Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
Sounds like a job for a monorail.
disneyland-monorail-00.jpg

I hear those things are awfully loud!
 
I wouldn't exactly describe something as a 'trick' that effectively doubles your amount of sensor analysis to perform - and I could also imagine that keeping track of already detected objects for more than two temporal steps would be useful, f.ex. for getting some info on accelerating/braking cars/motorcycles.

They (of course) do this, but not at the neural network level, but at the vehicle control software level, which is regular procedural-deterministic code, not neural networks.

Having the 'previous frame' as input to the neural networks gives the NN the ability to see the more dynamic aspects of the scene: velocity and to a certain extent 3D perspective vision.
 
  • Informative
Reactions: VValleyEV
Not on the highway it doesn't.
Where I live here in the states it get's cold, but not "really" cold... down to single-digit temps (F) is about the worst.

My S can lose 1/3rd range from cold alone. Even with straight highway driving. Add a little snow/slush to push through or some wind and you can lose 40%.
 
Then why did you object to my statement that cutting range in half for winter is not a realistic metric? It was a Texan making a matter-of-fact statement that "The first is trips--particularly trips up north in winter where the 400 mile range declines to a bit over 200." You know that that is not a realistic estimate of winter range loss, even in the coldest-winter places in the continental US.

I fully get why you want to emphasize winter range loss. And I agree! But do you also get why I don't want it overemphasized with unrealistic numbers?
A Texan that lived in Canada for over 25 years. (Admittedly not driving an EV during those years), so it's not as if I don't have first hand experience about those conditions (perhaps more than most Canadians due to the kind of job I had back then). And I've read other threads where people talk about their range experience in Canada (and by extension much of the northern U.S.).
 
My experience is that 3 hours is a very good time to stop for 15 minutes. 3 hours at 70 mph is 210 miles.

I can't speak for other parts of the country, but in the South and North East, NOBODY drives 70 mph on the interstate. In fact, the norm is creeping up constantly. I typically set AP at about 82 mph and I'm still being passed steadily. There are a few slow pokes over in the far right lane that I pass, but there is one of them for every 10 that passes me. I'd say the average speed is more like 80 mph (at least).

So, 210 / 80 = ~ 2 hours 37 minutes of driving - and then sit at a charger for 40 minutes. Most people simply aren't going to be limited like that when they've been driving an ICE that allows them 400 miles at 80 mph and a re-fuel takes less than 10 mins - AND you can stop at your will, NOT where there happens to be a supercharger.
 
I know many people, my wife being one, likes to drive 4 to 5 hours between stops. We often make a couple of 9 to 10 hour drives and she is hell-bent on one stop. I have to beg her to take the Tesla and she wines about it the whole trip. And that's coming from a person who drives a Tesla to work everyday and thinks it's the greatest thing on earth (and refuses to drive anything else to work). We're due for another "travel" type vehicle (something 6 people can ride in comfortably) and she is dead set against the X because of the range factor. So, if a person who owns and LOVES a Tesla doesn't want to travel in a Tesla, how will the skeptics be convinced?
 
China's sliding auto sales may be obscuring a change in buyers' tastes

"For some in the industry, they say it will be the smartphone-like interface of the new vehicles that will really attract buyers. Those consumers are increasingly using internet-connected services such as food delivery for daily life, especially in China."

I think that this is a significant aspect of the modern car, in addition to the EV powertrain and autonomous abilities, and a major attractor for the modern tech savvy consumer. Basically, the whole "iPad on wheels" or smartcar aspect, where much of the car can be controlled and improved/upgraded by the carOS, and where users have greater control and interactions with their car via the carOS and smartphone companion app.

Tesla takes pole position on this dimension as well, in addition to its lead on the EV powertrain and autonomous abilities. While the legacy OEMs may be able to compete on the latter two dimensions, they will have a hard time on the software front without partnering with a big software company, like Google, Apple or Microsoft.

I read that article and didn’t notice any mention of Tesla which seemed very strange. I did skim, so may have missed it.
 
TSLA down a bit in pre-market trading - I believe it's the effect of Citi downgrade from $273 to $238.

Citibank analyst Itay Michaeli is one of the permabears who never had a Buy rating for Tesla:

upload_2019-4-18_15-2-39.png


This downgrade is of course totally not related to the fact that today is Options Day, nor to the fact that the next trading day will be Autonomy Investor Day. ;)
 
"Tesla short sellers are using a white nationalist meme to spread their FUD. You read that right. They're elevating the same kind of people responsible for the horrible tragedy in New Zealand, just to make money."

