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

[Krugerrand]

Maybe you're right. I don't want to antagonize, but I also might lose interest.

100% HODL is freaking boring and it's precisely why dead people do best, on average, so they say. I'm not dead yet.

Thumbs up if I should tone down the day-trading stuff, betting. However, do not confuse this with wanting to blame others for an unfair stock price on days like these. I really thought I was adding to the fun.

Maybe a new hobby? And forget about TSLA for a bit. Bob can teach you how to ride a horse.

 

[GatorMeat]

2024 is the year y'all!! She means it this time!!

GM will be executing on its EV ambitions in 2024, says CEO Mary Barra

General Motors chair and CEO Mary Barra tells Yahoo Finance this is the year of executing on big goals.

finance.yahoo.com

I did see a GMC Sierra Denali EV pickup last night at a hotel with Michigan plates. Looks way better than the Silverado. It isn't on sale until summer, so it must be an employee? I'll have to go back today to see if it's still there.

 

[scaesare]

I'll preface this by saying that I'm not a computer architecture guy and I'm also not an AI guy. So my understanding is quite limited to my old-fashioned computer science training. But here is my best shot at this:

Dojo is competing with NVIDIA by using a different architecture. So there is a potential for Dojo to have a breakthrough that blows the doors off NVIDIA and the other players. Because we are so early in this field, it's possible for Dojo to become the undisputed leader in AI hardware. With such power would come enormous riches. If you can train AI models a lot more efficiently than everyone else, you will win the AI future. Whether that means "Dojo as a service" or just that Tesla has neural nets that vastly outperform everyone else, it boils down to huge profits.

But Elon is right when he tells us that this is a bit of a long shot. There are several players and we don't know who will win. So Dojo is high risk, high reward.

One of the advantages Tesla has as compared to Nvidia is being able to tailor their Dojo/D1 Chip for the specific workload they know they will be running. Much like the FSDv3 chips that that Jim Keller specifically architected for Tesla that primary target INT8 workloads, and thus can build the execution cores, pipelines, caching units, etc... tuned to exactly that workload, and jettison any cruft that would slow it down or consume additional die space or power.

Similarly, the D1 architecture, although supporting more data types that may be needed for training, and to produce the quantized INT8 end-data, can be tuned for a much more specific and narrower workload than a general purpose AI core that many customers might need for a wide variety of applications. Tesla actually introduced their own new data formats specifically for their workload, and built/optimized the D1 for them:

Tesla arithmetic formats and methods for the new 8-bit and 16-bit binary floating-point arithmetic in computer programming environments for deep learning neural network training.

Nvidia, has to try and target good performance on multiple formats, rather than excellent performance on fewer:

The Blackwell core is undoubtedly a nice product (after all Tesla uses em), it's simply that general-purpose will always suffer as compared to tailor-made, based on a per-whatever ($, watt, mm^2, etc...), metric.

 

[Singuy]

Speaking purely theoretically:
If each 9 has 10x the situations which each occur at 1/100 the frequency (1/10 overall occurance), then it would need 10x the compute at each step.
However, it seems unlikely that so many unique cases would be needed.
1. Don't hit things
2. Don't get hit by things
3. Don't break laws (in general) and get to destination

That's only if Tesla is throwing 10 passes at the gpu cluster to find fix the 1 problem.
However they only use the data collected to recall the unique situation and then throw that into the training cluster and then bolt the fix into the NN. At least that's how they make it sound.

 

[scaesare]

I think the F-150 Lightning, specification-wise, looks (or at least looked) like a real contender. But when you see that Ford is subsidizing it to the point of $40k (reported as the overall average loss on each of their EV's - IMO they are likely subsidizing the Lightning even more than this), it must be considered a loser.
Why the sales are dropping, given the massive ICE subsidy Ford is putting into the purchase price - I do wonder this as well.
Could the dealerships be getting tiny profits for moving these, since they sell overall at a loss, so dealers have little incentive to learn and sell them?
Did the Cybertruck "Osborne" the Lightning (I doubt this given the reported lack of knowledge of the CT out there in general public land)?
Did the lack of decent charging options until VERY recently put off folks? Combination of multiple of these? I am curious if anyone has insights on this as well. I expect GM is a lost cause, but I really want Ford to pull through.

Yeah, given that the end-consumer only cares about their cost, and don't care if the manufacturer is subsidizing them, I wonder why the lackluster adoption as well.

