$35,000 Tesla Model III Is Coming In 2017

[manly]

I'd say Tesla is a little wreckless with FSD. They're deploying before its ready and who know if the hardware in cars now will ever be capable of real FSD. Paying $6-7k is just nuts for something that might never work. That's not saying FSD won't happen, but Telsa is taking an approach that requires significantly better computer vision AI than its competitors.

This is a great summary IMO. By many accounts, Waymo is the industry leader in autonomous driving. They are testing "robotaxis" in Phoenix, public trials without a human safety driver. Meanwhile, Tesla struggles to navigate around parking lots. Yet some in this thread boldly assert Tesla FSD is years ahead of the competition. Talk about parroting Elon on this with no real-world evidence.

Speaking of add-ons, I've never been a huge fan of how Tesla chargers for package upgrades. For example, when I bought the car, EAP was $5k at delivery and $6k after. On the other hand, FSD was $3k at delivery and $5k after. Outside of the new computer for FSD, everything else is just a software authorization. (To note, the new computer wasn't even a thing when I ordered, so it wasn't given such a huge hike due to the hardware upgrade.) Honestly, I don't think there's a way to look at it other than a tactic to push people to upgrade their options to avoid having to pay more later, and I think that's a pretty scummy thing to do.

Of course, I paid for FSD because I didn't want to pay more later, and what ended up happening? Tesla accidentally dropped it to $2k when doing the EAP to AP switch, and my friend, who ordered his car without FSD, ended up paying less to upgrade his car post-delivery than I paid for the feature at delivery. I still wish I would've just ordered AWD instead of FSD.

We've talked about these changes before, and I think it's an indication of (business) confusion rather than scumminess. They obviously rejiggered AP+FSD in a way they could claim to have delivered partial self-driving functionality this year. Elon is still flapping his gums that they will deliver on-roads FSD sometime next year (pending regulatory approval), but if you use Smart Summon as a baseline, that seems like quite a stretch of the imagination.

I don't have any problem with allowing people a choice of paying less now, or more later. The essential problem that you hint at is that FSD is vaporware. So not only did they get a free loan from all early adopters, they also changed pricing in a way that punished early adopters.

 

[Aikouka]

As I use the screen for various things, I'm curious if Tesla will update the MCU again. At least in my experience, the screen sometimes feels like using a phone that's about 2-3 years old. If I recall, they're using a fairly anemic Atom chip, which probably explains a lot.

They obviously rejiggered AP+FSD in a way they could claim to have delivered partial self-driving functionality this year.

I'm not sure if it seemed that way, but I don't care about the reorganization of the features. Sure, it led to an accidental and unpredictable massive reduction in price that would've been better for me compared to the at-delivery cost. But as noted by my choice in words, who could've expected that would happen anyway?

I don't have any problem with allowing people a choice of paying less now, or more later. The essential problem that you hint at is that FSD is vaporware. So not only did they get a free loan from all early adopters, they also changed pricing in a way that punished early adopters.

I think it's worthwhile to keep in mind how there's almost a bit of a fervor around Tesla. You'll see people remark that they're fine with paying for things because it "helps the mission". I see Tesla as a company that has been branded as the "good guy" due to their green, Earth-saving initiatives. (Also add in how they're in the David-and-Goliath fight against the incumbent automakers.)

On the other hand, artificially raising prices for no reason comes across as being against that lofty ideal. I don't see any reason for that action other than to push people toward buying it at delivery instead of considering purchasing it later. That's the sort of action that I expect from one of those companies that invokes sneers from people at merely a mention!

In other words, there's technically nothing wrong with it, but I don't have to accept what I consider to be a consumer-unfriendly practice.

To note, it's probably obvious, but I don't consider myself one of those Tesla zealots. I'm simply someone that bought a car from them and expects to get what I paid for.

 

[ponyo]

This is a great summary IMO. By many accounts, Waymo is the industry leader in autonomous driving. They are testing "robotaxis" in Phoenix, public trials without a human safety driver. Meanwhile, Tesla struggles to navigate around parking lots. Yet some in this thread boldly assert Tesla FSD is years ahead of the competition. Talk about parroting Elon on this with no real-world evidence.

We've talked about these changes before, and I think it's an indication of (business) confusion rather than scumminess. They obviously rejiggered AP+FSD in a way they could claim to have delivered partial self-driving functionality this year. Elon is still flapping his gums that they will deliver on-roads FSD sometime next year (pending regulatory approval), but if you use Smart Summon as a baseline, that seems like quite a stretch of the imagination.

