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

[unk45]

My first reaction was this wasn´t for real... Never heard about that. But I found several sources including wikipedia and Science journal that make it seem legit:

Science | AAAS

www.science.org

My first awareness of the possibilities came from the Science article. Given the present efforts this does present fascinating possibilities.

Still, we should recall that the very small hydrogen molecule (H2), the smallest one, makes every process of storage, transportation and use full o major challenges, not least the enormous physical size that forces compression techniques that also greatly increase leakage. Those problems, plus the enormous potential for pollution from other molecules that end out with greenhouse gasses, suggest there is major innovation still to come even after the prospecting and extraction problems have been solved.

This is probably nearing the point. or even past the point, that SapceX and Tesla materials gurus can get it work. After all, this is nearing the point of being an engineering problem rather than a pure science one.

Is this ever likely to be better than pure evaluation of solar, wind and storage? I doubt it, just think of that infinitesimally small molecule.

 

[AndrewZ]

38:25 - "The connectivity is not there" - Imo, this is the major factor in this drive's performance.

Yeah either connectivity or necessary GPS accuracy when navigating such dense environments, or connectivity of the GPS in that area

We know HW4 comes with better GPS equipment

 

[Ogre]

BYD factory closure 3 days/week due to rumors being a lack of orders.

https://twitter.com/x/status/1637764631522598914

Whole Chinese auto market is in trouble right now. Apparently legacy auto companies have months of inventory in spite of discounting.

There is one company which is managing to grow. Can't recall which it is. Named after some dead scientist I think.

 

[SOULPEDL]

Yes, heavily congested streets will be the final boss of FSD beta as now it's pretty good in low/median density environments.

While Andrewz enjoys showing FSD fail, lets watch a human driver almost hit someone in real time from the same video.

Good catch, I had to back it up a couple of seconds... Notice the person in the x-walk used that raised-arm signal; this clearly caused the grey car to stop in time, despite the cross-walk warning. /s

Humans

 

[Singuy]

Dense metro areas are where robotaxis make the most sense

My main purpose behind posting that clip was highlighting this particular tester’s risk tolerance. I’ve been following Beta since day 1 and have seen nary a tester take risks like that, letting the vehicle stop in oncoming lanes of traffic without disengaging either before it attempted a bad maneuver or to quickly correct in the middle of a maneuver. And sure enough, hours later another clip comes up with a near accident after letting Beta go ham in an unfamiliar dense traffic environment.

I was just in this thread going on about how it seems hard to believe there could be ~31 Beta accidents on urban streets lol, and these two clips from one person already make it feel more believable if there are more testers taking risks like this.

Considering the media are sniffing for FSD accidents hourly, it's probably more hard to believe they haven't found a good one to report yet out of the 31.

Most fsd beta testers are extremely risk adverse because they don't want to damage their car and are not in it for the views. Using fsd youtubers as some kind of gauge for behavior is selection bias.

 

[Xepa777]

Great Point !
Waymo probably needs to raise $2 to $3 billion a year for the next decade to get to autonomy at scale. Their cash burn will get a lot worse before it gets better.

Tesla on the other hand only needs a million FSD subscribers to generate $2.5 billion a year in cash to pay for their robotaxi program which will be infinitely be cheaper to deploy and more profitable than anything waymo or cruise can do.

Difference is: Waymo is ready to deploy now, while Tesla is ??? years away from deploying. That exact same sentence you wrote was a true statement in 2017. Might still be a true statement in 2027. We don’t know, but based on beta now it’s not close. I would know, I’ve had FSD on my model 3 since 2019.

 

[Huskyf]

Yes, heavily congested streets will be the final boss of FSD beta as now it's pretty good in low/median density environments.

While Andrewz enjoys showing FSD fail, lets watch a human driver almost hit someone in real time from the same video.

That's not the corrected version it's the old one

 

[⚡️ELECTROMAN⚡️]

Dense metro areas are where robotaxis make the most sense

My main purpose behind posting that clip was highlighting this particular tester’s risk tolerance. I’ve been following Beta since day 1 and have seen nary a tester take risks like that, letting the vehicle stop in oncoming lanes of traffic without disengaging either before it attempted a bad maneuver or to quickly correct in the middle of a maneuver. And sure enough, hours later another clip comes up with a near accident after letting Beta go ham in an unfamiliar dense traffic environment.

I was just in this thread going on about how it seems hard to believe there could be ~31 Beta accidents on urban streets lol, and these two clips from one person already make it feel more believable if there are more testers taking risks like this.

Can you guys please take it to another thread? The serious investors here will try it out for themselves and base their investment decisions on it accordingly.

 

[GOVA]

Difference is: Waymo is ready to deploy now, while Tesla is ??? years away from deploying. That exact same sentence you wrote was a true statement in 2017. Might still be a true statement in 2027. We don’t know, but based on beta now it’s not close. I would know, I’ve had FSD on my model 3 since 2019.

Yea. I am ready to deploy now too. The outcome is quite easy to predict in case of Waymo and my deployment.

 

[Singuy]

Difference is: Waymo is ready to deploy now, while Tesla is ??? years away from deploying. That exact same sentence you wrote was a true statement in 2017. Might still be a true statement in 2027. We don’t know, but based on beta now it’s not close. I would know, I’ve had FSD on my model 3 since 2019.

Difference is Tesla de-risked themselves by not going the fast deployment money losing route. If waymo finds their way of robotaxis will never be profitable, that is way way way worst than being stuck as a awesome L2 system. Tesla can fail at autonomy and can still profit off it because at 15k its still a legit product even with today's performance. 15k will only be uncompetitive if another company can do better, but so far there are none.

 

[bpjod]

Other cars seem like they´re standing still ! Not official yet and no conditions for a record, but good to have a race driver give a thumps up:

Fun, wish I could drive like that.

Note that the video is posted by Sebastian Vittel (with an "i"), not Sebastian Vettel (with an "e") the recently retired 4-time Formula one champion.

If you want to drive like that without the risk of expensive vehicle repairs/maintenance and no risk to life or limb, sim-racing in virtual reality is an amazing experience these days. It does require a fairly high-end gaming PC with VR headset and a decent set of racing pedals/wheel (and I recommend hand brake and manual shifter, but my favourite racing is rally racing), but it's so much fun. Happy to respond to PM if some info on getting started is desired.

 

[SOULPEDL]

Difference is: Waymo is ready to deploy now, while Tesla is ??? years away from deploying. That exact same sentence you wrote was a true statement in 2017. Might still be a true statement in 2027. We don’t know, but based on beta now it’s not close. I would know, I’ve had FSD on my model 3 since 2019.

Wait... I swear this just arrived on my inbox. Is Waymo seeking help from the public to analyze their own data and solve FSD? WOT!
I did not click on any links, I'm not sure if it's real as it's so absurd. 10,000 Google Cloud credits if I win.

From: [email protected]
To: Me personally!

We’re excited to announce the launch of our 4th annual Waymo Open Dataset Challenges and the latest expansion of the dataset with new sensor data and labels.

Our 2023 Waymo Open Dataset Challenges include:

• 2D Video Panoptic Segmentation Challenge: Given a panoramic video of camera images, produce a set of panoptic segmentation labels for each pixel in each image, where each object in the scene is tracked across cameras and over time.
• Pose Estimation Challenge: Given one or more lidar range images and the associated camera images, predict 3D key points for pedestrians and cyclists in the scene.
• Motion Prediction Challenge: Predict the future positions of multiple agents in the scene given 1 second of past history. We have made lidar sensor data available as a model input.
• Sim Agents Challenge: Produce sim agent models that control agents in the scene, which will be evaluated against the goal of being human-like. This is the first competition on simulated agents, a fundamental but relatively underexplored topic in autonomous driving, now available to the research community.

First-place winners in each of the four Challenges will receive $10,000 in Google Cloud credits. Additionally, teams with the best performing or noteworthy submissions may be invited to present their work at the Workshop on Autonomous Driving at CVPR in June 2023. The 2023 Waymo Open Dataset Challenges close at 11:59 PM Pacific on May 23, 2023, but the leaderboards will remain open for future submissions. You can find all the details in the Official Waymo Open Dataset Challenge Rules.