One thought is that it may be similar to the EV-1 in that they are being downplayed, de-emphasized, etc... by Ford and their dealerships in that Ford doesn't want to lose $$$ on each one, and dealers want that service $$$ the F-150 won't deliver...

 

[mongo]

That's only if Tesla is throwing 10 passes at the gpu cluster to find fix the 1 problem.
However they only use the data collected to recall the unique situation and then throw that into the training cluster and then bolt the fix into the NN. At least that's how they make it sound.

Right, and in this (quite unlikely) senario there are 10x the situations at each level which then require the training set to grow by 10x. The raw data collection would need to increase by 100x to collect those situations.

 

[Webeevdrivers]

I believe Ford exited 2023 with a ~1,200/week run rate of the Lightnings. They've made a bunch of different scaling back announcements, but I think they're anticipating ~1,600/week this year? Could be less as I can't keep track of what they're scaling back and when.

Oh ok. I saw reference to about 2500 delivered in January which is around 83 per day. I don’t think cybertruck is far from that now.

But January can be slow as well soooo…

 

[Doit]

Could you expand on this? Are you saying Dojo will push Tesla to try to influence how TSMC designs and builds their next fabrication plant?

After watching Karpathy's recent interview I gleaned something I think is very important with regard to Dojo.

He was talking about how horribly inefficient today's GPU's are. Even NVIDIA's best are power hungry monsters that just don't do their jobs very well compared to the low-power human brain at only about 20 watts. It's obvious that when it comes to AI hardware, we are doing it wrong.

Clearly, there is unbelievable potential for improvement.

So the Dojo project can be justified on the basis that AI hardware is in its infancy. It's time for exploration. Like the gold rush, most will fail, but the one who digs in the right spot will be richly rewarded.

There's gold in them thar hills!

/**Follow the fashion of prefacing: I know close to nothing, pure speculation, some maybe just foolish and ignorant, etc etc **/

Dojo, as a startup (among many internal startups of Tesla), is designed to be a fabless chipmaker + data center + hopefully cloud computing. The only thing Dojo won't do and can't do is chip fabrication.

Not that long ago, semiconductor companies made their own chips. When Morris Chang tried to get the initial funding for TSMC, (then) heavyweights such as Intel, Texas Instruments literally laughed him out of the door. Today the manufacturing capability of TSMC vs others is totally flipped, sometimes a telescope is required

Manufacturing is hard, manufacturing of most advanced chips is very very hard, and Dojo manufacturing, from my limited understanding, is another level of hardness. It's not like, "Hey, TSMC, here's our Dojo, just make it." It's a complex process of %^&$+ghh@Nj (something far beyond my comprehension) and it requires sizable commitment (financial, engineering talents...) from TSMC to make it work, PROFITABLLY and COMPETITIVELY, for both TSMC and Tesla. What about yield, considering how complex Dojo tile is? What about scale? How many Dojo tiles are actually going to be made, to be making sense financially, when Nvidia's chips are made in millions per year?

Despite all the prowess of TSMC, manufacturing of Dojo is still a tough nut to crack. TSMC's willingness and ability to commit are not automatic, given Samsung is hot on its heels.

Again, pure speculation, progress of Dojo might have nothing to do with above.


Highly recommended for anyone has half hour to kill (play at 1.5-2x speed) :

Manufacturing @ MIT: TSMC founder Morris Chang on the evolution of the semiconductor industry

 

[Krugerrand]

I wonder if folks waiting for a native NACS port, rather than having to futz with an adapter, is playing in to this at all.

I'm rather disappointed, as the F-150 Light(e)ning felt like it light be a vehicle to helpe segu the average "truck guy" in to EV's. That and Farley seemed to be willing to put (some) company politics aside and cooperate with Tesla on things....

I had hoped it would be a runaway success that would force Ford to pursue EV's more significantly....

So where did they go wrong? Is the pricing the problem? Is there a performance metric problem? Is there a dealership problem? Something else or all of the above?

Perhaps it’s as simple as, Ford is not recognized by its proponents, followers, fans et al as an EV company in America? A Lightening isn’t a Tesla. Selling feature for hardcore fans of Ford and Tesla haters, but anyone serious about an EV, who’s done a bit of research - would they go out on a limb and be an early adopter? Or are a lot of Ford fans waiting a few years to see how it plays out?

