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

[azaz]

Opposing advertising is a highly contrarian position. Elon doesn't have a perfect batting record on contrarianism. There is almost certainly low hanging fruit here.

his contrarianism at times comes from hubris

 

[Remus]

I have no strong opinion on advertising either way. But do we know that Musk swear off advertising? How come I vaguely remember he said he said Tesla would consider it sometimes in the future?

 

[Artful Dodger]

This, is a huge potential for Tesla. And we saw evidence that they are working on it. But it's not slam dunk.

The problem is slightly different, computer vision takes cars of "eye" part. But the arm and leg part of the robot are sometimes not straight forward. For example, seeing a cable vs grabbing it and attaching it to the right place are two different things. The precision of the environmental perception and the eye hand coordination are difficult problem.

I remember Tesla filed patent to make the cable semi regid to make the robot's job easier. I take this as an evidence that they are definitely working on this.

Andrew Ng. has a start-up landing.ai in this exact area. They work on the simple stuff like using image recognition for quality control, that doesn't require hand eye coordination. I guess Tesla is more advanced than that

Thankyou for you insightful comments! I think the 2nd gen Tesla chip being designed right now likely addresses movement plannning and control tasks. Last week during Autonomy Day, Elon said a new chip was halfway through the design phase. Planning by NN instead of via hard-coding is the goal of Gen 2.

Without knowing the AI details (but having done undergraduate studies in neuroscience), it seems a useful way to organize the task is to replicate the sensing and movement tasks across 2 NNs in close communication. the same way nature has organized the human cortex. As usual, to understand a complex system, break in into components:

Chokron, Sylvie, and Gordon N. Dutton. "Impact of cerebral visual impairments on motor skills: implications for developmental coordination disorders." Frontiers in psychology 7 (2016): 1471.​

Fig. 1 from this paper provides a planning overview of "The Tree of Vision": (note the two highlighted areas "Route finding in crowded scenes" and "Innate ability to route find")

Also consider Elon's "Neuralink" project. As of 2018, Neuralink is headquartered in San Francisco, sharing an office building with OpenAI, another company co-founded by Musk.

I think his long term goal is the robotic, autonomous exploration of the cosmos.

Cheers!

 

[AcesDealt]

ASP declining is different than gross margin declining. GM is projected to continue increasing, along with production increases. Not sure I draw the same conclusion as you.

That's fine. If you bother to do the math you will change your mind.

 

[Doggydogworld]

I don’t think Waymo cars are able to make left turns. But the tesla is now able, this is significant:
Tesla Autopilot made a 90-degree left turn at an intersection

Please don't feel I'm picking on you, but this is exactly what I'm talking about. Amir Efrati (the anti-Waymo version of Anton Wahlman) once said Waymo struggled with unprotected left hand turns. This somehow morphed into "Waymo can't make left turns" here on TMC. Then a clip shows up with Tesla following a lane line on a protected left turn, and suddenly Tesla is "ahead" of Waymo.

It's pure wishful thinking. Tesla does not even attempt unprotected left turns yet.

Meanwhile, Waymo vans perform unprotected lefts every day. They are cautious when doing so. Perhaps too cautious. Last year Andrew J. Hawkins of The Verge rode in a Waymo van. It made several unprotected left turns, but he talks about one in particular that took a bit longer than he wanted:
Waymo Unprotected Left Turn

Is excess caution a problem? It is if it causes angry drivers behind you to riot. But insufficient caution is an even worse problem. You have to strike a balance, while also provide a smooth, non-jerky ride that doesn't scare your robotaxi customers. Waymo has worked on this and similar problems for years. They have better sensors than Tesla. They have more smart people than Tesla. Can someone tell me why it won't also take Telsa years to solve problems like unprotected left turns?

 

[keydiver]

That's fine. If you bother to do the math you will change your mind.

As they say in the shareholder letter above, Tesla continues to target 25% margins across their car lines. Sounds healthy to me. I think there has been way too much BS circulating that the base Model 3 has little to no profit.

 

[RobStark]

EV Sales: Europe March 2019

 

[defc0n]

Can someone tell me why it won't also take Telsa years to solve problems like unprotected left turns?

With lots of real world training data examples, which tesla has, it should be fairly trivial to train the neural net to handle any common driving case. The only thing to worry about with Tesla’s, or anyones, approach is the uncommon things that occur while driving. Tesla’s vision system, massive amount of training data, and now onboard processing power is a massive advantage.

Learn more about how neutral nets work and specifically how they are trained and you will understand Tesla’s advantage here. It is very very important. This isn’t something you throw man hours at to try and engineer your way through, or some algorithm you need to solve.

Disclaimer: I’m a software engineer

 

[EVNow]

Waymo has worked on this and similar problems for years. They have better sensors than Tesla. They have more smart people than Tesla. Can someone tell me why it won't also take Telsa years to solve problems like unprotected left turns?

LOL. You state as facts what are at best pure speculation. You know "smart" isn't even an objective term.

There are a number of areas in which you can again claim Google has had more time & smarter folks etc. But they haven't always won in the marketplace or produced desirable products. They have had more failures than successes - just like any other tech company with smart people. Why ?

 

[defc0n]

I should say the problem is already solved. Now it is just a matter of feeding it enough data.

 

[Artful Dodger]

...how many trucks would Tesla have coming into and out of GF1? Could Tesla design some sort of hybrid water/car carrier truck, to make the trucks useful in both directions?

We discussed trucking in water to GF1 back on the 2018 Investor's thread. TL;dr way too expensive. Pipeline is only efficient way. Problem is no water rights. 2nd problem after that is no water.

