[GutsyGibbon]

Originally Posted by SreeT

You do realize you are actually doing more environmental damage buying a new EV while your old one is fine right? The best way you can help the environment is to keep using your tesla till it breaks.

It is presumptuous to think I am parting with my Tesla. We are a 3 car household, and I have a gas guzzling pickup truck, that I would like to replace.

Clearly, you have done your online research, and I cant offer explanations to each of them, as I will end up looking like a Tesla PR boy. We can agree to disagree. You can have your driving experience, and your online facts, and I will stay content with my car.

[landcruiser123]

I’ve seen (and driven a few) Teslas and I’ll tell you that they aren’t finished too well. It gives people snob appeal to be part of the EV club, but I honestly think if Audi, Volvo, or Lexus starts making equally good EVs, Tesla will feel the pinch.

Model S will be successful in India, but I feel Model 3 is going to be overpriced in India.

Highly successful EV makers are going to be the VAG group and a handful of Chinese companies.

[extreme_torque]

Originally Posted by SreeT

It seems to me that your only source of information is telsa, and ofcourse they say they are right and everyone else is wrong (Read: https://scale.com/blog/is-elon-wrong-about-lidar). What's wrong with having additional data to work off of? It's not like they are LIDAR only. Again, toyota is very conservative when it comes to technology. Which is why their cars don't kill people.

Right another link and a link, no less, from a company who's bread and butter seems to be annotation for 3D sensor data from Lidar and Radar. They will come out and say Lidar is useless. Its why I said context is totally lost on you. How do human judge distances? Do you have a Lidar on your head? How accurate do you need to be when driving a car while you are judging distances?
What is a neural network? What is Tesla's self driving AI chip and why their's is better than NVidia's? Or that data is the king when it comes to training neural networks and who has the most amount of real world data?

Originally Posted by SreeT

Yes but consider this: The ID3 sold 7,754 cars in just 4 months, while Tesla sold 7,770 for the year. The reason apparently for tesla beating in December but not the previous months is because they try to sync deliveries with earnings reports or something to keep their share price high.

From Tesla looses to ID.3 to "consider this" but the context again is totally lost on you. Do you understand supply constraints? VW is being built in a VW factory in Europe primarily being sold in Europe. So why dont you compare US market sales where Tesla does have a factory vs ID.3 or even China? Do you know how often cars are shipped to Europe and how many and from which factory?

Originally Posted by SreeT

What do you mean by sporadic links? I'm linking facts that support my point. You can argue my facts if you think your viewpoint is correct, you can't just stick your head in the sand and hope to win arguments based on words rather than facts.

The issue is exactly that its not your view point, just some random links thrown around from the internet. Convenient facts. If what you say is true Tesla should go bankrupt soon? Lets see whose head is in the sand.

[vb-san]

Mercedes looks to be the first among the big 3 Germans to dive into the electric onslaught. The plan to launch as much as 6 models (in the EQ range) before the close of 2022 looks pretty aggressive, but probably the best approach to be leader of the chasing pack. EQS sedan in the first half of 2021, followed by EQA and EQB SUVs (based on GLA and GLB respectively), and then the EQE sedan to close off the year. And the bigger EQ SUVs based on the GLE and GLS will follow in 2022.

Source: Autocar UK

With this pace I feel they will have enough material to fight out with Tesla, provided they can find a solution to match Tesla’s supercharger network; I think that’s one area where Tesla have a distinct advantage. It will be a win-win situation if the major manufacturers reach an agreement to use Tesla’s charging network – something like a USB-C adoption in the EV world.

Considering how well the new-age Mercedes cars are designed, I feel they will have a clear advantage on the looks department when compared to their German rivals. BMW is on an over-styling drive and Audi derivates mostly will stay classy plain-Jane.

[SKC-auto]

Some how links PM emissions we have seen over centuries to electric cars and health risks in some country.

The research is not clear on the dust particles, as not just the cars weight, dust particles are air borne by wind as well, difficult to quantify how much is due to cars weight.

Also, dust particles are PM10 and above which due to their weight settle fast. PM2.5< which are emitted by ICE are light weight and are in the air for longer periods and high chances of people inhaling and subsequent health risks.

All this dust particles are irrelevant in Indian context because we have too much dust in the air already. Ever rode a 2 wheeler next to a Volvo bus rear engine, you will see the dust sticking your face.

Please follow the below discussion:

[CarJunki]

The biggest buyers of carbon credit are European automobile manufacturers such as VW, Daimler, FCA. So much so that One particular seller will make 1.8 billion USD from selling carbon credits just from FCA alone by 2022.

Any one who thinks a HD map download every second dependent on a wireless connectivity surely does not know what is wireless or HD map. The combination is deadly fun.

I have a lot of patience when it comes to technology advancements usurping or disrupting the market. Companies doing "catalogue engineering" will never be able to match Innovation leaders.
There is simply no other player currently having a full in-house design of self driving hardware right from silicon to software.

Surprisingly VW CEO has slightly better understanding of the product and advantages and their target is to come closer to Tesla (not surpass, but come closer).

It is also fun to watch all these incumbent luxury brands losing money on every EV they sell as their battery cost are significantly higher coupled with the fact that they need larger batteries. If more bad mouthing comes from those brands it will not be a surprise.

