-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
Have to wonder if this is their response to non-USS vehicles having absolutely garbage obstacle detection at present since restoring the old view has turned out to not be feasible in a reasonable amount of time.
Buckminster
Well-Known Member
Artful Dodger
"Neko no me"
Tesla Autopilot delivers on 3-D view promise:
jeewee3000
Active Member
Let's see how it works first before stating Tesla has "delivered". So far they haven't delivered even on Tesla Vision park assist. The distance measurements are still not accurate/reliable, especially in the front.
Apologies for being a Debby Downer on this but anything Tesla Vision is still unproven and without front bumper camera it appears millions of buyers (between USS removal and introduction of bumper cams (highland, juniper etc)) will just be "unlucky" to have paid for features that'll never work.
Tesla engineer seems to be very proud of his work.
Attachments
Artful Dodger
"Neko no me"
"Tesla Vision" (blog May 2020) is the project lead by Andre Karpathy, and it runs Autopilot today (not just FSD). It is much larger in scope than just a particular new feature for 'park assist'.
Why do you assume Highland and Juniper have (front) bumper cameras? They do not; it's only the Cybertruck has that extra camera, and it's for reasons of geometry, not vision.
Ultra-sonic sensors were removed for cost, just like radar. Have you forgotten why radar was deleted? False positives and phantom braking where the bane of standard Autopilot, and were largely sensor fusion issues (when vision and radar disagree, who do you believe?). As always, Tesla solved it over time, in software. It's their thing, ya' kno...
Apology unnecessary, you just have to look at the larger picture...
Cheers!
jeewee3000
Active Member
Obviously, but Tesla never publicly stated this. They touted it's going to be better in the long run. My backup camera disagrees at the first sight of rain. USS>>>dirty camera.
They should've just decoupled USS from Autopilot/FSD and kept USS for the distance measurements when parking.
I love the Tesla software improvements but you have to admit sometimes they mess up and this is one of those things.
Still waiting.As always, Tesla solved it over time, in software. It's their thing, ya' kno...
Sunlight
Member
As a long-term investor, I have found the Cybertruck reveal to be fascinating. After reading and watching as much as possible, it throws up some thoughts.
Clearly the CT is next step for Tesla and opens the door to future products pursuing a more sophisticated path. The previous models were 'old generation' Tesla!
Under the (controversial) skin, the CT could have been any format and it'd be interesting to hypothesise how the truck would have been received if it had looked conventional.
My first thoughts were, would it have:-
So is the s/s construction which largely shaped the truck, worth the 'disadvantages'?
I see the disadvantages as:-
Do I think that the s/s was a 'good idea'? Yes - as it is different and practical (and saves paint– I wonder how the two costs balance out?). I wonder how the production practicalities work out too. I'm sure that Tesla have diligently assessed all costs but even if the CT ends up being more expensive to build, Tesla feel they can recoup that cost somehow.
I do wonder if the original CT revealed in 2019 wasn't just a tester (with zero substance underneath – wasn't it knocked up in 90 days?) to see the reception it received and if it'd fallen on fallow ground, Tesla would have gone back to the drawing board. The response was so exciting that Tesla then had a find a way to build the bloody thing (“We dug our own grave” comment) – no wonder it took 4 years! And has ended up more costly!
In the final analysis we will just have to see how successful the CT is – sales and practicality. It could end up a 'niche' vehicle (even if it sells in truckloads) for people wanting a big vehicle with lots of carrying/storage capacity but hardly ever used as a truck – a sort of brutalist limo!
Maybe it ends up on every building site/factory as a super useful truck withstanding all the usual bashes and scratches.
Conceivably Tesla could take the guts of the CT in a few years and re-clad it in a 'conventional' shape (as Elon has eluded to often) keeping the CT to fulfill a certain market or quietly letting it fade away.
Get the popcorn.
Clearly the CT is next step for Tesla and opens the door to future products pursuing a more sophisticated path. The previous models were 'old generation' Tesla!
Under the (controversial) skin, the CT could have been any format and it'd be interesting to hypothesise how the truck would have been received if it had looked conventional.
My first thoughts were, would it have:-
- Been better accepted?
- Sold in greater numbers?
- Been cheaper to build?
- More aerodynamic
- Tough s/s skin/construction
- Iconic design
- Possibly lighter
So is the s/s construction which largely shaped the truck, worth the 'disadvantages'?
