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

[insaneoctane]

Unsafe, repetitive,and boring....Does this mean that the Tesla Bot will listen to GLJ and the mainstream media for us and then offer a summary?

 

[ZachF]

Ah yes, deposits. Good point.

If they did a demo of a Tesla Bot opening a fridge and delivering a beer to someone on a couch, they will sell millions of them.

Yeah, to the guy in the presentation who asked what the first use case for the robot is: personal assistant. First to the wealthier, then to the middle class.

 

[ZachF]

Unsafe, repetitive,and boring....Does this mean that the Tesla Bot will listen to GLJ and the mainstream media for us and then offer a summary?

I think even Tesla bot might want to self-terminate if stuck with that....

 

[BrownOuttaSpec]

I wonder how many pre-orders applications Tesla received after the presentation, I think that would be a cool number to have announced.

 

[Electroman]

doing so is so high any legacy-fleet retrofit costs if needed would be a rounding error.

.. and remember only those Teslas that the owners have paid for FSD need to be retrofitted. If you haven't bought FSD by that time, then the only option for them is to buy a new Tesla with the latest hardware.

 

[lafrisbee]

"Boring, repetitive, and dangerous"

Yeah?

That is too contrived. There is one aspect of work that is worse than all of those..even the "dangerous" aspect.
DEMEANING/Degrading.
Only one time have I seen a person whose job was to pick up litter not be in a demeaning position..well that and every street walker. (Luke I think you just pointed out why demeaning was not stated....)

 

[SOULPEDL]

The non-scary non-dangerous robot criteria conflict with efficiency of motion. Which turns into battery runtime.

I still recall nightmares about a robot spider on Johnny Quest. Centered mass and very light fast limbs use less energy. The complication of the human wrist is likely to get the right combination of speed and strength - not just strength. This says, sort of, efficiency.

If you parrot Elon "control authority" (means leverage but takes you down different thought paths) thinking, and think about the dynamics of moving weight, Popeye forearms can help balance, they have control authority, but they are energy expensive.

To get good balance, reliability and good control authority with an element of grace that does not scare people - you don't want it quicker or stronger than people. It just needs to be so inexpensive to make that the dangerous robots are priced out of the market by being 10X more expensive.

Here is where that "control authority/leverage" thinking leads:

1. 4680 batteries in the legs and forearms to provide dynamic balance leverage.
2. Simple, relatively stiff ankles where lateral articulation is done by core muscles positioning the knees. The angular displacement of the lower leg translates all the way to the little toe via structural stiffness. There are not many, if any foot muscles. This is a cost, reliability and time decision.
3. Proprioceptors in the knees (Proprioception - Wikipedia) would use some bandwidth.

I think this is one way to a not-scary robot that is so inexpensive the scary ones never become economically viable. I think that is what Elon is looking for. It really needs to exist now and be inexpensive. Accommodation for the inexpensive foot would be via snow shoes and the like. Software would compensate using the knee data and core muscles that position the knee through time. That should solve balance cheaply at first blush.



View attachment 699026

Edit: Would probably have alternative locations for 4680s so that architectural mistakes do not scuttle the effort. A full string can be configured with and without using the arm and calf battery compartments.

Looks to me you're applying for a job at Tesla Robotics. Go for it!

 

[StealthP3D]

I don't think that is the case, and that isn't what was said.

HW4 will always do better, but HW3 can do enough was the message,

HW4 will simply advance the march of the 9's

When HW3 was announced, Elon said FSD could be solved on HW2 but it will faster/easier on HW3. That's why I said if it was not yet solved on HW3 by the time HW4 was released, it would be solved on HW4 first.

 

[lafrisbee]

Last night after the presentation I had to walk the dogs. Where was my Dog-walkerbot?
Elon, git off yer butt...so I don't have to.

Added: I can only imagine what New York City would look like when there is a robot that walks dogs... more people will get dogs so more dog waste, but then the bots will do a much better job of cleaning up after them...
(And in a post I did earlier today I mentioned how one of the major aspects of a job that a TeslaBot will do are the demeaning ones... Dog *sugar* picker-upper... in the rain/snow. Easily a $50k robot, and it will become an amenity of the better residential buildings. A new level of "Dog friendly.")

Another "add".. I wonder when the dog will start loving the bot more than its owner...

 

[larmor]

Remember when there was huge deal made about automating a felt covering attaching to a piece, aka the 'fluffer bot'? With requisite telsa ending FUD of course. Is the tesla bot signalling they are going to solve some of their manufacturing hurdles with dextrous robots? The mantra has been the line is as slow as the slowest part of the assembly line...

Also spacex planning to make 100 starships, which needs to happen soon for the next Mars window...

 

[JRP3]

Last night after the presentation I had to walk the dogs. Where was my Dog-walkerbot?
Elon, git off yer butt...so I don't have to.

Dog-Bot. The best part is no part.

 

[StealthP3D]

I disagree, Robot revenue is possible next year IMO. My point stated earlier, I'm ready to buy one when it comes out, so would everyone here practically. Sight unseen, cash, done it before with the Model 3. Besides, it can sit with us and watch TV until it learns to do something, just like a baby. "Hey Tesla"... know any new tricks?

The "revenue" isn't "revenue" until it's recognized and it can't be recognized until Tesla delivers a robot.

Anyone expecting humanoid robot revenue in 2022 is going to be sorely disappointed. Successful investors can be incredibly bullish and bet on amazing things but those things have to be possible. You are severely under-estimating the time it will take to develop and test a robot and put it into production, even one that can only walk and gurgle like a baby.

 

[TheTalkingMule]

These clowns are gonna stumble back from the Hamptons in September and be able to buy below the S&P inclusion price. Unbelievable.

All retail combined simply doesn't have the share demand to overcome the MMs ability to naked short on behalf of themselves and their hedgie customers.

 

[insaneoctane]

That's the important part. Even if it were a sensor issue. The cost to add a camera to older models would be nothing compared to the value. Same with the computer as you mentioned.

The cost of replacing HW2 to HW2.5 to HW3 to HW4 for FSD owners is easily justified as the cost of developing FSD...Tesla has been putting the best HW they have in production into all vehicles to build the fleet and collect the data they need to perfect FSD. If that means upgrading HW for FSD owners a few times....drop in the bucket IMO.

 

[Marc_K_inNJ]

Hey guys, sorry for bringing this back on topic... looks like AWS 2.0 is confirmed...

Tesla to offer Dojo neural network training as a service in the future: Elon Musk

Tesla’s AI Day promised to be a tour de force of the company’s FSD strategies and most ambitious projects, and it definitely did not disappoint. Elon Musk noted that AI Day was designed to boost recruitment. And considering how well the event has been received by those familiar with the bleeding...

www.teslarati.com

 

[StealthP3D]

Deposits. I never even saw the Model 3 when I signed up. Sure, deposits aren't big revenue, but if we saw 1M reservations stemming from large corporations looking to save labor costs and start the training ASAP, what do you think the stock would do? I'd personally give more than $1K deposit and accept a Beta unit gladly, I'm in Alpha/Beta land now anyway.

True, deposits aren't big revenue, in fact, they are not even small revenue because they aren't revenue until they are recognized.

 

[Singuy]

Deep dive/first impression of Dojo. So mind blowing that they are skeptical but still can't believe their eyes.

TLDR: "Chips are not the unit of scale for Tesla, the 25 chip tiles are. This tile far surpasses anything from Nvidia, Graphcore, Cerebras, Groq, Tenstorrent, SambaNova, or any other AI training geared start up in per unit performance and scale up capabilities.

All of this seems like far out technology, but Tesla claims they already have tiles running at 2 GHz on real AI networks in their labs.

Rolling it all up, cost equivalent versus Nvidia GPU, Tesla claims they can achieve 4x the performance, 1.3x higher performance per watt, and 5x smaller footprint. Tesla has a TCO advantage that is nearly an order magnitude better than an Nvidia AI solution. If their claims are true, Tesla has 1 upped everyone in the AI hardware and software field."

Tesla Dojo – Unique Packaging and Chip Design Allow An Order Magnitude Advantage Over Competing AI Hardware

Tesla has flipped the hardware world on its head by creating a chip with GPU levels of compute, more than the highest end networking chips IO, and high levels of flexibility. They utilize a integra…

semianalysis.com

 

[M|S|M MYP]

Just to confirm....Max Pain is @$680 today right?

 

[Singuy]

As predicted, the street considers any kind of android announcement to be a distraction for the company.

 

[phantasms]

You're welcome.

https://twitter.com/x/status/1428765860173336582