I don't have any problem with allowing people a choice of paying less now, or more later. The essential problem that you hint at is that FSD is vaporware. So not only did they get a free loan from all early adopters, they also changed pricing in a way that punished early adopters.

Whether you believe Waymo is the FSD leader or not depends on if you believe HD maps and Lidar is the future and keys to solving FSD. You're obviously in the camp Waymo is the right way. I'm in the camp Waymo is the wrong way and Google is wasting their time and money with trying to apply the wrong solution. You drop that Waymo robotaxi in Atlanta Costco parking lot and lets see if it can do even half the things Tesla cars can do.

Tesla used to charge $5k for EAP and $3k for FSD. They rearranged the AP and moved some stuff from EAP to FSD and reduced the AP to $3k and increased the FSD to $5k. They have since merged the $3k AP into the cost of the base vehicle and increased the FSD to now $7k. Tesla can instantly recognize 100% of the AP revenue while they can only recognize portion of FSD they delivered. So it's kind of a wash and early adopters of FSD actually won since they paid less than people buying now. If you're early adopter of tech, you have to assume you're going to pay more. It's the price you pay to play.

Speaking of add-ons, I've never been a huge fan of how Tesla chargers for package upgrades. For example, when I bought the car, EAP was $5k at delivery and $6k after. On the other hand, FSD was $3k at delivery and $5k after. Outside of the new computer for FSD, everything else is just a software authorization. (To note, the new computer wasn't even a thing when I ordered, so it wasn't given such a huge hike due to the hardware upgrade.) Honestly, I don't think there's a way to look at it other than a tactic to push people to upgrade their options to avoid having to pay more later, and I think that's a pretty scummy thing to do.

Of course, I paid for FSD because I didn't want to pay more later, and what ended up happening? Tesla accidentally dropped it to $2k when doing the EAP to AP switch, and my friend, who ordered his car without FSD, ended up paying less to upgrade his car post-delivery than I paid for the feature at delivery. I still wish I would've just ordered AWD instead of FSD.

If Tesla actually solves FSD, your FSD option will be worth way more than the $3k you paid for it. True working FSD is probably worth $50-$100k. You basically bought call option that Tesla can solve FSD. In the meantime, you get to enjoy the beta features.

 

[Midwayman]

Whether you believe Waymo is the FSD leader or not depends on if you believe HD maps and Lidar is the future and keys to solving FSD. You're obviously in the camp Waymo is the right way. I'm in the camp Waymo is the wrong way and Google is wasting their time and money with trying to apply the wrong solution. You drop that Waymo robotaxi in Atlanta Costco parking lot and lets see if it can do even half the things Tesla cars can do.

I don't doubt the Tesla approach is the right one long term. However it requires significantly better computer vision and that's a huge issue. Lidar is just easier to work with. Once you get past sensor fusion its a pretty similar problem. Telsa has more real world data to work with, but I think google is ahead for now. I think who win depends a lot on the timeline. Google is more likely to be able to put out a FSD car earlier than Tesla. However due to Lidar its unlikely to be something a consumer will buy for at least a few years. Tesla has to be relying on FSD taking longer than the early 2020's or some break through in computer vision.

If Tesla actually solves FSD, your FSD option will be worth way more than the $3k you paid for it. True working FSD is probably worth $50-$100k. You basically bought call option that Tesla can solve FSD. In the meantime, you get to enjoy the beta features.

50-100k? Maybe to a commercial user. No doubt that level of system would sell for things like semi-trucks and taxis. It wouldn't be something you would own though. I'd be shocked if FSD was very successful at more than about $10k for the average buyer. That is an awful lot of taxi rides.

 

[kn51]

FSD is decades away. Fusion will come out before it.

 

[zinfamous]

They may have some impressive lane-keeping cruise control, but come back to me when it works in bad weather, poorly marked roads, and the insanity that is generally driving. Doing well under relatively well-controlled conditions doesn't mean much for the insanity that is driving. Remember, easy things for people are often very difficult for machines. The best self driving vehicles in suburban settings require highly detailed maps to really perform well. I remain highly skeptical of Tesla being able to get around this roadblock.