Didn’t Ford do a hybrid Focus or some such years ago? I vaguely remember it and something about problems or discontinued or something? Does anyone remember?

 

[SOULPEDL]

Best to keep in mind how your enthusiasm might lead someone to try doing the same without understanding the strategy, and the potential consequences.

I guess this is why "The Wheel" thread was created, so people are at least offered an opportunity to read all the information to first form a strategy. That opportunity for learning rarely happens when individuals post in this thread about short-term trading.

I'm actually quite tempered from what I'm thinking inside. I specifically avoid begging for lower prices and know full well there are people hurting out there. My excitement comes out when I'm hoping for a higher price. Notice that I'm fighting against the 140 thesis after winning 160? I don't even visit the Wheel but I do have one Call for Sept kinda flickering.

This game approach helps me cope, and at the same time I'm toying with the bastards that spew BS, especially when they come on here. New names pop up out of the blue and it's tag-team season all over again. I'm impressed BTW by how this group handles it, some with unusual tact that can also be quite effective.

If you had a system that seemed to be working, was amusing (if not educational), why hide it? People read what they want on any given day. Somehow they might think there's no risk to investing in TSLA, but HODL regardless? I think there could be a bigger risk in being in the market at all right now. This is also quite biased if you ask me, but historically has paid the best returns - we all know Tesla is just like any other company. Meanwhile, the investment odds are quite dynamic and course correction might be prudent.

I see something completely different I guess. It's partially subjective based on this crowd, along with FUD, Volume, and other indicators... all to try to understand how my risk is changing, sometimes daily.

Anyway, I hear ya'll but need to chew on a few things. Thanks for the input.

 

[FastEddieB]

I wonder if I should buy some TLSA soon, or wait.

Probably.

 

[cliff harris]

To maybe help explain the 'will they need more compute' debate in a different way:
I agree that Tesla may not need much more compute, but they may need more data. This is entirely consistent.

For example:

• Tesla now clearly has more than enough video clips of an empty road in decent weather. FSD aces this all the time.
• Tesla may not have enough clips of when a pickup truck loaded with inflatable toy unicorns sheds its load over the road during a snowstorm.

You can collect an absolute TON of data, and throw away all the boring stuff, and just be on the lookout for the increasingly rare situations that still cause disengagements.
Ultimately, the real 'gold' for people developing self driving cars is going to be real world videos clips (including metrics on how the driver responded) of the once-in-a-lifetime crazy things happening on the road. This gives a huge disproportionate advantage to Tesla due to its large number of deployed camera-enabled, connected vehicles. Nobody can catch up. MAYBE byd, in theory. But certainly not soon.

 

[Krugerrand]

I'm actually quite tempered from what I'm thinking inside. I specifically avoid begging for lower prices and know full well there are people hurting out there. My excitement comes out when I'm hoping for a higher price. Notice that I'm fighting against the 140 thesis after winning 160? I don't even visit the Wheel but I do have one Call for Sept kinda flickering.

This game approach helps me cope, and at the same time I'm toying with the bastards that spew BS, especially when they come on here. New names pop up out of the blue and it's tag-team season all over again. I'm impressed BTW by how this group handles it, some with unusual tact that can also be quite effective.

If you had a system that seemed to be working, was amusing (if not educational), why hide it? People read what they want on any given day. Somehow they might think there's no risk to investing in TSLA, but HODL regardless? I think there could be a bigger risk in being in the market at all right now. This is also quite biased if you ask me, but historically has paid the best returns - we all know Tesla is just like any other company. Meanwhile, the investment odds are quite dynamic and course correction might be prudent.

I see something completely different I guess. It's partially subjective based on this crowd, along with FUD, Volume, and other indicators... all to try to understand how my risk is changing, sometimes daily.

Anyway, I hear ya'll but need to chew on a few things. Thanks for the input.

There’s a risk being in the market every single second of every single day because it’s run by crooks with their own agendas. But there’s equal risk keeping your money under your mattress too.

There are no guarantees in life and a lot of people are just trying to get through today.

I feel you’re trying to beat the system called life instead of just playing the hand you were dealt. And what an incredible hand it is.

You know more about a single person and his companies than most people know about themselves. You’re holding aces before the flop and instead of being aggressive you’re trying to slow play, giving everyone else at the table a chance to hit big on the river.

