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

[SOULPEDL]

+ Regarding FSD V12 .

outside is the same, but inside ...

.... single end-to-end neural network trained on millions of video clips, replacing over 300k lines of explicit C++ code.


... giant leap ...

Fewer "coders" will be needed now for the same level of improvement. This job appears automated, held back only by compute time. I think they have plenty of data, but who really knows. Maybe they need new data and feedback as a matching set aligned to the current build.

That was my real question on SAY. I was seeking guidance on what to keep commenting (or not) after disengagement. More rain driving or not?

Nothing like taking a survey that seems to go nowhere - this affects human participation in the program. We could help more, but OTOH, revealing trouble spots could become fresh meat for MSM. So here we are, no matter where we go.

 

[Singuy]

So you're telling me I'm gonna have to specify yet another layer of software?

"I'm currently on software version v12.1 (2024.3.32.12), Full Self-Driving (previously known as Beta) Software v12.1.1, FSD 1.0"

I think we've got enough names in the current software menu

It is not uncommon for software in development to have code names prior to mass release of v1.0

If you think about it, those of us testing and in the know understand v12.1.2 or whatever but the common man would be like "why are people calling fsd v12?"

 

[Christine69420]

So why ask car journalists?

Dunno.

But it was the first time I heard any of those acknowledging that the Model Y will be the bestseller next year too.

It was the interesting part for me. No matter how many plushy car launch trips the other car makers have or not sent them on. No matter how many of their drinking buddies happen to work for various car dealer ships and/or importers. No matter how far up on the petrol head scale some of them might be.

 

[jerry33]

I don't have a problem being mistaken if you want to work off of real numbers, here is another fairly reliable source...

“The average electric vehicle has about 2,000 chips, roughly double the average number of chips in a non-electric car,” Raimondo, Commerce Secretary of the US

SOURCE: Chip Shortage Threatens Biden's Electric Vehicle Plans, Commerce Secretary Says

This is, of course, in reference to EVs in particular (as was the Indian article). I am not sure about the number of chips in Teslas in particular. If you have other sources I am open to learning.

My thought here is that Other BEVs are still tied to the supplier chain for programming and chips, so there is a dramatic increase in the amount of chips used because each supplier provides their own and it's up the manufacturer to tie them together. I wonder if the average even includes Tesla?

 

[SmokyPeat]

Damn. On days like today I wonder if the stock is going to $100 before $300.

 

[Musskiah]

Between this Panasonic to add new battery production line at Tesla Gigafactory Nevada and Panasonic's 30GWh battery factory in Kansas that was announced in 2022*, I think you are right, the batteries will be there.

As far as a THIRD NA factory, Panasonic seems to be reconsidering...

[Panasonic CTO Shoichiro Watanabe recently told Bloomberg in an interview that the company will deliver on its promise to “quadruple production capacity by the 2030 fiscal year.” And to make that happen, he said the company won’t need to rely on building new factories or pouring large investments into production plans. Rather, “we will expand battery capacity and improve productivity at the same time,” Watanabe said.]
SOURCE: Panasonic cools on idea for third US plant for Tesla batteries

 

[M|S|M MYP]

Damn. On days weeks like today so far this year, I wonder if the stock is going to $100 before $300.

FTFY

 

[Knightshade]

V12 is not yet mass released, so until retail version comes out, it can be whatever it wants to be.

What is "the retail version"?

It's already released to a small group of non-employees and is still labeled beta.

I'd bet real money it still will be when released to a larger group of non-employees too.

Imo v12 will be released as a different name. Fsd 1.0 perhaps. The back bone of v12 will have no more rewrites. This is how I take what Elon means by v12 not being beta.

Elon has previously given suggestions about when they'd stop labeling their ADAS stuff beta several times in the past--including as early as 2016 when he suggested they'd need 1 billion miles of data/training (they've long long long since passed that)-- none of which turned out to be true. It's certainly POSSIBLE that This Time It's Different, but it's hardly a sure thing.

That said, given beta has no objective legal meaning, I'm unsure why folks even care that deeply about the label being there or not. Heck even the auto-wipers are labeled beta.

