-
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
Knightshade
Well-Known Member
I literally quoted his own expert, James Douma, saying the NNs being used locally on-car now were too large to run in a single node anymore.
For a third time now I'm mentioning that- because it seems like people keep pretending it wasn't presented?
That's in addition to all the further evidence from Green- or simple understanding of what cross node costs you- like the fact it's flat out costly from a compute perspective to split anything across nodes, so the ONLY reason you'd want to do that is you don't have a choice.
Hell, one of the big innovations in Dojo was massively faster interconnects to avoid taking the large "move across chips" hits traditional designs like the FSD computer have on them.
At AI day we had Jim Keller (presumably we can agree HE is an expert on his subject?) telling us they designed not only for redundancy, but so that the FULL stack could do planning on each node independently- then check the two against each other before executing to insure everything was working correctly.
Which seems like a good plan regardless of SAE level- but again is not possible if you can't run the full stack inside a single node.
Indeed. It's certainly, theoretically, possible Tesla will somehow shrink all existing code by at least 50%, and then somehow add a slew of additional, higher complexity, functionality that doesn't exist today- like a complete OEDR among others and do that without taking up ANY added space. That's, technically, if you squint enough, something you can't 100% rule out as possible. It's phenomenally, requires lots of hopium, unlikely, but it's possible.
But he was trying to claim there wasn't evidence they'd run out of single-node compute NOW. Which there's an abundance of evidence for. That I've already cited. Including from his own expert.
The most interesting part of the Q2 deck is :-
Region= Various, Model = Next Gen Platform, Status = In Development.
They previously talking about ramping 2 Gen3 factories at around the same time,
We can make a guess that the Berlin expansion is for Gen3 production, and possibly for Model 3 Highland, though that might come later. We also can't 100% rule out energy storage battery production in Berlin at some future date, if cells were available.
Cells in structural packs seem to need to have steel cans, I think the extra weight of steel cans works best with a 4680 form factor and might be slightly counterproductive with 2170s.
We we try to guess which vehicles were designed for a structural pack, and where a 4680 structural pack is the only option, Cybertruck and Gen3 would be my guess.
So for various reasons, 4680 cell production in Berlin more or less has to happen for this production volume increase and making Gen3 cars to make sense.
It is also interesting to speculate on what is happening with Class B 4680 cells produced at Austin and Fremont, which are not good enough to use in vehicles.
My hunch is that the 4680 production at Sparks is intended for the short range Semi, but also for energy storage products, and Class B 4680 cells from Fremont and Austin might be shipped to Sparks for use in energy storage products.
Tesla also has to find a use for Class B 4680 cells produced in Berlin, IMO shipping them to Sparks or China for use in energy storage products doesn't make sense. An energy storage battery factory somewhere in Europe makes more sense.
Good point, I expect the Austin 4680 ramp to be progressing well no later than Q2 2024. Perhaps some 2170 or 4680 cells will be shipped from the US to Berlin? Or perhaps Berlin will use more Chinese packs.
There were some 2170 cells are Sparks that were used in Megapacks prior to the ramping of Lathrop, if there are Class A cells suitable for use in vehicles, those 2170s can now be shipped to Berlin. As Austin ramps, more Class B 4680 cells suitable for use in Megapacks might be produced and shipped to Sparks. Sparks can probably still make some volume of energy storage batteries. This is a good time to retool Sparks Megapack production to use 4680s.
Region= Various, could imply Mexico + Berlin, but to me the wording implies at least one other location.
Status = In Development. can mean well before the start of factory construction. The first phase of development is selecting the factory site, designing the factory and lodging construction plans for approval. So far Mexico and Berlin are the only Gen3 sites that seem to be in active development.
Last edited:
Xepa777
Banned
James Douma is a charlatan I’ve been saying it here for years, but because hes a former tmc member and he is like the only technical guy who is superbullish on FSD it’s like d blasphemy in the cult of tesla to point out how much of a fraud he is.yes... mere weeks after he'd told us how it was beneficial because it can see in situations vision can not (he specifically cited fog as one example)
The only thing that had changed was tesla announced radar bad, so James changes his story to match them.
This is the appeal to authority fallacy.
Especially weird when, as I cited, he's one of the original sources that proved Tesla had run out of HW3 compute in the first place.
I cited the source for that.
YOU claimed you had a source saying otherwise from him- but have been unable to actually provide it.
I'd love to- except you said you "couldn't remember" which one he made the claim you say he did- so how can I?
If you can't even recall where you heard it, it's entirely possible you also don't correctly remember what you heard.
Of course I do. I posted it.
Hilariously one of the sources was James Doumas own findings.
Since I guess you didn't bother to read it here's the exact words
They HAD TO use the 2nd processor because the NNs had grown too large to run otherwise.
Per YOUR "expert" on the topic.
Of course I have such evidence. Again I previously cited it as being called out by tesla on ai day as one of the architectural goals of the FSD computers entire design
You have Jim Keller on stage stating the intent is for both chips to recieve and process all of the video from the cameras entirely independent from each other....right after Elon explains the whole point is redundancy so one entire node can fail and the car is still fine to drive.
Keller goes on to say the two indepedent chips would then design a plan for driving from the input, then compare those independent plans to insure they match as a sanity check... which again requires full stack running redundantly on each node
All of this is in the roughly 8-10 minute range in the AI day video depending whose copy you use.
You seem deeply determined to ignore all actual evidence and just make up your own narrative that ignores all the available facts.
it'd be great if you could at least do so in the correct place- which is:
AI, Autopilot, & Autonomous/FSD
Discussion about AI, Tesla Bot, Tesla Autopilot (AP), the promise of Full Self Driving (FSD), as well as other Autonomous Vehicles.teslamotorsclub.com
Oh, but a dude who worked on teslas autopilot team 10 years ago said he’s good
Since the subthread regarding FSD - which, as several regulars plaintively were pointing out really didn't belong here - now is becoming toxic in its own right - viz. the @Xepa777 post currently directly above this one - it now formally is closed for further discussion here. Any more posts will be deleted; if it continues to dribble on anyway, Moderators will be issuing Red Tickets.
Todd Burch
14-Year Member
Simple explanation IMO: The stock oversold/overreacted to the ER. Facts:
1. Tesla beat consensus handily.
2. TSLA received a lot of analyst *upgrades* after the earnings call.
Stock dropped big on Thursday due to overreactions and the desire for MMs to bring the stock down to max pain.
Today was simply a recovery from the oversold state and a return to this week's max pain, which is higher than last week.
We don't know if Gen3 cars will be designed for 4680 structural packs. If they are, then the total cost that needs to be summed is, the cost of the final vehicle, especially in the US while the IRA is in play.
While it is possible to produce structural packs with other LFP form factors, we don't know if those packs would be a drop in replacement for 4680s.
We do know that by making 4680 packs with different chemistries, Tesla can balance range and cost considerations.
For energy storage, the IRA provides incentives for locally produced cells, To take full advantage of the IRA it is necessary to ramp energy storage fast.
If Tesla are making LFP cells, I don't know why they would bother with the complexity of a different form factor, when they could simply get one process running well for multiple chemistries.
Tesla overtook Toyota's California sale in Q2 2023 by 1730 units.
Q2 69,212 vs 67,482
YTD 123,482 vs 133,375
I just want to add my $0.02 given the latest developments.
First, I am going to take advantage of the the FSD transfer offer and upgrade my M3 to MY, that's my contribution for Q3.
Had to sell some shares to finance it today as the financing makes very little sense in the short term. And hey... it's only less than 2% of my holdings, so, I think it's ok to splurge a little bit every so often. And yes, I do think the FSD transfer offer is working as intended.
Now, back to TSLA. Given what Elon commented on FSD licensing, I want to go back to my post before back in 2021 (quoted above) and revisit the idea behind my reasoning and see if it makes better sense now. It wasn't the most favorable idea as I got nothing but downvotes, but still.
IMHO, Tesla's superiority in both software and hardware for EVs, and its (ease) manufacturing scale of such, is what gives Tesla the competitive advantage.
I'm going to stick with my original idea... that anyone, be it legacy automakers or Apple or whoever, should they choose to enter the EV scene, Tesla should take advantage of its manufacturing and software prowess and move with it.
With Giga Austin, Berlin and Shanghai all trying to reach the full production capability and Giga Monterey in progress, along with other future Gigas coming, it makes perfect sense for Tesla to branch out and create a different subsidiary that caters nothing but EMS (Electronic/EV Manufacturing Services).
Not only could it spread its expertise in mass-manufacturing, along the fix-cost of building Gigas, but it can also potentially achieve a much higher influence in the direction that the EV industry is moving.
Also, as we enter Gen 3 product line, there would be some CapEX needed as we move from Gen 2 to Gen 3 and Gen 2 becomes obsolete. Should there be additional players willing to use Gen 2 (which quite frankly, it's already state of the art relative to the rest of the industry, it's just Gen3 is better), the cost and timelines of shifting from generation to generation can be prolonged and make it less costly for TSLA.
He's putting out a lot of content (2+ videos per day). I agree he has a broader perspective with good inputs on macro economic data, which I do appreciate.
To each his/her own, but I find him to be too optimistic with his Tesla assumptions/prognostications for my taste. He recently has been going on about why Q2 auto gross margins were going to surprise to the upside (20-21%), and until Elon gave guidance for Q3 production numbers, Randy was trying to convince us that Tesla was going to produce 2.2M vehicles in 2023. He also has been suggesting (along with Herbert Ong) that Tesla has already built many bots which are already working in Tesla factories (which Elon also disputed on the earnings call).
IMHO, it feels like he came on late to the Tesla YT community and is trying to carve out a space as an Uber-bull in order to get more views/clicks. I don't fault him if that's the case. I just prefer a less biased, non-cheerleading message, e.g., Rob Maurer, Dillon Loomis. I stopped watching/listening to Warren Redlich for the same reason. But definitely give him a try if you haven't already.
Artful Dodger
"Neko no me"
↑↑ This is an typo. The true CA sales number is 208 units higher, obviously*. D)
*you can't make this *sugar* up.
Last edited:
Stretch2727
Engineer and Car Nut
This is huge. The number one brand overall in California for Q2 with 4 models. I have been watching this report for a while and did not expect Tesla to overtake Toyota so quickly.
Artful Dodger
"Neko no me"
So lemme get this straight, basically ur saying 'the factory is the product'?
Cheers to the Gig-lets!
My 2 cents. I have seen a few different huge run ups have also followed a few things collapsing. Many times I have seen prices go up/down based on news, events, legal stuff, deals with other companies, reports in media etc where cause and effect is obvious. Sometimes you see a big price movement and later you hear the news, which means that some people knew about it before it became public. Sometimes it just one whale having a major effect. Sometimes there is one single obvious cause of a big change. Sometimes there are causes but they are just very hard to find, some country far away, some obscure new policy, some new technical breakthrough etc.
What I have come to understand is that the price at a given time is just the market trying to find a balance point being buying pressure and selling pressure. The market consists of many actors, each trying to do their best to predict a future that is uncertain and everyone having different beliefs at different times. Many times I have seen bears be very confused about a price point, then a few years later it will be pretty obvious that the past price was actually a very low price given new fundamentals. Bulls just happened to be right about future fundamentals in their guesses that time.
If you see a price moving up/down for no good reason, many times it's just the sentiment of the actors in the market that has shifted their beliefs about the future which is a process of thinking, talking with friends, listening to analysis by other rather than just instant reaction to new data. A few good news, then a month of the market slowly digesting these good news.
