Just imagine if Tesla were to add Robotaxi to their existing app on Aug. 8th to give Uber a run for their money! With Tesla Insurance and a ride share function added, it would allow any Tesla owner to participate in the Robotaxi adventure! How is that for a new advertising plan? Just like the Uber phone app, people would add the Tesla app to their phone which would absolutely raise the user's awareness of the Tesla brand. That would actually be the Final Piece to the Robotaxi Puzzle!
This also shows the issue with more models which means that the bucket of bits from each camera is fundamentally different when you have multiple camera positions and camera specifications..
Latest News ( copied from CNBC Web page ) about TESLA Auto pilot: Federal authorities say a “critical safety gap” in Tesla’s Autopilot system contributed to at least 467 collisions, 13 resulting in fatalities. The findings come from a National Highway Traffic Safety Administration analysis of 956 crashes in which Tesla Autopilot was thought to have been in use. Tesla’s Autopilot design has “led to foreseeable misuse and avoidable crashes,” the NHTSA report said.
I think Nvidia suggests that you simulate the environment using their Omniverse Cloud, so you can train on whatever environment you want with as many virtual cars you want, which means that you do not need a fleet of real cars to train the AI system. I think there is value in that idea, as Nvidia can produce quite realistic models of the world, and is of course in line with how you train for example chess engines, where they do not train on real chessboards but use a model of the chess board. Even Tesla uses simulated environments to train. Real world pictures and environment data is something that is of course useful but a) that data is collected by Nvidia and other companies too, not only Tesla, and b) you can in fact simulate most traffic by models.
@@dwylhq874I believe simulation can be more useful. Chess engines didn’t learn watching video of people play the game. It’s the same for just about any other software program of this nature.
Hans, great content as always, but some constructive criticism from a video producer... Popping those half dozen text balloons up on the screen while someone is talking breaks concentration, and is inconsiderate of the speaker because the human brain can't read text and process verbal language at the same time, so either the text won't be read (not likely) or the text will be read and the viewer's attention will be pulled away from the speaker. Best to place those text messages on a screen by themselves at a stopping point within the video. This doesn't apply to very short written corrections to something that was said in error.
We have the best computers driving around on the streets, that are the best of the world! And Elong is probable the human with the most knowledge in manufacturing in the world! And his Digg is the most largest in the world, so I sugg it every day!
Used M3/Y prices will SKYROCKET. Jimmah's analysis NPV of autonomy is well OVER $100k to a fleet owner/personal owner in-network. So, all used will be bid up, up, up approaching the NPV by fleets such as Rental car companies & ride share like Uber. Price cuts by Tesla based upon demand will end...AND REVERT to increases in Price...due to NPV created by autonomy.
I presume most of the value increase would only occur if FSD came by default? Else, we may just see the price of FSD itself rise, or go to a pay per mile model
@@Sthielke None of Teslas currently on the road - and I've bought 3 and currently own 2 - will ever be a Level 5 autonomous vehicle. Current Teslas lack not only the software but also the hardware. Level 3... maybe. But Level 3 is far from full autonomy.
You seem quick to assume that there's a lot of headroom left in the FSD 3 NNP, but given how it was designed back in 2018 when Elon believed it would be a vastly simpler problem, I feel like there's room for skepticism. How do you know the end-to-end NN FSD12 isn't already using a large amount of the NNP's resources, since the code is entirely rebuilt? It's not like the same line of code went from a slow CPU in FSD11 to a hyperfast NNP for v12 - for all we know the FSD12 inference based code could be infinitely more complex than in previous versions, and could require the full brunt of the NNP, with room for only code based optimization.
I didn’t even think of that! All the murder mysteries can be solve easier if there is a Tesla parked near a crime scene! Holy smokes! This is at a different level! It will change the WORLD Indeed !!!
i don't think its going to work that way... Tesla already says its very selective with the data it trains on, which means it probably only keeps data from interventions or other strange activity, because its would take WAYYYY to much digital storage to keep a copy of every Tesla driving at every moment...
These people don’t say much about Waymo other than to say it’s ’too expensive’. They don’t acknowledge the cost of a Waymo is less than what they say a Tesla ‘robotaxi’ could generate. Doesn’t make much sense to me.
after watching elon at earnings it's clear that tesla is toyota, aramco, apple, nvidia and uber combined. fsd is pretty cool but is nothing compared to replacing human labor with humanoid robot labor. all the world's factories will be run on this stuff.
