Tesla Model Y

[Viper GTS]

The removal of radar is a really terrible move IMO. I won't be recommending 3/Y to friends and family any more; and if my 3P were destroyed tomorrow I would not buy a new one to replace it. If Cybertruck is similarly afflicted with this vision only nonsense I'll cancel that reservation too.

This combined with the launch of Lightning has taken me from not being willing to buy any vehicle that isn't a Tesla to likely to not buy another one unless they manage to pull an enormous software magic trick. The ability to see through weather is one of the few things lending legitimacy to the idea that a computer can be a better driver than a meat puppet.

If I've learned nothing from watching Elon turn $160M into Tesla and SpaceX it's that betting against him is not smart. But this just seems wildly premature.

Viper GTS

 

[ultimatebob]

The removal of radar is a really terrible move IMO. I won't be recommending 3/Y to friends and family any more; and if my 3P were destroyed tomorrow I would not buy a new one to replace it. If Cybertruck is similarly afflicted with this vision only nonsense I'll cancel that reservation too.

This combined with the launch of Lightning has taken me from not being willing to buy any vehicle that isn't a Tesla to likely to not buy another one unless they manage to pull an enormous software magic trick. The ability to see through weather is one of the few things lending legitimacy to the idea that a computer can be a better driver than a meat puppet.

If I've learned nothing from watching Elon turn $160M into Tesla and SpaceX it's that betting against him is not smart. But this just seems wildly premature.

Viper GTS

I'd be cool with it if it lowered the price of autopilot, but it doesn't seem like it has. It's still $10,000, and it seems like you're getting less for your money now.

 

[Kaido]

The removal of radar is a really terrible move IMO. I won't be recommending 3/Y to friends and family any more; and if my 3P were destroyed tomorrow I would not buy a new one to replace it. If Cybertruck is similarly afflicted with this vision only nonsense I'll cancel that reservation too.

This combined with the launch of Lightning has taken me from not being willing to buy any vehicle that isn't a Tesla to likely to not buy another one unless they manage to pull an enormous software magic trick. The ability to see through weather is one of the few things lending legitimacy to the idea that a computer can be a better driver than a meat puppet.

If I've learned nothing from watching Elon turn $160M into Tesla and SpaceX it's that betting against him is not smart. But this just seems wildly premature.

Viper GTS

As a computer nerd, I just don't see how a camera-only system is going to be as functional as also having radar & LIDAR. Elon is wishy-washy on his timelines, but so far he's delivered the S, X, E, and Y to consumers, along with a nation-wide charging network, solar panels, and rockets. So he's more in the politician-loved "directional truth" category of leaders. So far he's been pretty solid, just late on his timelines (extremely late with real FSD! lol). Supposedly the v9 update next month is going to be a huge leap from v8:

https://twitter.com/x/status/1392428986001592321

Ford will probably be on v2 of their F-150 Lightning by the time my Cybertruck becomes available for purchase, so I'll be curious to see if they have things like a 6.5 & 8-foot bed option, 400 & 500-file battery options, etc.

 

[Kaido]

I'd be cool with it if it lowered the price of autopilot, but it doesn't seem like it has. It's still $10,000, and it seems like you're getting less for your money now.

Get ready for DLC! Staring with a FSD subscription:

Elon Musk updates Tesla Full Self-Driving Beta v9 timeline and FSD Subscription once again

Elon Musk has again updated the timeline for Tesla releasing its highly anticipated Full Self-Driving Beta v9 software update and...

electrek.co

Plus "hostageware":

https://cleantechnica.com/2021/05/03/evs-hostageware-the-push-to-sell-stuff-over-the-air/

Apple grossed more than $64 billion from the app store, and all they have to do is run the servers & take a monster 30% cut. The math is interesting on digital subscriptions from Ford alone:

Auto analysts at Morgan Stanley say that Ford’s cloud computing platform that will enable digital subscriptions could become a $100 billion a year business — three times the company’s current market capitalization. Here’s the math according to Bloomberg: “$10 a month from the 75 million Ford’s on the road adds up to $9 billion a year — and an extremely profitable $9 billion a year at that.” [Of course, not all those Fords are OTA capable at this time.]

 

[thecoolnessrune]

I'll believe it all when I see it. So far, I've been hearing about v9 since last year, and the deadlines, like most Elon deadlines, just keep slipping. Hopefully the machine learning and all can really help pull a win, but I remain very skeptical about this camera-only approach. The entire point of giving this stuff to computers is because they have ways of seeing and working on data that humans are extremely inconsistent at. They must really feel that at the wavelengths detectable by the camera system's received light are enough to overcome the lack of having some sort of wavelength reflection system (LIDAR / RADAR). I guess time will tell, but I thought RADAR / LIDAR was a pretty important addition to self-driving cars.