I see the disadvantages as:-
- Vulnerable and expensive windscreen which may compromise site/work use
- Solar gain resulting in excessively hot interior and extra use of a/c
- Smaller frunk – in general use will the CT suffer in comparison to the F150?
- Quirky looks which may influence sales – which way I'm not sure!
- Rear sloping buttresses which may hinder practicality on site/in use
Do I think that the s/s was a 'good idea'? Yes - as it is different and practical (and saves paint– I wonder how the two costs balance out?). I wonder how the production practicalities work out too. I'm sure that Tesla have diligently assessed all costs but even if the CT ends up being more expensive to build, Tesla feel they can recoup that cost somehow.
I do wonder if the original CT revealed in 2019 wasn't just a tester (with zero substance underneath – wasn't it knocked up in 90 days?) to see the reception it received and if it'd fallen on fallow ground, Tesla would have gone back to the drawing board. The response was so exciting that Tesla then had a find a way to build the bloody thing (“We dug our own grave” comment) – no wonder it took 4 years! And has ended up more costly!
In the final analysis we will just have to see how successful the CT is – sales and practicality. It could end up a 'niche' vehicle (even if it sells in truckloads) for people wanting a big vehicle with lots of carrying/storage capacity but hardly ever used as a truck – a sort of brutalist limo!
Maybe it ends up on every building site/factory as a super useful truck withstanding all the usual bashes and scratches.
Conceivably Tesla could take the guts of the CT in a few years and re-clad it in a 'conventional' shape (as Elon has eluded to often) keeping the CT to fulfill a certain market or quietly letting it fade away.
Get the popcorn.
Just wait until the tradituonal truck fans finish their "5 stages of grief".
"Solving The Money Problem" made a quite funny video of that.
Artful Dodger
"Neko no me"
Uh, what? USS was never coupled to Autopilot, that was radar.
This is where we'll have to agree to disagree. Software is the future, it is necessary, and it is on the critical path to the future. When would you take on this problem? Or how about the switch to 48V? It's the same management issue: accepting technical risk vs. stiffled progress (that's why Detroit has been 12v for 70 yrs).
Tesla plans to put 20M units per year on the road. Over 10 years, that 200 MILLION extra sets of sensors, wiring, and assembly COGS/repair expense. It doesn't take a Form 4 or a 10-Q to know what's the desired outcome.
Tesla did NOT mess this up; they're not finished yet. But on the eve of introducing a new software solution, let's focus instead on how they messed up... lol!
This type of short-term thinking is what makes TSLA such a great buy.
Tesla will likely produce ~150M cars over the next 10 years (2024-2033). If a USS system costs $100 per car, that's $15B in COGS (and that's only if Tesla stops expanding production at 20M units/yr in 2029).
For scale, Giga Shanghai cost about $7B to build and tool. Giga Texas will cost ~$10B by the time current plans are complete (they doubtless add more plans).
So deleting USS is worth about the same as Giga Shanghai + Giga Texas combined.
Let that sink in.
JusRelax
Active Member
Bloomberg says that Ganesh was already no longer with the company about a month ago, so I would guess that this is unrelated.
Artful Dodger
"Neko no me"
Rivian R1T has 300 miles range on its 135 KWh battery, compared to Cybertruck at 340 miles on 122 KWh. And R1T has smaller tires, and a smaller payload.
And it won't matter on Mars (or on the Moon). Cybertruck (SpaceX ed.) is the official Mars Rover. NASA will likely use them on the Moon/Artimus.
How much this that worth in PR? Ask any boomer who watched the Apollo lunar rover drive around on the Moon.
Cheers to the SpaceX Martians! (powered by Tesla)
Last edited:
It's a fun thought experiment, but if Tesla were to produce that many cars, I'm sure they would vertically integrate USS and economies of scale would kick in. Your 15B figure is probably off by 2 or 3 orders of magnitude then...Uh, what? USS was never coupled to Autopilot, that was radar.
This is where we'll have to agree to disagree. Software is the future, it is necessary, and it is on the critical path to the future. When would you take on this problem? Or how about the switch to 48V? It's the same management issue: accepting technical risk vs. stiffled progress (that's why Detroit has been 12v for 70 yrs).
