Tesla Fans!

[ajdelange]

You are correct, I think the bigger problem is the EPA using 48 mph. I never drive that slow on the HWY but that is what they use so it's what we get.

"Highway" impliea vastly different conditions in different parts of the country. It is rare, in my experience, that one can get on the highway and cruise consistently at 70 mph. I can rarely do that on a superhighway (because of traffic) and certainly not on the regular higways (or superhighways here in Quebec where the speed limits are 62 mph or less). And I drive the NJTPK. In my experience "highway average speeds are perhaps not as low as 48 but not much higher. I find my Teslas turn in consumptions quite close to but a little better than the EPA rating, YMMV.

Sponsored

 

[ajdelange]

If I understood his comments correctly, he says that Tesla's results are consistently estimating their EPA results higher than their (InsideEVs') testing, la.

What is wrong with this?

IMO it is failing to emphasize that they (InsideEvs) are comparing their apples to Tesla's and the EPA's oranges,

 

[Moonjock]

"Highway" impliea vastly different conditions in different parts of the country. It is rare, in my experience, that one can get on the highway and cruise consistently at 70 mph. I can rarely do that on a superhighway (because of traffic) and certainly not on the regular higways (or superhighways here in Quebec where the speed limits are 62 mph or less). And I drive the NJTPK. In my experience "highway average speeds are perhaps not as low as 48 but not much higher. I find my Teslas turn in consumptions quite close to but a little better than the EPA rating, YMMV.

Yes, you are correct. Traffic is not an issue where I live. I have no problems maintaining what ever speed I want. My daily commute to work is an average of 60 mph. My trips to civilization very depending on time of year. Average of 65 to 80 mph.

 

[HimuraMOdo]

I will avoid writing a whole thesis on this here, and also state that I am not a participant, but a close observer in the energy industry with a very specific angle of view.

We are seeing that renewable energy solutions are having significant problems in generating sufficient power in the face of increasing energy demand. The financial and environmental costs of increasing renewable energy resources are unlikely to offset even just the increase in the energy demands.

EVs make up only 4% of American auto sales at this point, 3% of that is Tesla. The battery technologies are advancing at a rapid rate, and the efficiency of batteries is also superb. However, generating and storing enough energy to move vehicles the way we currently do is likely a losing battle, because renewable energy's efficiencies are far behind battery capabilities. So, there is a major delta that I don't think we figured out how to address. I am not even touching the recycling of batteries at this point. The word "recycling" is very misleading in this context - today.

In the meantime, increases in EV usage will have to be accommodated, because the majority believe it will positively impact climate change and our quality of life. However, it is estimated that to meet the demand of energy EVs will require will result in an increase in coal production to the tune of 26%. In addition, the financial and environmental costs of manufacturing batteries remain relatively steady, although a relatively small reduction is observed.

I am in the opinion that unless we address our mid-term energy production and electrical grid problems, EV movement will stall. This may happen at a point in which people who are already skeptics of EVs will not be able to use the technology due to reasons that are not directly related to EVs. For example, when people cannot use their EVs to move their families out of harm's way during a major storm or fire because the electric grid was down or electricity (and water) has been rationed due to a 30-year drought, they will turn against EVs on mass.

So, in this rather pessimistic scenario, will EVs ultimately reduce our carbon footprint? It may have the exact opposite effect if we continue to view the scope of the discussion as "EV," "renewables," etc., and not as energy production in the balance of environmental well-being.

We do not have a coordinated, multi-angle strategy as a country to address these issues and I do think we are ignoring some of the relatively easier solutions due to misinformation that has been spread over decades.

The real question in my mind, what will it take to make fundamental changes, while accepting that we will have to make uncomfortable compromises and sacrifices that won't be evenly distributed across society. And frankly, some of the answers lie with Tesla, not necessarily as a company, but as a lessons-learned study.

Well said. It's good to have both pessimistic and optimistic views on the same subject.
Here are a few comments regarding EVs:
1. EV adoptions are ramping up quickly and it will take up larger percentage of the overall vehicles on the road as time goes.
2. EVs are way more efficient than ICE vehicle. ICE engine is at best 30% or so in thermal efficiency. that's on par with the average coal power plant thermal efficiency(worldwide). However, new/modern coal power plant can have ~45% efficiency. Also, the increased energy demand won't be met by coal power plant only, there are other types of power plant which have better efficiency than ICE vehicles.
3. Electric grid innovation would be driven by the increased energy demand. EVs are part of the increase but not the only driving factor.
4. Transmitting electricity is way more efficient than transporting fossil fuel.
5. Most family would have both EV and ICE vehicle. (At least I will). That would be the most versatile solution. There might be a scenario where grid is down and EV won't be driven. But don't forget gas could have shortage as well.
6. Battery recycling would happen one way or the other. There are start ups that do this exact thing.