The 2020 VWs (the Golfs, anyway, because they are the first details to emerge) are equipped with a software package that responds to conditions of vehicles some distance up the road (if equipped with the same software), so that the car knows a slowdown is happening, there is a wreck, crap in road, etc.
This is one of those functions that I've always pointed to for defining properly-intelligent networked, fully automated cars. This is the kind of thing that removes a lot of the unnecessary guess work from each car, responding to their immediate surroundings in an instant. The sudden thing popping in front of your car is far more rare than the common road hazzard that has been sitting there for whoknowshowlong, just waiting for the next victim. And besides: humans are already terrible at adjusting to sudden things, so demanding that robots be flawless in this category is a bit short-sighted...but yes, it would be great.

"How do bees or ants move about?" That's how a properly-networked, fully-automated car fleet should operate in every second on the roads.

 

[zinfamous]

I did not say FSD is ready to go 100% now. It is getting closer to completion. I said when it is ready...

You might not remember but Amazon's same day delivery for most one can imagine was Science-Fiction not that long ago. Now, you just click and have Merry xmass every time you do.

Uh, this system also currently results in a lot of unnecessary maimings and death, so it definitely isn't merry xmas for a lot of people, hah...eh, heh? Amazon has of course cleverly contracted out all of this last-leg delivery, so you never see these accidents in any kind of marked vehicle (Amazon, anyway), and it's often private cars or rented vehicles, so the blame tends to mask the real culprit behind this kind of new, incredibly desperate driving from people making deadlines.

basically, Amazon put no thought into this system. They decided to just contract a bunch of non-Amazon employees, rather than deploy their own fleet with their own protocols and standards. ...and this is just hiring humans to do what they already do. So, it's not that innovative and it's actually insanely reckless.

 

[kn51]

Yeah, well TSLA hasn't exactly put that much thought in it either.

 

[ponyo]

I don't doubt the Tesla approach is the right one long term. However it requires significantly better computer vision and that's a huge issue. Lidar is just easier to work with. Once you get past sensor fusion its a pretty similar problem. Telsa has more real world data to work with, but I think google is ahead for now. I think who win depends a lot on the timeline. Google is more likely to be able to put out a FSD car earlier than Tesla. However due to Lidar its unlikely to be something a consumer will buy for at least a few years. Tesla has to be relying on FSD taking longer than the early 2020's or some break through in computer vision.



50-100k? Maybe to a commercial user. No doubt that level of system would sell for things like semi-trucks and taxis. It wouldn't be something you would own though. I'd be shocked if FSD was very successful at more than about $10k for the average buyer. That is an awful lot of taxi rides.

If you can solve FSD with computer vision, then what's the point of HD maps and lidar? You just wasted all that time and resources for nothing then. What is Waymo going to do? HD map all the roads on entire earth and update it real time 24/7? Waymo might work for buses that travel the same route the same way everyday. But Waymo is not the solution for worldwide FSD.

True working FSD will be worth $50-100k easily even for average car owner. Musk thinks $100-200k but he's always optimistic. If you own FSD car, you would let it operate as robotaxi anytime you're not using it yourself. Thus, your car wouldn't sit idle at your home or at work. It would be out driving and making money for you while you work or sleep. One car might make you $30k a year working as robotaxi while you're not using it. Tesla could charge $150k for FSD cars and people would fight each other to buy it. True working FSD is total game changer and if Tesla solves it first, Tesla would instantly be revalued at like trillion dollar valuation. And Uber, Lyft, Didi, etc will be toast.

 

[ponyo]

Yeah, well TSLA hasn't exactly put that much thought in it either.

Yeah, totally. That's why Tesla added all those cameras, radar, and designed and built their own self-driving hardware 3 computer from ground up and included it in every car whether people bought FSD or not. Because Tesla totally spent zero time thinking about the FSD problem.

 

[kn51]

Yeah, totally. That's why Tesla added all those cameras, radar, and designed and built their own self-driving hardware 3 computer from ground up and included it in every car whether people bought FSD or not. Because Tesla totally spent zero time thinking about the FSD problem.

So why does it run into fire trucks?

 

[ponyo]

So why does it run into fire trucks?

Because people are not paying attention and driving their car into it?

 

[ponyo]

Couple more pricing changes.

1. FSD increased from $6k to $7k.
2. Model 3 Performance price increased another $500 from $56,490 to $56,990.
3. Model 3 dual motor long range price increased $500 from $47,990 to $48,490. The battery range increased from 310 miles to now 322 miles. Existing owners will get OTA firmware update that will boost the range to 322 miles and also increase the HP by 5%.

 

[kn51]

Because people are not paying attention and driving their car into it?

So they are using autopilot?