I advise you to stop it. It’s all fun and games until someone slow plays you and simply gets lucky.

I know you’re mostly holding and just playing with a bit. But for what purpose? Go find something else to entertain yourself that doesn’t have you questioning the man and the company you’re seriously betting on. It’s as if you’re looking for a reason to doubt yourself. I suggest Parcheesi.

 

[jerry33]

So where did they go wrong? Is the pricing the problem? Is there a performance metric problem? Is there a dealership problem? Something else or all of the above?

Perhaps it’s as simple as, Ford is not recognized by its proponents, followers, fans et al as an EV company in America? A Lightening isn’t a Tesla. Selling feature for hardcore fans of Ford and Tesla haters, but anyone serious about an EV, who’s done a bit of research - would they go out on a limb and be an early adopter? Or are a lot of Ford fans waiting a few years to see how it plays out?

Didn’t Ford do a hybrid Focus or some such years ago? I vaguely remember it and something about problems or discontinued or something? Does anyone remember?

I'd guess that far fewer people know about the Lightning than know about the Cybertruck. In addition, because it just looks just like every other truck, it doesn't stand out at the dealers (who probably don't want to sell it anyway due to it's fewer service requirements) so there won't be any walk-in interest in it either.

 

[Krugerrand]

I'd guess that far fewer people know about the Lightning than know about the Cybertruck. In addition, because it just looks just like every other truck, it doesn't stand out at the dealers (who probably don't want to sell it anyway due to it's fewer service requirements) so there won't be any walk-in interest in it either.

I have been seeing a number of commercials about the Lightening lately. Maybe that will help?

You make a good point that it looks like every other F150 so nothing to draw people to it on lots. Maybe CT will have F150 fans begin asking about an electric pickup from Ford?

 

[rallykeeper]

At 0%, 5 years = 60 months * $200/ month = $12k.
At 6%, a $12k loan = $232 a month.
Loans payments are required to keep the car, continuing the subscription is optional.

And a weird quirk. If you're in a state with sales tax, paying for FSD upfront will incur sales tax -- which you'll then have to finance.

The $200/month doesn't have sales tax.

 

[2daMoon]

I have been seeing a number of commercials about the Lightening lately. Maybe that will help?

Is "the Lightening" sorta like Global Warming, but about loss of mass instead?

Typical, to get hit with two global catastrophes simultaneously.

And they say lightning doesn't strike twice in the same place!

/S

 

[Krugerrand]

Is "the Lightening" sorta like Global Warming, but about loss of mass instead?

Typical, to get hit with two global catastrophes simultaneously.

And they say lightning doesn't strike twice in the same place!

/S

Spellcheck for the win! Perhaps some subconscious dislike for the company sneaking in as well.

 

[Stretch2727]

And a weird quirk. If you're in a state with sales tax, paying for FSD upfront will incur sales tax -- which you'll then have to finance.

The $200/month doesn't have sales tax.

In New Jersey I pay sales tax on a FSD subscription. If I purchase an EV I do not pay sales tax. Not sure if it is part of the car, but my guess is you would avoid sales on a purchase if FSD is part of the car.

EDIt: Actually I did buy EAP as part of the purchase of a Tesla in 2018 and there was no sales tax on this portion. So pretty sure this has not changed. If I try to add later on the app it is showing sales tax.

Every state is different on sales tax.

 

[Skryll]

It's

So where did they go wrong? Is the pricing the problem? Is there a performance metric problem? Is there a dealership problem? Something else or all of the above?

One anecdotal my contractor told me about his buddy who bought one and complains about it all the time makes me think they sold it too expensive with not enough value and not enough freebees and guidance to excite and turn early ford EV adopters into brand ambassadors like Tesla did. I remember when Tesla service centers would 'take care of you' doing stuff for free all the time in the beginning.

Fords early adopters could have sparked a wave of adoption but instead are now anti ambassadors.

My contractors friend told him: I paid $110k for a truck that needs charging all the time and lost value like crazy and it wasn't worth it. Maybe he doesn't even have an L2 charger in his garage ?

Also there are stories of people having tried powering their home and because of inferior software integration relying on a Bluetooth connection despite being plugged into the car found it to be hit and miss to get it running.

Then again we have this f150l owner joining our Tesla meetups and he is just happy about it especially now with NACS.