Tesla plans to put 20M units per year on the road. Over 10 years, that 200 MILLION extra sets of sensors, wiring, and assembly COGS/repair expense. It doesn't take a Form 4 or a 10-Q to know what's the desired outcome.
Tesla did NOT mess this up; they're not finished yet. But on the eve of introducing a new software solution, let's focus instead on how they messed up... lol!
This type of short-term thinking is what makes TSLA such a great buy.
Tesla will likely produce ~150M cars over the next 10 years (2024-2033). If a USS system costs $100 per car, that's $15B in COGS (and that's only if Tesla stops expanding production at 20M units/yr in 2029).
For scale, Giga Shanghai cost about $7B to build and tool. Giga Texas will cost ~$10B by the time current plans are complete (they doubtless add more plans).
So deleting USS is worth about the same as Giga Shanghai + Giga Texas combined.
Let that sink in.
(or, more simply: keep the USS now that they don't have an alternative, same-quality solution, and eliminate when they have it).
Is that confirmed to be true? Source?
Btw, what is Artimus? I only know of Artemis
Artful Dodger
"Neko no me"
2 orders of magnitude on a $100 subsystem implies a price of $1 each. 3 orders of magnitude is $0.10 each, for the whole system, including labor.
Yeah, -10% sure, maybe, but not 2-3 orders of magnitude lower. And they still need to do the software. When should they start that?
cliff harris
Member
Whenever anyone goes into a long post about how removing USS and making the cars objectively WORSE, they are always 100% of the time, not from Europe. And no matter how many times this is mentioned, its still dismissed.Uh, what? USS was never coupled to Autopilot, that was radar.
This is where we'll have to agree to disagree. Software is the future, it is necessary, and it is on the critical path to the future. When would you take on this problem? Or how about the switch to 48V? It's the same management issue: accepting technical risk vs. stiffled progress (that's why Detroit has been 12v for 70 yrs).
Tesla plans to put 20M units per year on the road. Over 10 years, that 200 MILLION extra sets of sensors, wiring, and assembly COGS/repair expense. It doesn't take a Form 4 or a 10-Q to know what's the desired outcome.
Tesla did NOT mess this up; they're not finished yet. But on the eve of introducing a new software solution, let's focus instead on how they messed up... lol!
This type of short-term thinking is what makes TSLA such a great buy.
Tesla will likely produce ~150M cars over the next 10 years (2024-2033). If a USS system costs $100 per car, that's $15B in COGS (and that's only if Tesla stops expanding production at 20M units/yr in 2029).
For scale, Giga Shanghai cost about $7B to build and tool. Giga Texas will cost ~$10B by the time current plans are complete (they doubtless add more plans).
So deleting USS is worth about the same as Giga Shanghai + Giga Texas combined.
Let that sink in.
I am well aware that software is the future, I've been progarmming for 43 years, since I was 11. The current Tesla software implementation of park-assist for non-USS cars is UTTER GARBAGE. I have owned a £71k car for a year and its still vastly, laughably inferior to a £20k car when it comes to parking. Tesla support assured me that working parking assist was 'coming soon'. Total lies, and then they stop replying when called out on it. Terrible customer service.
For *investors* to try and pretend that this is a cunning 4D chess move by Tesla is madness.
The USS removal was a dumb, dumb mistake. If Tesla had offered European buyers the choice of a non-USS car for $100 cheaper not a single person here would have chosen it.
Can be at least accept that this was a mistake? I know multiple people who refuse to choose Tesla because of this. How many times to people have to point this out before Elon listens?
Artful Dodger
"Neko no me"
Elon tweeted about the Mars rover shortly after the Cybertruck reveal event 4 yrs ago:
Elon Musk on X: "Tesla Cybertruck (pressurized edition) will be official truck of Mars" /X (Nov 21, 2019)
Summary here:
Tesla Cybertruck on Mars: Elon Musk playfully ponders sending a pickup to the Red Planet | teslarati.com (2019-12-30)
lol, it's his kid brother from another mother. Thx 4 da spiel chk.Btw, what is Artimus? I only know of Artemis
Last edited:
I stand corrected on the orders of magnitude, but I still think Tesla could pull off a 10$ USS...
But, still.
If this is a deal breaker for customers (and as @cliff harris it is for a lot of European ones) why not leave it? At least as an option?
Meanwhile, they can work on the software as much as they want, until they reach parity.
Similar threads
- Locked
- Locked