 

[Midwayman]

If you can solve FSD with computer vision, then what's the point of HD maps and lidar? You just wasted all that time and resources for nothing then. What is Waymo going to do? HD map all the roads on entire earth and update it real time 24/7? Waymo might work for buses that travel the same route the same way everyday. But Waymo is not the solution for worldwide FSD.

True working FSD will be worth $50-100k easily even for average car owner. Musk thinks $100-200k but he's always optimistic. If you own FSD car, you would let it operate as robotaxi anytime you're not using it yourself. Thus, your car wouldn't sit idle at your home or at work. It would be out driving and making money for you while you work or sleep. One car might make you $30k a year working as robotaxi while you're not using it. Tesla could charge $150k for FSD cars and people would fight each other to buy it. True working FSD is total game changer and if Tesla solves it first, Tesla would instantly be revalued at like trillion dollar valuation. And Uber, Lyft, Didi, etc will be toast.

The point of Lidar and maps is computer vision is a much longer term problem. level 4 cars with lidar are just around the corner. Anything using cameras is going to take significantly more work. I don't think its a waste of time. being first to market is worth a ton. The lidar approach might not be the right one for level 5 driving, but that's the stage that's probably a decade off.

Who are these average car owners you think can afford $50-100k on top of the price of a car? Hell most people have a hard enough time affording a new vehicle. Letting your car run as a robotaxi undermines that whole point of a private vehicle. You no longer have a private space. You can't keep stuff in the car. You can rely on someone not damaging it, or leaving a mess. You have to worry about commercial insurance and how the car is going to be maintained/fueled/cleaned etc. what happens when someone damages your car and its out of service?

Then you have the issue of just scheduling rides either you're independent and have a tough time scheduling rides or you lease out through uber or something. They could provide logistics and riders, but at that point was is the point of even owning as car? You're essentially providing the capital for uber to run a taxi service. (Or whatever brand rideshare you hook up with)

In any case, you're proving my point that at $50k-100k its only viable as a commercial venture. FSD will be a game changer even at that price, but mostly because SDCAAS type taxis/busses/etc will eliminate a lot of private car ownership- especially in the most congested areas.

 

[Kaido]

My buddy just picked up an AWD long-range Model 3, custom order, fresh from the factory...it's been in the shop 4 times already. First time, it woke him up in the middle of the night because he smelled smoke, went into the garage, and it was just pouring smoke out from underneath the car, had to call the fire department & everything. Went back & forth to the shop a few times. Just had it delivered back after the third repair - they turned it on, drove it off the car carrier, and it started smoking again with the delivery guy still in the vehicle, so it's back in the shop again lol.

He's pursuing Lemon Law on it right now...he does want another one, but there's no way I'd drive a battery-powered car that kept smoking on me. Yikes!

 

[ponyo]

The point of Lidar and maps is computer vision is a much longer term problem. level 4 cars with lidar are just around the corner. Anything using cameras is going to take significantly more work. I don't think its a waste of time. being first to market is worth a ton. The lidar approach might not be the right one for level 5 driving, but that's the stage that's probably a decade off.

Who are these average car owners you think can afford $50-100k on top of the price of a car? Hell most people have a hard enough time affording a new vehicle. Letting your car run as a robotaxi undermines that whole point of a private vehicle. You no longer have a private space. You can't keep stuff in the car. You can rely on someone not damaging it, or leaving a mess. You have to worry about commercial insurance and how the car is going to be maintained/fueled/cleaned etc. what happens when someone damages your car and its out of service?

Then you have the issue of just scheduling rides either you're independent and have a tough time scheduling rides or you lease out through uber or something. They could provide logistics and riders, but at that point was is the point of even owning as car? You're essentially providing the capital for uber to run a taxi service. (Or whatever brand rideshare you hook up with)

In any case, you're proving my point that at $50k-100k its only viable as a commercial venture. FSD will be a game changer even at that price, but mostly because SDCAAS type taxis/busses/etc will eliminate a lot of private car ownership- especially in the most congested areas.

If and when Tesla solves FSD, you will be able to opt in or out of the Tesla Network. That will basically be Tesla operating the robotaxi in Uber like app. You're not going to be independent and you're not going to use it on Uber. Tesla is going to take a cut of every ride on the Tesla Network just like Uber. But robotaxi rates are going to be so cheap that Uber won't be able to compete with its human drivers. So I suspect Tesla Network will put Uber, Lyft, Didi, etc. out of business.

If you can make $25-$50k per vehicle by using it on Tesla Network, why wouldn't you do it? I can see people buying as many as they can to operate their own fleets on the Tesla Network. Tesla could manufacture and keep all the vehicles for themselves to use and operate on the Tesla Network. Tesla is already not letting people buyout the leased Model 3 at the end of the lease. Tesla wants to keep those cars for Tesla Network. So Tesla is thinking by 2022, they'll have solved FSD and can start operating their own robotaxis.

 

[ponyo]

My buddy just picked up an AWD long-range Model 3, custom order, fresh from the factory...it's been in the shop 4 times already. First time, it woke him up in the middle of the night because he smelled smoke, went into the garage, and it was just pouring smoke out from underneath the car, had to call the fire department & everything. Went back & forth to the shop a few times. Just had it delivered back after the third repair - they turned it on, drove it off the car carrier, and it started smoking again with the delivery guy still in the vehicle, so it's back in the shop again lol.

He's pursuing Lemon Law on it right now...he does want another one, but there's no way I'd drive a battery-powered car that kept smoking on me. Yikes!

That's pretty crazy and yet speaks volume about how good Model 3 is. You have someone pursuing Lemon Law and wants to buy another Model 3. How often does that happen for other cars?

 

[Kaido]

That's pretty crazy and yet speaks volume about how good Model 3 is. You have someone pursuing Lemon Law and wants to buy another Model 3. How often does that happen for other cars?

Yeah, he did get to try Autopilot on the highway once before it smoked out, and he loved it haha.

I mean, stuff happens. My Jeep Renegade was a non-stop lemon. There was probably a loose wire somewhere in the system that kept shorting everything out, who knows...because that's not everyone's experience with the same machine, and mass-produced stuff is subject to all kinds of weird variability. I really loved the design of the car...it worked fabulously in the winter, was roomy, and was great for driving in the city & fitting into small spaces. If Jeep's customer service hadn't been so terrible, I would have happily upgraded to a later model, like a 2018, with a stick-shift transmission, because I personally didn't care for the 9-speed automatic. FWIW, Jeep now has a "Social Care" specialist team that handles complaints on various forums, so it sounds like they're trying a little bit harder these days, thanks to the influence of social media:

Jeep Social Care Specialists (JeepCares) Are Here to Assist JL Wrangler Owners – Jeep Wrangler Forums (JL / JLU) – Rubicon, 4xe, 392, Sahara, Sport – JLwranglerforums.com

www.jlwranglerforums.com

 

[herm0016]

That's pretty crazy and yet speaks volume about how good Model 3 is. You have someone pursuing Lemon Law and wants to buy another Model 3. How often does that happen for other cars?

It's so good it does not matter that it's trying to destroy his house, he cant use it because its always being fixed... if it had a GM badge, you would be calling for all of them to be recalled and crushed.

true fanboy here.

 

[Midwayman]

If you can make $25-$50k per vehicle by using it on Tesla Network, why wouldn't you do it?

I already made that pretty clear. You are completely ignoring most of logistics of running your car as a business. Its the same fallacy of why people are actually willing to be Uber drivers. They're either doing something illegal like running incorrectly insured or bad at math for the long term.

 

[ponyo]

It's so good it does not matter that it's trying to destroy his house, he cant use it because its always being fixed... if it had a GM badge, you would be calling for all of them to be recalled and crushed.

true fanboy here.

His friend wants Tesla to buy his Model 3 back under the Lemon Law and yet he wants to buy another Model 3. You don't find that crazy? Why would he do that unless the car is amazing when it works.

 

[ponyo]

I already made that pretty clear. You are completely ignoring most of logistics of running your car as a business. Its the same fallacy of why people are actually willing to be Uber drivers. They're either doing something illegal like running incorrectly insured or bad at math for the long term.

So you run it as a business. If I could buy like 10 cars and have it make me $25k-$50k /car a year, you bet I would do it. So would most people. No one is going to turn down $250k-$500k a year running taxi service.

 

[herm0016]

So you run it as a business. If I could buy like 10 cars and have it make me $25k-$50k /car a year, you bet I would do it. So would most people. No one is going to turn down $250k-$500k a year running taxi service.

if being wealthy was that easy, everyone would do it. if it was that easy to make money with something like this, why would tesla even sell you the car?

 

[ponyo]

if being wealthy was that easy, everyone would do it. if it was that easy to make money with something like this, why would tesla even sell you the car?

That's also a possibility and I mentioned that Tesla might keep all cars it manufactures to use as robotaxis. But until Tesla or someone else actually solves FSD, it's pointless exercise.