I love playing the game of investing, it's an amazing game. You are free to play if you want to, nobody will force you to play it. You can have different beliefs than your opponent and even if their arguments sound nice or they have noble blood, in the end the person who is right will win the game. If you are right you get a reward that is meaningful and if you are wrong you lose something you care about, so you can learn from your mistakes and grow as a person. Play the game knowing that you can lose and don't take it too seriously. If you just focus on playing the game well you will be rewarded in the long run. It a much nicer game than for example debating on the internet or tv, where the person who is wrong often wins... And imo if you think that the market is wrong about something and easily can get skin in the game and still chooses to not get skin in the game, then you either are not as sure as you think you are or you are stupid.
Artful Dodger
"Neko no me"
Today, we witnessed an awesome display of Tesla's (industry leading) cash conversion cycle.
30 days ago on June 23, 2023, Ford CEO Jim Farley lamented 'Cybertruck isn't a real truck' (but Ford builds real trucks for real men, yada-yada...)
Now 1 month later, we witness the 1st product response from the Joint Engineering product placement team at Tesla Motor Vators: the F-150 wrap for Cybertruck!
Now, back to the real issue: the CASH. Indeed, Tesla's multidisciplary team of social scientists and Socal engineers were able to go from receiving a raw unfinished Meme, turn it into a viable (epic) troll play, and ship the deliverable in less than 30 DAZE! A veritable gold meme!
This has tremendous benefits for the CASH conversion cycle (*paging @The Accountant ) since the Tesla team can deliver the finished troll up to 45 days before the company is required to shift funds from Accounts Payable (based on an industry standard 75 day billing cycle).
This memes that Tesla can achieve tremendous (indeed, geomentric) growth at it's troll-farm skunkworks, because they get cash for their trolls before they have to pay for the memes.
This matters.
only partly /sol
P.S. Note that this is just the 1st salvo in the coming "wrap battle". The 2nd exciting product from Tesla's Joint Engineering Team / High-Reliability Organization (J.E.T.H.R.O.) will feature integrated flat-panel LCD displays on all CT exterior surfaces. Combined with Computer Vision, it's traffic-aware generative AGI will troll adjacent traffic or pedestrians based on the appearance of bummy clothes, or the hooptie they're rollin'. External speakers will optionally dis your mother, depending on neighborhood crime heatmaps (update: now only at stop lights, as required by the NHTSA).
Here's an early prototype: (ignore those eco-conscious ride-share passengers, just focus on the side-mounted LCD-racks on the rear of this 2nd Gen FORD truck)
It's The Future, Earl...
30 days ago on June 23, 2023, Ford CEO Jim Farley lamented 'Cybertruck isn't a real truck' (but Ford builds real trucks for real men, yada-yada...)
Now 1 month later, we witness the 1st product response from the Joint Engineering product placement team at Tesla Motor Vators: the F-150 wrap for Cybertruck!
Now, back to the real issue: the CASH. Indeed, Tesla's multidisciplary team of social scientists and Socal engineers were able to go from receiving a raw unfinished Meme, turn it into a viable (epic) troll play, and ship the deliverable in less than 30 DAZE! A veritable gold meme!
This has tremendous benefits for the CASH conversion cycle (*paging @The Accountant ) since the Tesla team can deliver the finished troll up to 45 days before the company is required to shift funds from Accounts Payable (based on an industry standard 75 day billing cycle).
This memes that Tesla can achieve tremendous (indeed, geomentric) growth at it's troll-farm skunkworks, because they get cash for their trolls before they have to pay for the memes.
This matters.
only partly /sol
P.S. Note that this is just the 1st salvo in the coming "wrap battle". The 2nd exciting product from Tesla's Joint Engineering Team / High-Reliability Organization (J.E.T.H.R.O.) will feature integrated flat-panel LCD displays on all CT exterior surfaces. Combined with Computer Vision, it's traffic-aware generative AGI will troll adjacent traffic or pedestrians based on the appearance of bummy clothes, or the hooptie they're rollin'. External speakers will optionally dis your mother, depending on neighborhood crime heatmaps (update: now only at stop lights, as required by the NHTSA).
Here's an early prototype: (ignore those eco-conscious ride-share passengers, just focus on the side-mounted LCD-racks on the rear of this 2nd Gen FORD truck)
It's The Future, Earl...