In addition to the 2023 Challenges, we are releasing newer versions of the Perception and Motion datasets, as well as introducing a new dataset structure, the Modular dataset. You read more about the dataset updates here and download it here.

The technical achievements in the autonomous driving space would not be possible without the underlying innovative research that helps drive progress forward. Thank you to all of the Waymo Open Dataset users for your invaluable contributions to fields like computer vision, behavior prediction, and machine learning – the AV industry would not be where it is today without your support.

Good luck and we can’t wait to see what you accomplish! As always, please don’t hesitate to reach out to [email protected] with questions or comments.

Sincerely,
The Waymo Open Dataset Team

 

[aronth5]

If the vehicle is operating autonomously then there is no human driver, and any human in the car can watch a video just fine without breaking any laws.

Autonomous vehicles are already legal to operate in a number of US states (and some of the largest population ones too). Currently you can ride in such a vehicle in several cities as a robotaxi, though the only vehicle for sale to consumers with any legally-autonomous capability in the US right now is from Mercedes, and it's a pretty heavily ODD-restricted L3 vehicle whose approval for use will be state by state (as all such approvals are in the US right now- though in most of the states where it's possible to certify there's really nothing required to do so beyond the car maker saying it's autonomous and saying it can follow all traffic laws and the car maker has some huge insurance bond or policy behind the system). But in the very limited circumstances it operates in L3 mode, yes, you can legally watch a video, text, whatever since the car, not the human, is legally driving.

in fact Mercedes calls this out in the operating manual:



It does note availability might vary by location- and that'll be true of using L3 at all too since that's regulated state-by-state in the US.


I'd expect similar if/when Tesla offers true autonomous driving- given there's already a bunch of video and game features in the car.

I am not necessarily disagreeing with you but it should be pointed out that existing driverless cars are in specific geofenced services where the cars are clearly robotaxi's and the local police/state troopers are well aware of their existence. (i.e. Waymo). Plus they have formal regulatory approval.

What I'm looking for the state(s) where the driver (i.e. Tesla owner) is seen by a state trooper/local police watching a video on a random drive. Will they just let the car pass or will the office pull the car over? By state law they should ignore the car but what will actually happen initially? When the public realizes there is actually a solution without risking getting stopped FSD and/or similar solution will generate enormous revenue.

 

[growler23]

"EVEN THOUGH IT’S CARBON-FREE, hydrogen has its faults as an energy source. One kilogram of hydrogen holds as much energy as a gallon of gasoline (just under 4 liters). But at ambient pressures, that same kilogram of hydrogen occupies more space than the drum of a typical concrete mixing truck. Pressurized tanks can hold more but add weight and costs to vehicles. Liquefying hydrogen requires chilling it to –253°C—usually a disqualifying expense."

Full stop.

As a liquid, yes, full stop - a clearly disqualifying expense!!! But the article goes on to discuss compression of the gaseous form, which could be useful. Still lossy due to energy required? Quite. Still leaky due to H2 being the tiniest possible molecule? Yes. But in theory those are engineering challenges that could be solved in certain places and for certain cases.
BTW, no, I am not putting any money into hydrogen as a fuel at any point in the near future. Just wanted to point out the rest of the article that focuses on gaseous H2.
This could ring a bell for our rocket fuel aficionado, @Gigapress - does that article showing "gaseous H2 almost for free" help any of your cost assumptions about renewable rocket fuel? You know, for the SpaceX Model S?

 

[Bet TSLA]

I think its more likely big money realizes that the wide release of FSD beta v11.3.2 over the weekend also just added $0.08 EPS to 2023 Q1.

There has been no "wide release" yet. About 10% so far.

 

[bkp_duke]

Difference is: Waymo is ready to deploy now, while Tesla is ??? years away from deploying. That exact same sentence you wrote was a true statement in 2017. Might still be a true statement in 2027. We don’t know, but based on beta now it’s not close. I would know, I’ve had FSD on my model 3 since 2019.

Ready to deploy now? LoL. To TWO small geo-fenced areas . . .

What a joke.

Oh, and guess which one tops the charts for accident rates . . .

Waymo tops NHTSA’s list for autonomous vehicle crashes

The feds released a first-time report aimed at developing a clearer picture of autonomous vehicle safety.

www.freightwaves.com

 

[growler23]

Wait... this space program?
(Edit: not to throw shade at what you're exploring... it just kinda reminded me of this.)

