Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

This site may earn commission on affiliate links.
Tesla has the stats, we don't. Tesla has to balance amongst many things while still growing 10x and make a profit.
They have been building moar superchargers which get us (Tesla owners) and more brands covered.

Once they figure out that more range offerings make sense, they will do it. Many who want to pay for more range are still a overwhelming minority in the grand scheme of things.

My subjective thought: lately I took another interstate drive and I was looking to get more stops than suggested by Tesla. I am glad I did, 15 minutes extra actually helped to relax a bit and made no difference overall in 6-8 hour drives.

Tesla has been the only EV maker since S60 that knew how much range is needed. No need to pay more and carry extra weight for 90+(?) % of Teslas.
 
As expected, a year without expectations that can be constantly beat to death by the analysts and media leaves a huge, gaping void.

The adage of:
"TSLA goes down on bad news, goes down on good news, and goes up on no news"​

may be in the process of being demonstrated once again.

Those analysts who thrive on FUD stories to plan their easy money moves triggered by emotional reactions among the buyers and sellers of stock which are engineered to move the market have nothing to feed their need for controversy in an environment where Tesla's management has flat-out told them not to expect much this year. But, this won't stop them from looking, just the same.

By taking away the easy targets to place their levers beneath, such as guidance, Tesla has created a data vacuum that may, instead, be filled by reports on their products from independent sources. Sources like those youtubers trying out FSD's latest version, and news releases on Tesla Energy projects, and the pieces reporting on Battery progress in regard to both the manufacturing and supply sides.

Some might even stoop to actually looking at the financials and comparing them to the other players in Auto, Energy, and AI out of sheer boredom.

What is a FUD-based analyst supposed to do with that sort of information? They can't call Tesla a liar if nobody at Tesla was the source. Instead, a ground-swell of positive news is trickling in that may easily grow into occasional flash floods over the coming year.

How do the FUDsters get a lever under that? Ever tried to pry water? Something fluid is much more challenging to divert as it requires something solid to be built in the path of the flow, and FUDsters aren't particularly skilled at building anything solid.

HODL
 
What happened with the S40.
The S40 was _supposedly_ a software limited S60. i vaguely recall there _may_ have been a few sold, (doubtful) but it was perhaps 10+ years ago.

Back then you needed an S85 to make the "jump" between the Delaware, US Interstate 95 supercharger at the Delaware rest stop to the next supercharger in NYC, so the S40 was _very_ short lived sales wise as superchargers were few and far between and 1/2 the S85

this was around the time that a father daughter driving duo did a transcontinental US, New York City to California trip beating Tesla by a week
 
For more information, clickerize upon these here colorized letters and all will be revealed. (and yes, they are sixers)

Awww man... no shipping... but that's awesome...

I need to drink some of that Red IPA while wearing my shirt!:

1708530557174.jpeg
 
Long term: one stack
Short term: don't introduce regressions in Navigate on Autopilot (limited access highway mode).

But wasn't single stack for both introduced a year ago in v11.x?:

Tesla has started a wide rollout of FSD Beta v11.3.2 for testers, which has a single stack for highway and city streets.
 

Sigh...so many "news" articles like this lately.
Smells like desperation to me.
I'm not sure if Toyota's efforts are desperation or stubbornness. On one hand they might feel they waited too long and are too far behind on the EV transition so they need to hope for a miracle breakthrough to allow them to out Tesla Tesla. On the other hand, they might be too stubborn to take their hat in hand and be a follower in the EV transition..

I think it's the later and that's a shame because they have the talent and resources to make it but they are going to squander the opportunity, They are hoping they can milk the hybrids long enough for some new technology to step in and same them. They have such a need to be the leader they are risking the survival of the brand.
 
But wasn't single stack for both introduced a year ago in v11.x?:
Yeah, that was the trasition to v11 single stack.
Now the process repeats with v12.
Once v12 city streets is good enough it will all be single stack again.
(Or I'm off base, which is quite possible)

?
Is the concept of a "stack" still relevant with V12 and the neural net?
Sure, the FSD software can still be called a stack
V11 stack: NN -> C++ planning
V12 stack: NN -> NN planning
 
Yeah, that was the trasition to v11 single stack.
Now the process repeats with v12.
Once v12 city streets is good enough it will all be single stack again.
(Or I'm off base, which is quite possible)


Sure, the FSD software can still be called a stack
V11 stack: NN -> C++ planning
V12 stack: NN -> NN planning

So you think V12 wasn't built on the v11 single stack architecture, but was a parallel code base effort?
 
Yeah, that was the trasition to v11 single stack.
Now the process repeats with v12.
Once v12 city streets is good enough it will all be single stack again.
(Or I'm off base, which is quite possible)


Sure, the FSD software can still be called a stack
V11 stack: NN -> C++ planning
V12 stack: NN -> NN planning

And additionally, the "stack" still has upper level components feeding it, such as cameras, accelerometers, steering angle data, etc... and then lower level cotrol layers like, acceleration profile calculation, brake application, steering control...

The NN layer(s) are sandwichecd in the middle and decipher the upper layer data handed to them, and then pass it on to the lower layers that then figure out how to command the hardware...
 
  • Like
Reactions: AZRI11 and mongo
So you think V12 wasn't built on the v11 single stack architecture, but was a parallel code base effort?
Both, v12 is a branch of v11 and is now developed in parallel
V12 is (at a macro level) V11 with NN planner
The NN planner training breaks NN back compatibility with v11
FSD Beta toggles between V12 and V11 depending where it is
(Again, my impression/ expectation)
 
Agreed.

And just as I think the hub-bub about Tesla "recalls" for fixing a font are annoying, I prefer to not pile on other manufacturers for little stuff like this.

Besides, they give us enough serious fodder that it's not like we are without anything to razz them for...
The thing that really got my goat about the text size recall, was Tesla being required to send out paper mail to owners, notifying us of the "recall", and then telling us there was nothing to do about it because it has already been fixed via an over the air update! What a waste of time and resources. :mad::rolleyes: