in a good way I am not seeing the same level of thesis being floated that would so surely suggest they got this figured out better than the rest of the world. The thesis I’ve seen rely a bit too much on the belief that just because Elon got a few other things right which he certainly did, that Tesla would get this one right too. That remains a leap given the execution we are seeing.
I get the idea that with Tesla’s hardware deployment once they get the software right they can just deploy it out there. However even that equation changes when we consider the possibility that current hardware suite might not be enough eg for regulators. Also since we don’t know how behind Tesla is from the current lead it is hard to say where the rest of the autonomous market is by the time Tesla ”hockeysticks” on this...
Put it this way: Model 3 and Gigafactory hockeysticking have been plausible because the Model S and battery execution beforehand was solid. Roadster before Model S and so forth. It is easy to see how that advatange turns into reality despite expected hardships. With Autopilot 2 there is no such visibility to the driving software being a solid smaller volume version of the eventual big thing though.
In this case there is such visibility into Waymo. They seem to have the solid smaller volume version of the eventual big thing. Waymo One is their Roadster. The eventual Chryler and Jaguar ”tens of thousands” partnerships are their Model S... it is not hard to see how Waymo could scale what they are doing now into to the big leagues. Waymo has a competitive team already now they just need to take it out there. Sure their way of taking it out is a slower one than with Tesla but they have the team.
With Tesla they have laid the ground work for joining the big leagues further than Waymo but we still have no idea if and when will they have a competitive team — we have not seen any of it.
Last edited:
Elon has a track record of successfully disrupting established industries by, for the most part, simply boldly doing the things the established companies were too complacent or risk-averse to do, or which would hurt their core business in the short term, but which were right in front of their faces. He did this with Paypal, he did it with BEVs, and he's doing it in aerospace with SpaceX. In all of these cases, he disrupted by applying mostly existing but risky technology that was just beyond the frontier of what would be considered proven and safe technology.
In none of these cases was much progress in basic science or fundamentally new invention required. All of the technologies he used to disrupt (particularly at Paypal) were basically already at a level where an engineer skilled in the field could lay out the steps required to bring it to production, and probably be pretty accurate at a high level.
L4/L5 cars on public roads, on the other hand, are a completely different story -- especially to do it without lidar. This is not a question of just hiring the right engineers and pushing them hard to apply what they already know. This is a fundamentally new domain, not just on the other side of the technology frontier, but deep, deep into it -- particularly back in 2014 when AP1 was first released and Elon started talking about all the amazing things it would eventually do. The frontier is a little closer now but frankly it's still quite difficult to see how much farther back we have to push the frontier before this will work. This is not like Elon's success stories, and frankly the man has no idea what he's talking about (or else he's just an outright fraud). AP2 will be remembered as Elon's greatest failure, even while all his other disruptive projects ride that hockey stick. (And, eventually, AP4 will work, at which point Tesla will be late to the game and not disrupting anything.)
The thing about things deep beyond the technology frontier is that you can't usually say how long it will take to reach them. There are too many unknown unknowns. A person with appropriate vision and imagination can certainly say that a thing is possible, that it does exist beyond the frontier -- and I think any reasonable person needs to admit that camera-only L5 is in fact somewhere beyond that frontier -- but if it's not almost within reach already, it's just folly to commit to any kind of timeline. And when you combine that folly with accepting payment for something that does not exist, you have crossed an important line.
People with TSLA stock will always spin everything to conclude "Tesla is right".
The Tesla proponents have laways tried to push this idea of exponential performance increase. But we have seen no indication of that. The same with the benefit of tesla's autonomous data. No one has actually proven or shown any evidence of this benefit.
Ark Invest is literally a scam whose main and only goal is to cheerlead and fawn over Elon.
Its beyond laughable when your lead autonomous cars research investor doesn't even know how many lidar sensors Waymo has.
Saying it has 3 LIDARS when it clearly has 5. Lets also ignore the camera discrepancy. Saying it has 8 cameras when it has 9 camera modules which contain about 18 cameras.
A simple google search and a first grade deductive reasoning would give you the answer. But this is Ark Invest.
"Tesla is three years ahead on autonomous data, three years ahead on autonomous hardware, no one can catch them at this point" - Tasha Keeney (Ark Invest)
This is as clueless as you can get.
Its amazing how little they actually know. Yet they continue to pump TSLA.
Its beyond laughable when your lead autonomous cars research investor doesn't even know how many lidar sensors Waymo has.
Saying it has 3 LIDARS when it clearly has 5. Lets also ignore the camera discrepancy. Saying it has 8 cameras when it has 9 camera modules which contain about 18 cameras.
A simple google search and a first grade deductive reasoning would give you the answer. But this is Ark Invest.
"Tesla is three years ahead on autonomous data, three years ahead on autonomous hardware, no one can catch them at this point" - Tasha Keeney (Ark Invest)
This is as clueless as you can get.
Its amazing how little they actually know. Yet they continue to pump TSLA.
lunitiks
Cool James & Black Teacher
Dey even forgot Tesla's 12 sonars.
Establishing a protective cocoon around the car
Establishing a protective cocoon around the car
Last edited:
This is ONE of the most ignorant article i have ever read but par for course for the Tesla apologists.
Tesla’s Autonomous Opportunity is Severely Underappreciated
And this is what they call "Research".
Tesla’s Autonomous Opportunity is Severely Underappreciated
And this is what they call "Research".
Last edited:
Tam
Well-Known Member
Tam
Well-Known Member
Tam
Well-Known Member
CES 2019: nVidia Drive AutoPilot is huge sitting in the trunk!
Tam
Well-Known Member
CES 2019: ZF
I didn't know ZF makes Autonomous Vehicle components too. So is it nVidia using ZF or ZF using nVidia?
I didn't know ZF makes Autonomous Vehicle components too. So is it nVidia using ZF or ZF using nVidia?
Last edited:
I'm surprised nobody posted these
UPDATE: Disengagement Reports 2018 – Final Results
And these from CA DMV: Autonomous Vehicle Disengagement Reports 2018.
There's also this news: Apple Confirms 190 Layoffs From Self-Driving Car Project and rumors of Ford, VW continue talking; doubt raised that agreement is near on self-driving car pact and Waymo reportedly in talks with Renault-Nissan Alliance on self-driving partnership.
UPDATE: Disengagement Reports 2018 – Final Results
And these from CA DMV: Autonomous Vehicle Disengagement Reports 2018.
There's also this news: Apple Confirms 190 Layoffs From Self-Driving Car Project and rumors of Ford, VW continue talking; doubt raised that agreement is near on self-driving car pact and Waymo reportedly in talks with Renault-Nissan Alliance on self-driving partnership.
strangecosmos
Non-Member
Reporters like Amir Efrati have revealed that the disengagements data is almost meaningless as a metric of autonomous driving capability or progress. Each company reports their disengagements differently but the numbers get compared side by side anyway. Companies can choose not to report the vast majority of disengagements.
Similar threads
- Article