View attachment 919542

Now that is unfair. Everyone who's watched SpaceX videos in the last few years knows that rockets never explode

 

[Tony73]

"EVEN THOUGH IT’S CARBON-FREE, hydrogen has its faults as an energy source. One kilogram of hydrogen holds as much energy as a gallon of gasoline (just under 4 liters). But at ambient pressures, that same kilogram of hydrogen occupies more space than the drum of a typical concrete mixing truck. Pressurized tanks can hold more but add weight and costs to vehicles. Liquefying hydrogen requires chilling it to –253°C—usually a disqualifying expense."

Full stop.

Hydrogen probably has no long term future in road transportation but it can be incredibly valuable in energy intensive industries like cement and steel and for seasonal storage of energy.

 

[nativewolf]

To work out profit margins we need to consider income and expenses.

On the income side, in a competitive market, all operators will charge roughly the same amount per mile.

Fixed overheads possibly roughly the same, but fleet size determines the effective cost per mile of fixed overheads.

Then we come to the vehicle cost component, which impacts on cost per mile and fleet size.

It is a fairly safe assumption that the "all in" cost to produce a Gen3 Robotaxi is less than $25K, perhaps 20-30% less than $25K.

I doubt that the cost of a Robotaxi for a competitor is much less than $100K, certainly not less than $50K.

The competitors per vehicle cost is 2X-4X what Tesla is paying.

Let's assume electricity and insurance is roughly on par, even though I expect Tesla has a small advantage there.

For things like tires and cleaning I expect Tesla to have good solutions.

My hunch if every operator is charging roughly the same amount per mile, Tesla is making much higher profits than the competitors and will rapidly have a much larger fleet, which then results in a higher level of service. A Robotaxi turning up when you want one and where you want one is all important.

IMO fleet size, vehicle cost, and geographic spread of the service are important, because cost and convenience are important.

You are talking about 2-3k less depreciation per year? I don’t honestly know what cars depreciation schedule is here in the USA.

I don’t see a couple of thousand a year in extra costs being any hindrance. Because if a couple of thousand makes a difference than the operators are running money losing pigs like uber.

I see waymo being able to monetize robotaxi in many different ways. Much easier than others. Lots of data, ad serving, meal suggestions shop etc due to the data store they own. I could see that alone being worthy of a long hard fight.

 

[Bet TSLA]

It does look like Tesla is finally pushing FSD Beta V11 (in the form of 11.3.2) to a wider population.

Looking at TeslaFi, I see it tracking 19000 vehicles. Of these, 5432 are running some version of FSD Beta. That's 28.6%.

At the moment 5246 are on 10.69.25, 10.69.25.1, and 10.69.25.2, mostly on the last of these. A few, 162 so far, have updated to 11.3.2. There are 428 that have the update pending. That's only 3.0% of those on FSD Beta updated so far, with another 7.9% pending.

So it's not a "wide release" yet, at even potentially being only barely over 10% of the potential customers, but maybe by morning we'll see it go wide. It does indicate to me that Tesla is willing to sample a bigger audience than its chosen few, and maybe 11.3.2 will really go wide rather than us waiting for 11.3.3.

Sadly, it's 2022.45.11, so not based off of the latest branch. Meaning that anybody who has updated to anything 2023.xxx is not going to be able to opt in at this point and get the latest FSD Beta. On TeslaFi, that's 64% of all the tracked vehicles.

Note that these numbers are from around 11:52PM PDT, and are subject to rapid change. But they haven't changed much in the last half hour.

And now, a bit more than 10 hours later, I see 364 updated to 11.3.2 and 232 pending. That's 6.7% and 4.3%. So no significant change. Still small numbers. No wide release yet. Maybe later today?

The longer we see no significant increase in pending installs, the more likely it becomes that they've seen some significant issue and that we'll be seeing 11.3.3 before a wide release happens.

I won't be posting about this any more on this thread. The only reason I've brought it up at all is that this release is kind of a big deal and attracting a lot of attention. We probably ought to be aware that (as usual) what's being reported is mostly wrong.