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.
Dojo is an alternate hardware platform to Nvidia chips. Tesla can run the same training and autolabling software on either, just at different price/time/energy usages.
Dataset improvements are independent of platform; though faster compute leads to faster itteration and/or larger dataset support.
Yup.

But I believe Tesla outlined the next significant step in training capability and scale would be when Dojo came online, given it's performance as compared to traditional GPU based architecture. And I think the first exapod was supposed to be online about now.

Hence my wondering if the focus on the improved dataset/labeling size in the notes implies we may be seeing some of the results...
 
That's wrong-headed thinking because the original Roadster was a proof-of-concept EV in order to get media publicity and sell the upcoming Model S to original Roadster buyers and their family, friends and co-workers/acquaintances, etc. Not any old publicity would work, it had to be good publicity. If people viewed the Roadster as overly impractical and/or couldn't even get in it for a test drive/ride, it wouldn't have been nearly as effective as it was in getting pre-sales of Model S. Tesla was depending upon those pre-sales in order to go public most profitably. Being seen as the maker of overly impractical cars could have been the difference between reaching critical mass and people losing interest.
Since it's a long weekend I'll indulge. The change in door sill height did not transform the Roadster from a tiny niche targeted 2 seat sports car with new technology that few would buy into something else. Many would say it is still difficult to get in and out of and hardly a practical vehicle for most people. There was plenty of publicity at the time saying that exact thing.
 
I put a disagree here, not because the Thwaites risk is not real, but because the risk is not imminent hence Curt's western Florida solar-powered project will probably long outlast his presently young life.

Once the Thwaites ice shelf collapses there still will not be catastrophic sea rise since the present melting is under the ice shelf. The catastrophe follows the shelf collapse over time as the land borne ice rapidly moves to the sea, so the dual effects of new fresh water in the Antarctic ocean and rise of ground level in Antartica do definitely produce catastrophe worldwide, with this handy mapping device showing the consequences:

It is worthwhile to note that the consequences of such and event would decimate most of the world's most consequential cities. from New York, Washington and Miami to Shanghai, London, the Netherlands, Bangladesh, Kolkata and so much more.

All the TSLA considerations disappear, as does the planet as humans have known it.
'Doomsday' is obviously not alliterative. The real questions are: one, is there anything that can be done to stop that; two, what about Greenland?; three, what about the arctic regions and permafrost melting?; four; how can mankind stop destruction and simultaneously resuscitate the health of our planet?

On a pessimistic day I am convinced Elon Musk may well be the only really influential person who understands we are destroying our planet. Thus on slight less pessimistic days I HODL TSLA and and some others. Surely those will endure well beyond my lifetime. I rather doubt my nephews and nieces will live so well, if they survive. My spouse and I are the only ones who live in edifices above that zone of rise, and our second home is well below the future water level.

Of course all this will not effect any material change in human behavior until the catastrophe strikes.
I put a disagree here, not because the Thwaites risk is not real, but because the risk is not imminent hence Curt's western Florida solar-powered project will probably long outlast his presently young life.

Once the Thwaites ice shelf collapses there still will not be catastrophic sea rise since the present melting is under the ice shelf. The catastrophe follows the shelf collapse over time as the land borne ice rapidly moves to the sea, so the dual effects of new fresh water in the Antarctic ocean and rise of ground level in Antartica do definitely produce catastrophe worldwide, with this handy mapping device showing the consequences:

It is worthwhile to note that the consequences of such and event would decimate most of the world's most consequential cities. from New York, Washington and Miami to Shanghai, London, the Netherlands, Bangladesh, Kolkata and so much more.

All the TSLA considerations disappear, as does the planet as humans have known it.
'Doomsday' is obviously not alliterative. The real questions are: one, is there anything that can be done to stop that; two, what about Greenland?; three, what about the arctic regions and permafrost melting?; four; how can mankind stop destruction and simultaneously resuscitate the health of our planet?

On a pessimistic day I am convinced Elon Musk may well be the only really influential person who understands we are destroying our planet. Thus on slight less pessimistic days I HODL TSLA and and some others. Surely those will endure well beyond my lifetime. I rather doubt my nephews and nieces will live so well, if they survive. My spouse and I are the only ones who live in edifices above that zone of rise, and our second home is well below the future water level.