 

[DaBooster]

Damn. On days like today I wonder if the stock is going to $100 before $300.

not beyond the realm of possibilities

 

[SOULPEDL]

My thought here is that Other BEVs are still tied to the supplier chain for programming and chips, so there is a dramatic increase in the amount of chips used because each supplier provides their own and it's up the manufacturer to tie them together. I wonder if the average even includes Tesla?

I seem to recall a comment long ago about Tesla's have even more chips than ICE vehicles. And maybe 2,000 was cherry picked from a Cadillac with lots of buttons.

OT - Def'n of a "chip" includes any active peripheral (regulators, motor drivers, D/A, A/D, inertial sensors, mux's, buffers, radios, and many more). I would not include any of the passive components like diodes, R/Cs, LEDs etc... but it's certainly a gray (grey) area. One could argue that a crystal is a chip... because some do come with regulators built in and are powered like any other chip.

 

[cliff harris]

I've been an AI programmer (among other things) in my day job for decades. It used to be a joke in AI when someone asked what was new in AI, the response would be 'ha. Nothing.' It was an arcane field, full of dispirited developers constantly trying new algorithms to try and get AI to seem a bit smart.
Now its suddenly gone bananas, thanks to machine/reinforcement learning managing to beat the world champion at Go, something that was considered impossible when I was in my first job. Ironically my old manager lead the team that did it.
Things have only accelerated since then. Beating Lee sedol at go was step one, and almost seems quaint compared to what AI can do now.

I bought FSD on my model S in 2015. I totally get that its late, real late, and many promises were not met. But I'm not saying this time is different because of wishful thinking. I think, amusingly, that this time REALLY IS different.

 

[SOULPEDL]

Damn. On days like today I wonder if the stock is going to $100 before $300.

Picked up some at 208. Hope for the best!

 

[Knightshade]

I bought FSD on my model S in 2015.

FSD was not a thing you could buy until ~October 2016, on HW2 and newer cars. It was never available for 2015 Model anything which only ever had AP1

All Tesla Cars Being Produced Now Have Full Self-Driving Hardware | Tesla

www.tesla.com

 

[cliff harris]

The terminology may have changed, but it was absolutely being described as self driving.

 

[uscbucsfan]

The terminology may have changed, but it was absolutely being described as self driving.

It was AP.

 

[jamaatta]

I'm really not an expert, but I have understood that the shift to the end-to-end is a big milestone.

"James Douma explains that for self driving cars and other challenging problems for neural nets, the reason that prior attempts to jump to 100% complete AI systems did not work was the training signal was too weak. The subsystems need to be improved with strong training signals and the system for getting good training video data needs to be built and refined. Once each of the components of self driving (planning, control and visual perception) are all very good then the complete system can be improved with weaker training signal data."

Now it's more like computing, less programming..?

James Douma Explains the Challenge and Benefits of Tesla's End to End Neural Network Self Driving | NextBigFuture.com

Businesses want a 100% complete AI systems that can be trained directly with just data. This is what OpenAI was able to do with ChatGPT and large language

www.nextbigfuture.com

 

[Knightshade]

The terminology may have changed, but it was absolutely being described as self driving.

It wasn't though....

Dual Motor Model S and Autopilot | Blog | Tesla Motors

Yesterday, we unveiled the world’s first dual electric motor production car and announced that new safety and autopilot hardware is standard on every new Model S.

web.archive.org

That's the AP1 release announcement from 2014. It explicitly says it is NOT self driving and not intended to be

Model S will be able to steer to stay within a lane, change lanes with the simple tap of a turn signal, and manage speed by reading road signs and using active, traffic aware cruise control. It will take several months for all Autopilot features to be completed and uploaded to the cars.

Our goal with the introduction of this new hardware and software is not to enable driverless cars, which are still years away from becoming a reality. Our system is called Autopilot because it’s similar to systems that pilots use to increase comfort and safety when conditions are clear. Tesla’s Autopilot is a way to relieve drivers of the most boring and potentially dangerous aspects of road travel – but the driver is still responsible for, and ultimately in control of, the car.

FSD didn't come until Oct 2016, on entirely different HW (which itself was found inadequate to the task and requires upgrades of both cameras and driving computer to even get todays L2 ADAS FSDb)

 

[cliff harris]

I stand corrected!