 

[Kaido]

I'll believe it all when I see it. So far, I've been hearing about v9 since last year, and the deadlines, like most Elon deadlines, just keep slipping. Hopefully the machine learning and all can really help pull a win, but I remain very skeptical about this camera-only approach. The entire point of giving this stuff to computers is because they have ways of seeing and working on data that humans are extremely inconsistent at. They must really feel that at the wavelengths detectable by the camera system's received light are enough to overcome the lack of having some sort of wavelength reflection system (LIDAR / RADAR). I guess time will tell, but I thought RADAR / LIDAR was a pretty important addition to self-driving cars.

It looks like they are keeping the ultrasonic sensors, so maybe ultrasonics + cameras will do the trick in v9. Per the website:

* 12x updated ultrasonic sensors (has twice the detection range of previous sensors)
* 250 meters of range (820 feet)
* Detects both hard & soft objects

I mean, they do have a strong team at Tesla HQ. Getting rid of the radar seems logically odd, but I doubt Tesla would ruin their reputation by investing in something super dumb if they weren't confident in a better path forward for the future. Then again, the pressure from up high (Elon) is probably strong & he's been sticking to his "no lidar" guns pretty heavily. As far as Elon deadlines go, there's a pretty interesting tracker website for them:

https://elonmusk.today/

 

[ultimatebob]

As far as Elon deadlines go, there's a pretty interesting tracker website for them:

https://elonmusk.today/

Poor Elon... he's such a misunderstood soul

 

[bigi]

The removal of radar is a really terrible move IMO. I won't be recommending 3/Y to friends and family any more;

....

Viper GTS

Dude. Just relax. Can we at least see how this plays out?

What are you going to get? A bolt? leaf? kona ev?

Not personal, but overreacting is never good. It ain't the end of the word for sure.

 

[bigi]

FYI. It's been known for a while that Q2 production has been already sold out for weeks. If you order Y now, the wait time is around 10 weeks.

Watch out for the 10th of this month. Tesla will start postponed deliveries of new S and X. They might also announce a few new things like 350KV supercharging and 1st deliveries of the Semi Truck. Pepsi already installed superchargers for the semi.

The 250KV superchargers can already charge at 1000 miles / hour.

 

[thecoolnessrune]

FYI. It's been known for a while that Q2 production has been already sold out for weeks. If you order Y now, the wait time is around 10 weeks.

Watch out for the 10th of this month. Tesla will start postponed deliveries of new S and X. They might also announce a few new things like 350KV supercharging and 1st deliveries of the Semi Truck. Pepsi already installed superchargers for the semi.

The 250KV superchargers can already charge at 1000 miles / hour.

I'm curious to see how the new battery cells handle higher supercharging. Even at 250kW, we're pretty much at the max of what you can have the Model 3 battery pack take in without severe degradation. While faster charging will be good for the Model S Plaid and top-end Cybertruck, I think most people are still going to be charging at sub-250kW rates, if anything to ensure the batteries last.

 

[bigi]

^well, 350KV is still a rumor. The charing is not linear anyway so even current 1000 milels per hour does not happen during the full charge cycle. The taper still applies.

 

[thecoolnessrune]

^well, 350KV is still a rumor. The charing is not linear anyway so even current 1000 milels per hour does not happen during the full charge cycle. The taper still applies.

Agreed. I only supercharge my LR Model 3 maybe every couple of months and otherwise just charge from home or public charger. But the last time I used a Supercharger, it only stayed at 250kW from about 10% - 25%, then started falling off. By the time I was at 40%, I was back to the same ~120kW rates that almost everything can charge at. For the big battery packs, this is going to no doubt make a difference (500+ mile range), but currently, the car spends like 5 minutes at that Charge Rate. I know people have said of course taht Charging needs to get faster, but for me, by the time I'm done having a stretch, getting rid of trash, using the restroom, etc, I'm at 80% and ready to go again.

 

[drnickriviera]

I was back to the same ~120kW rates that almost everything can charge at.

Looks at my Nissan Leaf and Bolt with a 50kw max charge rate. "Why can't you be like them?"

 

[bigi]

Based on delivery dates reported by current buyers of Tesla Model Y, it is possible that Q3 production is nearly sold out.

Currently Vins 214xxx range are being delivered.

 

[bigi]

Tesla Radar was the Weak Link in FSD

https://twitter.com/x/status/1407039878425157642

 

[Kaido]

Tesla Radar was the Weak Link in FSD

https://twitter.com/x/status/1407039878425157642

Yikes! I wonder how compact LIDAR would do. Camera-only (well, plus the ultrasonics) seems like it wouldn't do too well in inclement weather like rain or snow...

 

[repoman0]

Tesla Radar was the Weak Link in FSD

https://twitter.com/x/status/1407039878425157642

JFC these guys need some help on their sensor fusion and estimation / filtering. Seeing easily solved stuff like this result in the sensor getting canned rather than filtering software fixed and made robust makes me trust FSD even less.

 

[heymrdj]

Yikes! I wonder how compact LIDAR would do. Camera-only (well, plus the ultrasonics) seems like it wouldn't do too well in inclement weather like rain or snow...