Of course all this will not effect any material change in human behavior until the catastrophe strikes.
Not nearly enough credit to the tens of thousands of scientists that accumulate data and create the models and work on the underlying science- tedious mind numbing science. Elon added urgency to Tesla but he was not acting in a vacuum. Even that handy model you link to has lifetimes of research behind it.
 
These several major deals all seem conveniently timed for Investor Day, which, with its expanded scope, seems quite likely to show how far TSLA is advancing to match the 1910-1920 era Ford. The difference is that Ford emphasized production efficiency to the near-exclusion of product refinement while Tesla is simultaneously advancing product, 'fueling', production technology and distribution.

The world has many challenges but TSLA clearly is staying well ahead and advancing its lead. Every other competitor that seems similarly integrated, BYD especially as an example, does not establish an 'eco-system' that integrates sales, service and 'fueling'.

All those who see Tesla reaching the limits of growth had the same 'eco-system' myopia that so many have held about Apple. Apple, though, outsources most manufacturing, much of sales and support. Among those of us who have. like me, held one or both of them for a decade or more, it's instructive to note how truly durable such advantages can be.

Consequently, for longer term success I shall continue concentrating more with TSLA, both for financial rewards and as one of the very few capitalistic forces for positive environmental impact.
 
A bit off topic - does anyone have an opinion if this could still be traded? There is a trading halt till tomorrow or announcement (whichever comes first). If I place an order now, would you expect the stock goes up for a couple of days or longer so it is tradable or would I likely pay peak price and not much of an appreciation afterwards?

Unless you know the financials of the company and have done research into them, and maybe financials of this deal, buying their stock is extremely risky. It is a mining junior, low market cap. One offtake deal does not a company make. Did you realize they are positioning themselves as a battery maker?

This ties into the gambling vs investing debate here…
 
Yup.

But I believe Tesla outlined the next significant step in training capability and scale would be when Dojo came online, given it's performance as compared to traditional GPU based architecture. And I think the first exapod was supposed to be online about now.

Hence my wondering if the focus on the improved dataset/labeling size in the notes implies we may be seeing some of the results...
Agree on the Q1 2023 Pod target, but I don't recall the step change comment, when was that stated?
10.69.1 (methodology step change) also had large increases in dataset size. I think the infrastructure changes also contribute to the ability to further expand the datasets.
Upgraded Occupancy Network to use video instead of images from single time step. This temporal context allows the network to be robust to temporary occlusions and enables prediction of occupancy flow. Also, improved ground truth with semantics-driven outlier rejection, hard example mining, and increasing the dataset size by 2.4x.
Blending of compute:
 
Take for example, last week. It doesn't take a lot of brainpower to realize that TSLA stock is going to take a hit with the announcement of a "recall".

This is amusing because that "hit" was due to macro, it had nothing to do with the recall announcement.

So...your "obvious" confidence was...luck.

Screen Shot 2023-02-20 at 6.22.58 AM.png
 
Investing in the market is really sanctioned gambling.

Only if you don’t do any significant research. If you research a company and its market, then the risk can be mitigated to the point where the risk level is comparable to any other major decison you take in life, like buying a house, marrying someone, or making decisions at a company. Every major decision carries uncertainty risk, stock market investing included. But it isn’t gambling if you do a lot of research to mitigate the risk.

As ChatGPT said, true gambling is a completely random outcome.
 
Only if you don’t do any significant research. If you research a company and its market, then the risk can be mitigated to the point where the risk level is comparable to any other major decison you take in life, like buying a house, marrying someone, or making decisions at a company. Every major decision carries uncertainty risk, stock market investing included. But it isn’t gambling if you do a lot of research to mitigate the risk.

As ChatGPT said, true gambling is a completely random outcome.
Such as what the stock price will be next week.
 
This Teslarati article focuses upon a known flaw in the human system. Essentially, "familiarity breeds contempt" is a very real issue. Here, IIHS comments upon the problem of this tendency of people to become complacent when using automated driving aids.

You really cannot depend upon many/most drivers to exercise the level of vigilance needed as the march of 9s strives to reach a functional full autonomy. There will be those instances of tragedy, and uncounted instances of near tragedy, as the data is gathered to accomplish this worthy goal.

On the flip-side, there will still be many, many lives saved which will not be counted, simply because you don't tend to notice when a system is working well.

 
This Teslarati article focuses upon a known flaw in the human system. Essentially, "familiarity breeds contempt" is a very real issue. Here, IIHS comments upon the problem of this tendency of people to become complacent when using automated driving aids.

You really cannot depend upon many/most drivers to exercise the level of vigilance needed as the march of 9s strives to reach a functional full autonomy.

Which begs the question “is it safe for Tesla to beta test fsd on the highways”
 
You really cannot depend upon many/most drivers to exercise the level of vigilance needed as the march of 9s strives to reach a functional full autonomy.

Not sure if you have a vehicle equipped with a cabin camera, but let me tell you that Tesla is doing a pretty good job of enforcing vigilance. I'm afraid (as I should be) to even pick up my phone while FSDb or NOA are active, because it can dole out strikes very quickly. And it's very good at sounding an alarm at me when it catches me staring at the infotainment screen for too long.
 
Not sure if you have a vehicle equipped with a cabin camera, but let me tell you that Tesla is doing a pretty good job of enforcing vigilance. I'm afraid (as I should be) to even pick up my phone while FSDb or NOA are active, because it can dole out strikes very quickly. And it's very good at sounding an alarm at me when it catches me starting at the infotainment screen for too long.

Unfortunately, I'm still ICE-ing about, until my TSLA holding gets to a point that I can transition to BEV. Hoping this will happen before my CT reservation is filled. :)

From what I've read here and elsewhere, Tesla is great at quickly pivoting toward deployment of safeguards as the Beta-Data reveals the need.
 
Last edited:
Former Tesla CEO shares his thoughts on the company

BI in the last few days has come out with a few anti-Elon, Pro-Eberhard hit pieces. If they run out material they dig from their archives and recycle this crap.

Next, they are going to dig deeper and write a series on the life of the (deranged) whistleblower that was shattered by evil Musk.
 
Not sure if you have a vehicle equipped with a cabin camera, but let me tell you that Tesla is doing a pretty good job of enforcing vigilance. I'm afraid (as I should be) to even pick up my phone while FSDb or NOA are active, because it can dole out strikes very quickly. And it's very good at sounding an alarm at me when it catches me staring at the infotainment screen for too long.
I concur. The cabin camera works very well.
 
Hour long audio-only of interview of Tony Seba by Bobby Llewellyn. I haven't watched it yet, but seems to be largely about cost curves, electric vehicle adoption progress


I really love RethinkX presentations with the charts to support the narrative. Tony and RethinkX have done brilliant work.

In an audio-only format this was more difficult to follow as Tony's (English-as-a-second-language) presentation without the graphics to knit the concepts together wasn't as impactful, for me.
 
Which begs the question “is it safe for Tesla to beta test fsd on the highways”
Where do you suggest they beta test FSD then? And by whom? If not in the real world by regular people. 🙄

People are too stupid to live if they can’t pay attention long enough to get from point A to point B with a driver assist program as advanced as FSD *as they’ve been told they MUST do by the very program* never mind bloody common sense.

I don’t actually want you to answer the questions. It’s a ridiculous and entirely circular argument you propose with your question.
 
Agree on the Q1 2023 Pod target, but I don't recall the step change comment, when was that stated?

Hmmm... feel like that was at one of the AI days...

(rummages through transcripts...)

This sounds like it... The "doubling" is what I was thinking of when I said "significant change", which of course doesn't preclude incremental changes to their existing GPU cluster...

TESLA AI DAY 2022 TRANSCRIPT

So after proving our performance on these complex real-world networks, we knew what our first large-scale deployment would target
Our high arithmetic intensity auto-labeling networks
Today that occupies 4,000 GPUs over 72 GPU racks
With our dense computer and our high performance, we expect to provide the same throughput with just 4 dojo cabinets
And these 4 dojo cabinets will be part of our first exapod that we plan to build by quarter one of 2023
This will more than double Tesla's auto-labeling capacity
The first exapod is part of a total of 7 exapods that we plan to build in Palo Alto right here across the wall
And we have a display cabinet from one of these exapods for everyone to look at
6 tiles densely packed on a tray, 54 petaflops of compute, 640 gigabytes of high bandwidth memory with power and host defeated
A lot of compute
And we're building out new versions of all our cluster components and constantly improving our software to hit new limits of scale
We believe that we can get another 10x improvement with our next generation hardware


10.69.1 (methodology step change) also had large increases in dataset size. I think the infrastructure changes also contribute to the ability to further expand the datasets.

Blending of compute:

Yeah, there may be no real dojo-factor in this release, the notes emphasizing the dataset/labeling improvement stuff just jumped out at me and got me wondering...