That is some serious journalism there - 18 minutes worth to boot.
I don't think his story is unique. I give him credit for being accountable and trying to make the most of his situation.
-
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
I rated this like, but I disagree precisely because the more acceptance of BEV enters the mainstream the better it is for both Tesla and the vision.
Until BEV are, say, >50% of total global vehicle market competition is relevant not due to Tesla competition, but to predominate acceptance of BEV vs ICE.
Many of us still do not understand that. It is almost irresistible to discuss market share of EV, but the only relevance is market share of total vehicle sales by BEV and then Tesla.
For reference I follow Norway, but not so much, say, South Africa or Peru.
How many times need we remind ourselves that Tesla Supercharger buildout drives Tesla market share? So long as that is true we are not even coming close to widespread BEV adoption. Thus, opening Tesla Superchargers to non-Tesla use will advance the mission, and generate income in the process.
Thus competition should be relevant, even if it mostly isn't today.
After reading that all I can is this:
I sure am glad I didn't sell all of my TSLA and put it all into crypto!!!
And generate headlines "Tesla superchargers experience long waits for charging, leaving several EV owners frustrated and upset"
Felt more unsafe than a gas station? Yes I have to plan my charging stops carefully, but I’ve never seen anyone robbing the cashier, had my credit card skimmed, or seen a fuel leak/fire/explosion at a supercharger. Stupid article.
Elon actually said in the recent Ron Barron interview that no one is even close to Tesla in solving generalized autonomy. This statement implies they are years behind not just a few years.
There are a lot of barriers to entry for the competition. First, they need a fleet of it least 200,000 cars that have cameras all around to create the occupancy network. They also need to gather that data and train the network with a quick feedback loop. They need the computer hardware built out for training. They also need to hire the engineers which Tesla has in order to do this.
A competive company also has to figure out how to produce EVs at scale and at similar profit margins to Tesla. Because the winner in the robotaxis space will be the ones that can offer the lowest prices.
It’s clear to me that Tesla will win. This is a winner take most situation here. Just like google search. I highly doubt another company will be able to compete.
Last edited:
I agree that one day, all cars will be self-driving, but the time difference between when Tesla gets there and when the 2nd company gets there is going to be many, many years. For instance, how long until Cruise can scale to Austin? Then another city? How many rides will Cruise offer once Robotaxi starts in San Fran? What will their market share be? Then consider, how long would it take them to pivot and adopt Tesla FSD? I don't think that would be feasible, but will be the only way to stay in business.
It is interesting as I think I have the same information as you, but we have come to very different conclusions.
I look at it more simply as this:
Q: Is it possible to create a generalized FSD solution without enormous capital to iterate on an enormous amount of compute?
A: Hells no
Q: When will Tesla offer FSD/Dojo training as a service?
A: As soon as Tesla not only has enough engineering time to spend turning it into a platform, but also has enough spare compute that they can't better spend it on their own products.
Answer in years: 2035 (or whenever ICE cars cannot be sold anymore in a specific region/country)
Elon knows that only his companies will 'do the right thing' with Tesla technologies. Turning that over to another company does not inherently further the mission.
I think having this on the front page of the Times was bad enough. No interview with a Tesla driver that would have told that woman who got stranded if she was driving a Tesla the navigation would have told her exactly where the next Tesla supercharger was. when she started her trip. How many miles it was. Exactly how to get there and pre condition the battery when she got close to the charger.
Guess who is on the editorial board of the NYT times? Please inform yourself. Getting good information about electric vehicles from the NYT while he sits on the board is about as likely as getting water from a stone.
John M. Broder
John M. Broder joined the editorial board at the start of 2018. He most recently served as deputy climate editor and director of news surveys, with responsibility for overseeing the New York Times Poll and polling analysis.
FSDtester#1
Craves Electrolytes
I think the biggest edge Tesla has is 3 + million vehicles driving around collecting data around the world with the cars equipped with the hardware for fsd.
Someone made a good point about shadow mode. It allows Tesla to get telematics, kinematics, etc from vehicles not using any software upgrades. This data will keep them so far ahead. As we have seen on this thread, the phrase "second place will need a telescope " might be an understatement
!Edit: even if any company had some of Tesla's code, they don't have millions of cars driving everyday to compile the data, let alone do something useful with said data.
Last edited:
Yes no one is even close to solving generalized FSD. That doesn't mean the industry wouldn't catch up the minute it's solved, especially if they all want to buy teslas fsd computer and start submitting specs for camera placements.
I mean it will be a huge blow to competitors like Nvidia and Mobile eye, but not VW and Toyota because their plans was the wait and see approach anyways.
Catch up in a minute?
It takes several years to produce a fleet of 200k cars with the right hardware. Then curate the data to find the corner cases.
They got to be successful poaching from Tesla and hiring the necessary talent. That could take years.
Also, I don’t think it’s a guarantee at all Tesla would sell their FSD computer.
I do think Tesla will sell their FSD hardware and software to other OEMs but then most of the self driving profit goes to Tesla anyways.
Thanks. I have looked him up. Didn't know anything about him. Read about the controversy in 2013. Looked up the board now but didn't see his name.Guess who is on the editorial board of the NYT times? Please inform yourself. Getting good information about electric vehicles from the NYT while he sits on the board is about as likely as getting water from a stone.
John M. Broder
John M. Broder joined the editorial board at the start of 2018. He most recently served as deputy climate editor and director of news surveys, with responsibility for overseeing the New York Times Poll and polling analysis.www.nytimes.com
Read what I wrote. The industry will BUY from Tesla and it'll be a plug and play solution and bring them to Tesla parity overnight(well not overnight because they still have build the cars with Tesla fsd computers and the sensor suites).
Of course Tesla will charge a hefty fee and even profit sharing because Tesla probably mandates all Tesla hardware enabled fsd robotaxis must be on the Tesla network.
Or they can do their own thing and use Dojo which tesla will help them with training which will take years. Either way the winner of this race will dominate, but I don't buy the narrative that other car manufactures won't see FSD in their own cars for like another decade.
bkp_duke
Well-Known Member
After reading that all I can is this:
I sure am glad I didn't sell all of my TSLA and put it all into crypto!!!
Where is that user, bigpeepee or something? Isn't this exactly what he was advocating back in Jan?
I agree about the data, but is there any way to copy or replicate the resulting NN? I mean, data are the means to an end, and the end is a NN. I don't understand for example why all the text2image algos appeared at the same time... very tight race with different teams developing their own NN with similar datasets? or there's more to that?I think the biggest edge Tesla has is 3 + million vehicles driving around collecting data around the world with the cars equipped with the hardware for fsd.
Someone made a good point about shadow mode. It allows Tesla to get telematics, kinematics, etc from vehicles not using any software upgrades. This data will keep them so far ahead. As we have seen on this thread, the phrase "second place will need a telescope " might be an understatement
Edit: even if any company had some of Tesla's code, they don't have millions of cars driving everyday to compile the data, let alone do something useful with said data.
Amazing how two pages from the NYT have different factual information. Who would have guessed?
Tesla could open source the data used to train FSD, then open source the model, then open source the in-car hardware specs and then write a white paper on how they all fit together.
Sure, how long would it then take for a company to build a solution do you think?
powertoold
Active Member
How long will it be before another entity reliably lands orbital rockets?
Similar threads
- Locked
- Locked