 

[lascavarian]

I placed the order in the Munich Tesla store, where I was promised that Tesla will at a later stage offer to upgrade my car with a tow bar

I wonder if it will be a Tesla product?

 

[Remus]

Thankyou for you insightful comments! I think the 2nd gen Tesla chip being designed right now likely addresses movement plannning and control tasks. Last week during Autonomy Day, Elon said a new chip was halfway through the design phase. Planning by NN instead of via hard-coding is the goal of Gen 2.

Without knowing the AI details (but having done undergraduate studies in neuroscience), it seems a useful way to organize the task is to replicate the sensing and movement tasks across 2 NNs in close communication. the same way nature has organized the human cortex. As usual, to understand a complex system, break in into components:

Chokron, Sylvie, and Gordon N. Dutton. "Impact of cerebral visual impairments on motor skills: implications for developmental coordination disorders." Frontiers in psychology 7 (2016): 1471.​

Fig. 1 from this paper provides a planning overview of "The Tree of Vision": (note the two highlighted areas "Route finding in crowded scenes" and "Innate ability to route find")

View attachment 401824

Also consider Elon's "Neuralink" project. As of 2018, Neuralink is headquartered in San Francisco, sharing an office building with OpenAI, another company co-founded by Musk.

I think his long term goal is the robotic, autonomous exploration of the cosmos.

Cheers!

Uh... I think that NN chip is for driving. It appears to me that using such powerful chips on robots control might be an overkill. Usually for robots to "see" you don't really need 360 vision that reaches far, the complexity should be much lower.

The eye hand coordination part however, I guess the complexity is how to experiment it off line and apply it online, iterate on it without causing production line shutdown for very long. This is more of a task for human mind than computer chips. And the high pressure on production volume won't help unfortunately.

 

[AcesDealt]

As they say in the shareholder letter above, Tesla continues to target 25% margins across their car lines. Sounds healthy to me. I think there has been way too much BS circulating that the base Model 3 has little to no profit.

You just made a strong statement so clearly you've done the math. What is the average cost to build a model 3 in q1, and what is the cost to build the base 35k$ version? What are the levers for further cost reduction and how significant are they?

 

[EVNow]

I should say the problem is already solved. Now it is just a matter of feeding it enough data.

I think what we still don't know is - how fast can they incorporate new scenarios and improve the driving. As Karpathy mentioned, training is expensive. I don't know how much they can automate. If all this needs mostly manual effort, then it becomes a linear progress vs an exponential one.

In Software generally the progress is exponential slower - and I think that is what Google is facing.

 

[Zaxxon]

Holy crap, people. I had a crazy work day and didn't check this thread for 16 hours...

...and you provided 350+ posts in that span, more than a post every 3 minutes.

It's Sunday. Nothing of merit happened today relating to Tesla's stock. The closest is the referral program boost, and that's not exactly market-moving.

Chill out. The poster who earlier today said it's impossible to keep up with reading--much less contributing to--this thread while having a life is correct. Try to self-moderate a bit. If you find yourself coming in for the 6th post of the day on a newsless Sunday, you are [part of] the problem.

 

[Remus]

Waymo has worked on this and similar problems for years. They have better sensors than Tesla. They have more smart people than Tesla.

If you replace Waymo with Microsoft and Tesla with Apple, didn't we hear similar things about this before?

Or Intel vs ARM?

BTW I am a Microsoft fan for decades, which means I am a weirdo.

 

[Tim S]

In my experience with Turo, the owners don't check the cars, but end up blaming the damage from prior renters onto the renters they don't like that didn't cause any damage. Turo is a scam company for scam vehicle owners to charge 4x fees to renters. Furthermore, Turo doesn't offer flexible hours or extensions on rentals, and charges triple fees instead. Don't use Turo. (Use a traditional car rental company, or a new peer to peer company that doesn't encourage scammers.)

I used Turo to rent a Model 3 in Southern California with no problems. The owner checked the car before and after. He included free supercharging. It was about the same price as a rental car, after gas savings.

 

[MP3Mike]

“Tesla’s are cool but I would rather buy a Bolt made in North America then see my money go overseas to an Italian company”. I had to stop and correct him but he then corrected me and told me that no, they are assembled in California but all the parts come from Italy. I was busy, had to leave it at “you might want to research that”

I would have been so all over that. "Have fun driving your Korean powered car, since the drivetrain and electronics are all made by LG in Korea.")

Most people are surprised that a Model 3 contains ~75% US/Canadian/Mexican parts.

 

[Artful Dodger]

The mods need to start deleting posts. This forum is impossible to keep up with and still have a life. I just scrolled through about 50 posts FROM THE SAME MEMBER that all said the same thing... "Elon shouldn't tweet".
I started coming here in December to find Tesla news. I have wasted way too much of my life the last few months reading useless posts (like this one). I'll go back to Electrek, Clean Technica and Teslarati for news. I'll check back in when we're back up to 300 and everyone is in a better mood. Good luck to all the longs.

MODS You seriously need to reconsider the changes made to you policy with the 2019 Investor's thread.

We had 1,250 posts in 24 hrs on thursday. All of which were unusable to Investors because there were 1,250 post on thursday.

KEEP IT ON TOPIC: NOT NUCLEAR PHYSICS IN THE INVESTORS THREAD.

DELETE ruthlessly until people stick to TSLA investment discussion.

Use the BAN hammer when people post the same thing 50 times. TIMEOUT!

We have lost too many good commenters in this cacophony of random kvetching.