I still believe EVs will reduce the total carbon footprint comparing to no EV at all.

 

[Riventures]

@HimuraMOdo hard to argue with your points. I'll still pick one, which has been my main argument outside of these forums: EVs and other electricity-hungry approaches will certainly drive electric power plant constructions and expansions. However, such constructions and improvements to the electrical grid take significantly longer time than the increasing demand partly fueled by EVs and will lag by 5 to 8 years. That is very worrisome to me.

 

[Denver_Paulie]

The obsession with EV range is for Tesla fanboys to spout on the internet how much better a Tesla is than any other EV because it can potentially go 300 miles in ideal conditions.

What people don't understand is that it is not about range, it is about charging speeds. My inefficient Porsche Taycan will get 175 to 200 miles of range on the highway, which is not as far as a more efficient Model S will go. However, Electrify America chargers are 150 or so miles apart, and I can charge up to 275kW. So, I can drive from Denver to Vegas, or Phoenix, or Kansas City just as fast, or faster than any Tesla even if they can go 100 miles further on a charge because I can charge faster and spend less time sitting at a charging location.

Don't get me wrong, I ordered the larger pack for my R1T, but I also spend time in southern Utah in the Lake Powell area, and there are no CCS chargers close by. I am betting that 95% of Rivian owners will not venture more than 5 or 10 miles off a paved road so there should be no concern over range.

Charging at home and expanding the CCS charging network, that is what all future Rivian owners should be concerned about, not how far a Model S will go at 70mph on some highway.

 

[htownsportsfan]

I have a leased Dual motor Model 3 for two years now and I really enjoy the car. It's fun to drive and provided a great experience. I've only had one issue with the car that resulted in a mobile service appointment. Overall I would say I'm super glad I got the car. It's saved me a lot of wear and tear around town on my pickup. In another year I will be turning in the Model 3, selling my GMC Sierra pickup and *hopefully* will have taken delivery of my R1T by then.

I have a love/hate relationship with Tesla's FSD software. I love the capabilities that the car has today, but frustrated that I've been paying for "Full Self Driving" for two years now and likely will never get to experience the released product by the time I return the car in late 2022.

I really like were Tesla is taking FSD however. It truly is a unique differentiator for them from every other auto maker using LiDAR and hi-def maps. I hope the general public clues into the differences here. Perhaps Rivian changes their approach on this as well at some point.

I still highly recommend Tesla cars, if you haven't driven one I would recommend it. Maybe rent one for the weekend from Hertz!

 

[Temerarius]

The obsession with EV range is for Tesla fanboys to spout on the internet how much better a Tesla is than any other EV because it can potentially go 300 miles in ideal conditions.

They lost that crown...

 

[SeaGeo]

I really like were Tesla is taking FSD however. It truly is a unique differentiator for them from every other auto maker using LiDAR and hi-def maps. I hope the general public clues into the differences here. Perhaps Rivian changes their approach on this as well at some point.

Having seen more than a few recent FSD videos from people I trust, it shouldn't be available to the public in it's current state. At least without a significant amount of training. You should also have to disclose to your insurance that you're using beta software to control the car.

It's very cool when it works, but also very disappointing how frequently it doesn't and *how* it doesn't.

 

[boneil1]

Having seen more than a few recent FSD videos from people I trust, it shouldn't be available to the public in it's current state. At least without a significant amount of training. You should also have to disclose to your insurance that you're using beta software to control the car.

It's very cool when it works, but also very disappointing how frequently it doesn't and *how* it doesn't.

What kind of training would you have them do? Drivers ed? If they have a drivers license then they already went through driver training. We don't require parents to go through some special training when teaching their kids to drive.

 

[Riventures]

In large part, my original question starting this thread wasn't about the vehicles, but the perception towards the product and the contribution of socio-economic factors if any. I stated my question very broadly to be able to get as diverse input as I could. I am still in awe of the Tesla fan base.

The human-machine interface is particularly interesting. It appears that it is not generationally limited, multiple generations from Baby Boomers and on like it and hate it. Various income levels seem to do so as well. I would have thought that increased automation would be a turn-off for many, but it appears to be gaining increasing acceptance.