 

[EQC_]

HW4 & HW5 Discussion V12 on HW4 early

FSD V12 (end to end AI) Outside the garden wall

HW4 & HW5 Discussion Investor crazies only

https://twitter.com/x/status/1749303451278786670

For fun, I'll predict the answers to those 6 questions, if asked on the call:

• Question: Given that you moved the start of the next generation compact vehicle production to Austin, has the timeline improved so that we might see next generation platform vehicles in 2025?
  • My predicted answer: this call is really not the place to discuss future products and their timelines. We will have a separate event when the time comes to announce those products.
• Question: When will Tesla start construction on the Giga Nevada expansion and Giga Mexico, and when can we expect each of these to produce their first product such as 4680, Semi, and next generation vehicles?
  • My predicted answer: construction has begun for both -- earthwork and other preparation is ongoing at both sites. As to precise timelines for products from each, that will depend on many variables in both construction, machinery delivery and commissioning, and supplylines.
• Question: Should retail investors be concerned that Elon has stated he is uncomfrotable expanding AI and robotics at Tesla if he doesn't have 25% of voting?
  • My predicted answer:
    • Elon says: It wasn't an ultimatum or threat, and my relationship with the board is good. In context, I was saying that AI and robotics are critical technologies with large potential for both benefits and harms to humanity. I love humanity. There needs to be balance, so that I have enough voting power to guide the projects, but that I can also be over-ruled if enough shareholders disagree or think I am making a mistake. It would be a mistake to let institutional investors with only short-term profits in mind, and not the long term welfare of humanity, gather a controlling stake and push things in the wrong direction.
    • Somebody else from the Tesla board or leadership team: Yeah, Elon's viewpoint here shouldn't be a concern. Media spin without context didn't present the whole story here, and we are all on the same page of doing what is best for both Tesla and humanity.
• Question: How many cybertruck orders are in the queue and when do you anticipate you will be able to fill existing orders?
  • Side note: this question states "orders" not reservations...which is a little awkward since they are really only offering orders for the Foundation series currently, if I remember correctly... I do suspect that the answer will side-step that a bit.
  • My predicted answer: We have well over a million pre-orders for the Cybertruck, we aren't doing any traditional advertising, and we have lines at our showrooms just to view Cybertruck. We are currently only allowing confirmed orders for the very special, limited Foundation series. Even once we ramp, it would take years to serve all of the pre-orders. At this point in time, we are not concerned about the number of orders, but in ramping production as quickly as we can to deliver this truly special product.
• Question: Can we get Tesla Energy volumes reported in the P&D?
  • My predicted answer: I think the time is right for that. Megapack volumes are increasing rapidly, with quasi-infinite demand, well beyond our current production. Yeah, I think we can look into reporting that on P&D.
• Question: What has been the barrier to ramping 4680 cells to the multi-million cells per week rate and when do you expect to get there?
  • Side note: I like this question, and I don't have an obvious answer.
  • Potential predicted answer: With any new technology, prototypes are easy and production is hard. There are many pieces that need to line up, and many things that only become apparent as you try to ramp up production. With 4680, not only are we developing our own cell production technology, but we are also taking steps to optimize and improve with each new generation of cells. Right now, we think we've got the current generation figured out, but we still have to install the production capacity for it. We have one line running now, with four more production lines expected in the coming months, and another four after that. But raw materials and other supply lines will need to ramp up in parallel. We will get there as quick as we can.

 

[totoro722]

So why ask car journalists? Have these particular journalists hit the pavement and spoken to thousands of regular people on the street to get a good understanding of what ‘the people’ think?

I’m just trying to determine why these four are the best people to ask these questions of. Why should we take what these four people say in this regard over someone else perhaps not specifically a car journalist. Indeed, is anyone in the world actually qualified to answer these questions accurately?

I mean, these people don’t even actually know when a facelift Model Y will hit the market. Which means they don’t know anymore than the rest of us. So what does that say about the rest of their information and conclusions. Why when asked that question they didn’t answer, ‘We don’t know. Nobody but those working on it at Tesla would know.’?

How did they come by the conclusion that the refresh Model 3 didn’t take off? Where are their numbers to back that up and given various circumstances that have/could have affected those numbers? When did Tesla give them that data?

Journalists are available, articulate, know a bit about the subject, and know how to help each other out.