Last edited:
Not meaning to poop on our Cyberporridge, but… has more than one Cybertruck been produced yet?
There are a few, but the # doesn't matter. The # they will have produced and delivered by December 31 will be interesting but also not real important. What matters are specs, pricing, reviews, and the 2024 ramp-up.
Yes.
At Austin? Also Yes
2 cybertrucks left the factory in a enclosed car carrier (I've never seen this type of car carrier). Source greggertruck
MartinAustin
Active Member
(market closed OT)
Vehicle wrapping is already a thing, but the Cybertruck wrapping business will be huge, it'll scale up the wrapping business and drive innovation. I don't think there will be on-site wrapping at the factory, but there will be Tesla approved wrap shops. Factory-standard wrap designs as well as 3rd-party designs, youtubers with their how-to-DIY videos (the flat panels will make it easy) - so the Cybertruck won't be always thought of as a stainless steel silver vehicle. As well as various types of classic & current artwork, there could be time-based wraps that are switched out after a few days or weeks - political elections, advertising, product placement, jokes, memes, optical illusions, artwork as well as vehicle themes e.g. Batmobile, Warthog, Pink Panther mobile, dragsters, classic cars, other current vehicles (assuming the impersonated don't sue/block), user-submittable designs e.g. for your business, or for your amateur artwork like your Call Of Duty team logo. The Cybertruck will change the business model of paint - and of how you describe your car on official documentation. (do you say it's silver? what colour do cops decide your vehicle is when talking about it over the radio or in documentation? Do you have to update everybody when you change your wrap? etc.) Owners of competitor vehicles will begin to want wraps, the manufacturers will wonder why they are offering so many colours if the buyers simply want to wrap. Perhaps a "primer" option will become the base on some vehicles. (The only downside I can think of is it's a lot of business for 3M, BASF etc. and the oil business, and probably plastic film going into the landfill when owners want to redecorate their vehicles.)
Yet again Tesla will lead the business in a new direction.
Vehicle wrapping is already a thing, but the Cybertruck wrapping business will be huge, it'll scale up the wrapping business and drive innovation. I don't think there will be on-site wrapping at the factory, but there will be Tesla approved wrap shops. Factory-standard wrap designs as well as 3rd-party designs, youtubers with their how-to-DIY videos (the flat panels will make it easy) - so the Cybertruck won't be always thought of as a stainless steel silver vehicle. As well as various types of classic & current artwork, there could be time-based wraps that are switched out after a few days or weeks - political elections, advertising, product placement, jokes, memes, optical illusions, artwork as well as vehicle themes e.g. Batmobile, Warthog, Pink Panther mobile, dragsters, classic cars, other current vehicles (assuming the impersonated don't sue/block), user-submittable designs e.g. for your business, or for your amateur artwork like your Call Of Duty team logo. The Cybertruck will change the business model of paint - and of how you describe your car on official documentation. (do you say it's silver? what colour do cops decide your vehicle is when talking about it over the radio or in documentation? Do you have to update everybody when you change your wrap? etc.) Owners of competitor vehicles will begin to want wraps, the manufacturers will wonder why they are offering so many colours if the buyers simply want to wrap. Perhaps a "primer" option will become the base on some vehicles. (The only downside I can think of is it's a lot of business for 3M, BASF etc. and the oil business, and probably plastic film going into the landfill when owners want to redecorate their vehicles.)
Yet again Tesla will lead the business in a new direction.
Just a question. From the 10-Q.
Tesla has $3.17 billion in deferred revenue attributted mostly to FSD. They state:
If Tesla thinks that FSD will be mainly solved at the end of the year (version 12). Why don’t they expect to recognize much more of the deferred revenue - if not all? There seems to be some inconsistencies here.
Tesla has $3.17 billion in deferred revenue attributted mostly to FSD. They state:
If Tesla thinks that FSD will be mainly solved at the end of the year (version 12). Why don’t they expect to recognize much more of the deferred revenue - if not all? There seems to be some inconsistencies here.
Similar threads
- Locked
- Locked