I had strong objections when manufacturers started incorporating more and more electronics into cars. I thought that is because I am old and I learned to drive, interact and use these tools (cars, trucks, etc.) in a different era. My objections, obviously, had no effect. Now we are facing a level of automation that I could not even imagine. The next wave is going towards cloud-based mobility service products utilizing OTA. Well if I (or my company) want to get on that bandwagon, I better understand the general consumers' preferences better than I do today.

As a personal preference, however, I truly am concerned about a developer writing code in the corner of a dark room to control the behavior of the vehicle I am driving, where s/he has nowhere near the understanding, ability, or experience that me on how to drive a vehicle with four wheels. I won't be able to stop it, so I have to accept and move on

 

[SeaGeo]

What kind of training would you have them do? Drivers ed? If they have a drivers license then they already went through driver training. We don't require parents to go through some special training when teaching their kids to drive.

At a minimum, some basic training on the psychology associated with helping train machine learning in an environment like this (for reasons discussed below), but also what the strengths and weaknesses of the system, how to anticipate where it's likely to have issues, how quickly to overrule the system, and honestly probably some simulation time training people to understand how quickly the system can and will go wrong and what it's like to override it in a safe environment.

Most States require additional training to be a drivers ed instructor. As a parent, at least in the States I've lived in, you don't replace drivers ed. That being said, there are significant differences between a teenager the FSD system. First off is most teens I know of have been observing driving *with a human brain* for somewhere between 14 and 18 years before starting driving. Generally people are progressively advanced into environments that start off as low risk and those are graded to become more and more difficult (ie, start in a parking lot and progress to neighborhood streets, then the City). As people learn, they develop *skills*. FSD does not develop skill in the same way. Once a human learns how to make a turn, they're highly unlikely to fail at a routine turn after that point. When a parent is riding along with a child who is in the process of learning (for example, training hours) the general driving skills aren't usually the major concern. You can typically anticipate pretty easily where your kid may or may not struggle or need help. Additionally, most teens, despite having poor risk management skills, still have a sense of self preservation. They don't try to barrel without warning into a monorail pole.

Conversely, FSD works fine in many environments, and then in a similar environment it decides to completely fail a left turn. So riders can be lulled into a false sense of comfort. And a not-insignificant amount of the issues I've seen are behaviors that require a raised sense of awareness to reliably intercede. Those can be breaking hard in the middle of the road, Or dart left/right for no apparent reasons, struggling with turns in general, trying to run into objects. Etc.

Additionally, the users with FSD right now end up playing a game of chicken with FSD for whatever reason wanting to see if it can work its way out. The video linked below is an example of this. Granted, this is an older version of FSD (from this year), but the FSD behavior observed here is absolutely dangerous, and the response the driver ends up making when FSD about runs into a planter (for example) is to stop in the middle of the road. Which doesn't surprise me, but very well could also have been dangerous.


Here's the longer video of the same drive, I've timestamped to share a portion before FSD tries to run *into a monorail pole* more than once seemingly because the software didn't see the massive column in the middle of the road. Before that occurs, the software does a terrible job even navigating getting making the right turn. You can see the steering wheel jumping back and forth. And then when it finally tries to make the turn and the driver (rightly) disengages FSD he then says "awww I shouldn't have..." YES you should have dude. The car made you uncomfortable because it appeared like it was going to run into a 3-ft wide concrete column. At 17:03, while driving on a straight road, it tries (again) to run into a column. At this point, the driver was obviously alerted to the issue because it failed navigating the column on the right hand drive. But what if he wasn't and was in a situation where FSD had been performing well for the prior 20 minutes? There's *very* little time to react in a situation like that, and the typical reaction time for a human driving a car is on the order of seconds.

 

[Scott]

As a personal preference, however, I truly am concerned about a developer writing code in the corner of a dark room to control the behavior of the vehicle I am driving, where s/he has nowhere near the understanding, ability, or experience that me on how to drive a vehicle with four wheels. I won't be able to stop it, so I have to accept and move on

I think you might be surprised how little of the autopilot system is hand coded. A lot of it is using machine learning (ML) platforms and feeding in the right training data so the ML learns what to do. generating this data and classifying it well is a big part of the problem. I am sure there are many hand coded things but the real hard problems here are likely around computer vision. Identifying objects from camera feeds and placing them accurately in a 3d map is hard as hell. That is where Tesla mining customer data gives them a big leg up. As a general rule with AI: the company with the best training data will have the best AI. That said, relying on pure ai and less on a developer in a dark room probably doesn’t increase anyones comfort level.

Sponsored