"Either TSLAQ have some white nationalists amongst them, or Russian bots. It's definitely used to spread nazi propaganda."

Alter Viggo on Twitter
Lisa G-Punkt on Twitter

Not overly surprising, since Mark B. Spiegel has been using alt-right arguments in the past. I guess this is yet another sign that the TSLAQ fish rots from the head down?

The TSLAQ cult is attracting the best, brightest people ...
 
Saw a Reddit post of a short video by a Dane.

He argues Denmark SR+ sales should be through the roof.

The SR+ meets the criterion of EVs below DKK 400,000 ($61,000) to only pay 20% sales tax vs. 150% sales tax for a normal ICE. This should mean a $50k Model 3 is competing against $22k ICE economy cars.

Can anyone from Denmark confirm this? Is there some loophole that people use to avoid the 150% sales tax, such as importing "used" cars from Germany with 50 km on them?

Denmark has 200k+ passenger vehicle sales per year:


$22k ICE vehicles is at the low end of the car price spectrum - the Model 3 SR+ should capture a big chunk of the sedan market, which should be beyond 50% in Denmark.
 
I read a little bit about waymo. 600 cars (total) 400 passengers (in total) drivers still chaperone every car. Only working in 1 city. valued at 175 billion dollars.
Investors are such idiots. The minute they sit in a HW3 tesla and hear how many are being shipped to *paying* customers every day...any of the sane ones should reach for the BUY button.
I think tesla have been a bit lax in reaching out to hype the stock to investors. They are probably guided be elons perfectly rational belief that managers of billions of dollars do perfectly unbiased research. Meanwhile google is wining and dining those same people and keep whispering 'waymo' in their ears...
 
My understanding is that no, it is never told about the 8 cameras. It learns what to output through machine learning.

Maybe it's my background in astronomy that confuses me, because there you really care about where you are looking - and to that end the pointing of the telescope is important.

I really have to think that also in the traffic is not enough to be looking, you also need to know in what direction you are looking - and to that end the a priori knowledge of each camera's pointing is kind of natural.
 
There's nothing wrong for calling for accurate numbers. Pretending that "winter highway driving halves the range" when in the worst realistic driving case outside of extreme climates (Minnesota in the winter in -20F / -29C) you're actually getting the range 2/3rds (sub-300 vs. around 200) does nobody any favours. Two thirds is not one half.

The loss of range in cold weather absolutely should be discussed. It's a very significant factor to owners in places with cold winters - your 300-mile-car becomes a 200-mile-car in a Minnesota January. But it shouldn't be discussed with hyperbole - e.g. as though a 300-mile-car becomes a 150 mile car, or worse, that this happens in any below-freezing weather. Don't you agree that it's important to give accurate numbers?
The post you were responding to in this context was when @jerry33 said: " ...the 400 mile range declines to a bit over 200".

If you are striving for accuracy, you missed the "a bit over" part.

Several folks here, myself included, have experience where 40% loss in highway winter driving can happen... 240 is in the ballpark of his claim. Toss in a couple of stop/starts, and just over 200 is entirely accurate (you were the one that added the highway qualifier, not him).
 
But where are Tesla's limited resources better allocated?

Dan
Given that virtually every person I talk to at shows says 400 or more miles is required (e.g. I want to drive from Dallas to Houston without stopping. 85 mph speed limit, mostly driving 90-95 mph--and that doesn't mean arriving in Houston with zero miles) before an EV is a considerations, I'd say that it's pretty important use of resources. EVtripplanner indicates that in an LR M3 you'll use 350 rated miles for this kind of trip and 425 rated miles in my S85.
 
The post you were responding to in this context was when @jerry33 said: " ...the 400 mile range declines to a bit over 200".

If you are striving for accuracy, you missed the "a bit over" part.

Several folks here, myself included, have experience where 40% loss in highway winter driving can happen... 240 is in the ballpark of his claim. Toss in a couple of stop/starts, and just over 200 is entirely accurate (you were the one that added the highway qualifier, not him).

I don't want to drag this out further - hence I didn't reply to your first post on the subject - but that you've now written two posts on the subject, so be it. What's your full cycle summer range (e.g. not "drove X miles and then extrapolated" - as close to ~100% to ~0% as possible), and what's your full-cycle winter range (same as before, no "extrapolated half the cycle"), under what conditions, and are you including any phantom drain in there?