Norway 2020 full year sales:
BEVs: 76,800 (up 27%, at 54.3% market share)
Of this Tesla Model 3 sold 7,770 units
ID.3 sold for the full year 7,754 units

The total size of Norway car market close to 142,000 cars. Quite a big automobile market. Hmm..

[GutsyGibbon]

Originally Posted by CarJunki

Many players are depending on radar & mobile connectivity to attempt full self driving by constantly downloading HD maps of current local area to identify objects that are around.

AFAIK, gone are the days of excessive downloads of bitmaps. Vector graphics changed all of this. Bunch of instructions like "goto x,y" and "line to x,y" constitute the graphics layer of the map, they are simply rendered by the GPU. If nothing is updated, the vectors are simply updated by the GPU without downloading anything. You can pinch, tilt, rotate the map without anything being downloaded.
Connected cars are cognizant of the cost of data, Tesla uses Wifi at home to download updates. People parking on the street far away from home wifi have issues with this. Most data sessions are bunched up together into one short burst of data, they are known to use crappy QCI (quality of service class identifier). Not the voice or real time gaming kind of data, this is worse than buffered video streaming class.

[CarJunki]

Originally Posted by GutsyGibbon

AFAIK, gone are the days of excessive downloads of bitmaps. Vector graphics changed all of this. Bunch of instructions like "goto x,y" and "line to x,y" constitute the graphics layer of the map, they are simply rendered by the GPU. If nothing is updated, the vectors are simply updated by the GPU without downloading anything. You can pinch, tilt, rotate the map without anything being downloaded.

This is not about vector graphics, most companies are not putting GPU in the cars and rather depending on server back-end to do object motion calculations.

Object detection is low load and thus can be done on lower capacity GPU. If simultaneous motion calculation of multiple objects has to happen in real time this will need high powered GPU's. Roads are not empty and multiple things are moving at the same time. With humans and kids around this becomes even more complicated.

The main gripe of Tesla with Nvidia was that run of the mill GPU was not delivering enough processing to crunch 8-10 high def camera feeds simultaneously and in real time, and forced them to look for own design of GPU. From 2018 onward all Tesla cars are coming with Tesla designed GPU (manufactured by TSMC). The competition is proceeding with map based approach partly due to lack of processing power of onboard GPU.

Sometimes this lack of GPU is also just supply and demand gap. A company is competing with crypto miners to procure GPU. Recently Nvidia shipped more than 80% of their latest GPU to miners directly. There is real supply side crunch and in user programmable GPU market there are very few choices right now. AMD is also production constrained (huge demand from cloud server market).

[maddy42]

What’s the lead on the chip side which Tesla has over their rivals? I do see we have a thread on chip shortage, will that affect Tesla?

Tesla keeps selling FSD to their customers. Can someone enlighten on the tech aspect of this and is it a bunch of horse poo 💩 as one side claims it to be?

Lastly there is a bunch of data being stored by Tesla. Are they using it to gain any revenue currently ?

Maddy

[CarJunki]

Lead: Simply none of the competition is having their own chip. Thus they have to be content with what their vendor can supply. In this case Nvidia has almost a monopoly and for them automotive demand is a small fraction. So lower priority. The chip shortage is applicable to companies who are sourcing off the self parts from others. In case of Tesla they are getting it contract manufactured, so no such impact.

Tech aspects of FSD: One thing is very simple. Let's take a similar example. Google voice recognition and translation is better than competition for simple reason they have more data to train their AI model (more phones collecting user data and feeding the AI). Similarly Tesla has more driving footage than anyone else on this planet.
AI is very data hungry, the more you feed it the better it gets. Conversely with lack of data you will have not so good AI engine. There is very real scenario where most cars/automobiles will have FSD in future. When we have autopilot in commercial aircrafts already today there is no reason why automobiles can not have FSD. There is enough self flying capability (both hardware and software) where swarms of self flying drones act as collective organism. So yes, very real. Calling it poo or whatever is personal choice.

Currently people are buying FSD package for some real money (i guess 10000USD which is 20% of a 50kUSD car and even higher portion for lower end Model3) , thus it is safe to say that all the collected footage is helping the cause. Why because more data means better the AI gets and thus more attractive the FSD functionality becomes.

As currently there is no regulation to certify if FSD has been achieved, all players will continue to call their implementation as beta. In general regulation will always lag technology.

[RoadFiend]

Interesting take from famous Kellogg Prof Mohanbir: TESLA vs PORSCHE https://www.linkedin.com/pulse/tesla...1td3x0ig%3D%3D

[extreme_torque]

Originally Posted by RoadFiend

Interesting take from famous Kellogg Prof Mohanbir: TESLA vs PORSCHE https://www.linkedin.com/pulse/tesla...1td3x0ig%3D%3D

Thats actually fair enough. I feel Model S is too long in the tooth now and we have to remember that it launched way back in 2012 and has had the same basic shape and exactly the same interiors for all these years, new features via software not with standing. Grapevine is that Tesla is working on a refresh. I hope its just not a nip and tuck and more comprehensive update. The same goes for the Model X as well and I am pretty sure it will be updated as well once Model S is.