This is what I’m waiting for proof on. Completely white out roadway, no lines available and signs obscured by ice. Don’t interact with the car and show what it will do under those circumstances with no input.

 

[thecoolnessrune]

This is what I’m waiting for proof on. Completely white out roadway, no lines available and signs obscured by ice. Don’t interact with the car and show what it will do under those circumstances with no input.

It's interesting because there's been several times that I've been through winter storms bad enough (and fresh enough) that the only way I drove home was a combination of:

A. Having driven the road before
B. Guestimating the distance of my car from the guardrail to infer if I was on the road.
C. Feeling the vibration in the road to know if I was getting close to the outside berm.

In all of those cases, "Vision" only would have helped based on inferring how wide the road must be and how close I was to the guard rail. Radar wouldn't help here at all. For Tesla in its favor, it could utilize GPS and maps to help determine where it's at on the road, but the problem with that is the "final say" still has to be based on what the cameras see, because Maps won't help in construction or road hazards.

At some point, it comes back to the classic condition that Humans are absolutely awful when it comes to judging risk. There will no doubt come a time in self driving cars where people go "I could have navigated that" when the car says it won't, just like there's people who drive through deep water and end up with a destroyed, floating car.

On another note, I think some of this indicates how useful it would be to have some sort of marker system for roadways for autonomous guidance. Bonus if the flexibility would enable dynamic information for construction or accident avoidance. But you know that if such a thing were done, it'd open up a whole can of worms around security and safety because there's some waste of oxygen it would try to do something destructive with it.

 

[heymrdj]

It's interesting because there's been several times that I've been through winter storms bad enough (and fresh enough) that the only way I drove home was a combination of:

A. Having driven the road before
B. Guestimating the distance of my car from the guardrail to infer if I was on the road.
C. Feeling the vibration in the road to know if I was getting close to the outside berm.

In all of those cases, "Vision" only would have helped based on inferring how wide the road must be and how close I was to the guard rail. Radar wouldn't help here at all. For Tesla in its favor, it could utilize GPS and maps to help determine where it's at on the road, but the problem with that is the "final say" still has to be based on what the cameras see, because Maps won't help in construction or road hazards.

At some point, it comes back to the classic condition that Humans are absolutely awful when it comes to judging risk. There will no doubt come a time in self driving cars where people go "I could have navigated that" when the car says it won't, just like there's people who drive through deep water and end up with a destroyed, floating car.

On another note, I think some of this indicates how useful it would be to have some sort of marker system for roadways for autonomous guidance. Bonus if the flexibility would enable dynamic information for construction or accident avoidance. But you know that if such a thing were done, it'd open up a whole can of worms around security and safety because there's some waste of oxygen it would try to do something destructive with it.

But part of that is where radar comes in. I’m not interested in what I can see on a foggy day, high resolution eyes work fine to the limits of complete vision obscurity. Radar tells you what’s hiding beyond the fog. It tells you there’s a stopped car in the middle of the road when your vision is less than 100ft. It still provides what vision and LiDAR doesn’t. Not much for flat landers, but those in the mountains should certainly be appreciating it.

 

[bigi]

It will be fine. Moaners will moan.

Dirty Tesla dude has been very transparent in his videos. The latest one shows that no radar MY has done fine.

More to come...

 

[ponyo]

It will be fine. Moaners will moan.

Dirty Tesla dude has been very transparent in his videos. The latest one shows that no radar MY has done fine.

More to come...

I don't worry too much about Tesla haters anymore. They were proven wrong before and they'll be proven wrong again in the future. You have people here claiming they're some expert on AI and driving but in reality, they're full of shit. Tesla may or may not be on the right path for self driving. We'll know soon enough in couple years as everyone else seems to be going lidar and HD mapping route vs Tesla with their vision only AI. It could very well be both end up at the same endpoint solution but with different paths chosen.

I have sizable investments in both Tesla and Xpeng. Tesla is vision only AI. Xpeng is lidar and vision. I think Tesla will solve FSD before anyone by couple of years but I'm not smart enough to know the exact future. I don't worry about small bumps in the road or what may look like short-term setbacks. Progress is never smooth and linear. I just know there are lot of smart people working on FSD problem and some company will eventually solve it. And we'll be so much better off as society when that happens.

 

[repoman0]

I think Tesla will solve FSD before anyone by couple of years but I'm not smart enough to know the exact future.

Weird, when I told you shills years ago that there’s no way this would be done by 2019 or 2020 you let loose with the personal attacks and sure seemed “smart enough” then.

Maybe knowledge is worth more than blind faith after all.

 

[heymrdj]

It will be fine. Moaners will moan.

Dirty Tesla dude has been very transparent in his videos. The latest one shows that no radar MY has done fine.

More to come...

A clear sunny day. Shill be shillin.

 

[bbhaag]

Yup ready for prime time indeed.......paving the way.

https://www.washingtonpost.com/technology/2021/07/01/tesla-plaid-fire/