Hardware 3 might have headroom in the “processing” department, but the video QUALITY is Capped due to inferior resolution hardware…without that added resolution, no amount for “processing” will make HW3 able to compete with HW4 in terms of READING SIGNS which will be mandatory for Unsupervised FSD…
@@joelarrarte8486 because the discussion was on HW3 hardware…of course if you upgrade to HW4 hardware then it would no longer be HW3….wild that you needed the explanation…
Hardware and data are overstated. As AI get better each year and semiconductor too. What was very imposible and expensive two year ago is now cheap and easy. Iphones were the firsts, but there is many brand of 100$ phone with similar capabilities.
Hans u made a key mistake. The intense Matrix mul is used in the triaining of AI models. This is what Teslas Dojo chip does as well as GPUs. The inference chip though is just an accelerstor chip that is loaded up with the end result of the training which give the best wieghts and the best model(s) ( which is comprised of the transfer function(s) of the and a graph. The graph is just a list that connects up the NN instances together and the weights. Each instance is a function call to a given tranfer function. Training is an optimation process of the weights, given an NN archtecture. Before the trainig, u need to design the arctecture of the NN(s) to five u the estimared behaviour. The arctecture is a combo of what kind of transfer function each layer in the NN will have. Each Neuron can have different transfer function ie a Relu function or any appropriate non-linear function. Each neuron will have a number of inputs each with a given weight. Ie. Below is a simple Neuron with two kinds of input vectors InA(i) and inB(i), weight vector Wa(i) and Wb(i) and Neuron NN and output O. i is from 0 to X. InA(i)-----Wa(i) - - - N __o InB(i)-----Wb(i) - - -N
The issue with simulating cars is not physics, but the insane complexity of the world, and especially the countless other intelligent agents operating in it. Tesla uses a hybrid approach, their simulations are variations of interesting scenarios from the real world. As for humanoid robots, for them simulation actually works well. Physics simulation is pretty advanced now, and there are even AI models that can do it with much less resources. The usual issue with simulation is that the AI tends to get very sensitive to the exact workings of the simulator, which trips them up in the real world (literally in case of walking). The solution for that has been demonstrated by OpenAI years ago. Adding small random noise to every part of the simulation prevents overfitting, and forces the AI to generalize instead. And robots will work in far simpler environments, and can fail safely, at least at first, that's why simulation is adequate. For example factories are already fully simulated and they are designed to be highly ordered and predictable.
@@andyandreou i dont think thats true..... nobody is stupid enough to think you couldn't monetize FSD. what people think is that FSD will either never work or its far in the future..
@@sportbikeguy9875i'd like to believe this is true but i'm sure there will be people denying that cars can drive themselves even when robotaxis without steering wheels are driving right in front of them. I expect Wallstreet to be remarkably obtuse
Even if perfect FSD drops tomorrow good luck getting the required government approval quickly, also what insurance company will be the first to write a policy covering it!? Years away.
Robotaxis are a great idea and I'm sure Tesla will get there with the AI but, I don't want one and I don't know anybody that does. The market for this type of vehicle seems limited but I look forward to an Uber without the driver.
I remember all to well when computers just started to emerge in business use, my old boss (a seemingly very intelligent business man) said to me... I cannot fathom why anyone would want a pc for home! 😂😂
By the time the general public wakes up to the scope of what Tesla is doing, the stock will be through the roof. The analysts are asleep too, as are the OEM's that watched Tesla evolve over at least a decade. By the time they realize they should have taken action ages ago, they will be out of business.
@@michaelnurse9089 The roads won't cope, with empty taxis added to the mix. They fail miserably when dealing with rush hour as all the cars will be be in the wrong place at the wrong time and useless the rest of the working day. It's a fools errand to believe that they'll be any cheaper than uber. The driver is the cleaner
@@MegaWilderness The empty taxis will park in parking spaces. Rush hour traffic will be identical because the same number of people are travelling. AI can easily solve the wrong place problem. I suspect you are just giving a hot take and have not thought this through or done any investigations.
Tesla is a JOKE. You realize major tech companies spend 5 to 6 TIMES more on R&D than Tesla? Lol. You realize the supercomputer running TikToks algorithm has 3x the procuring power of "Dojo"? Haha. Tesla is a joke.
It's not the amount spent on R&D only but how efficiently you spend that money. Nobody is as frugal or hungry as Tesla was and still is in efficiencies of operations with R&D definitely at the forefront. But hey, you do you.
As an AI practitioner I am somewhat confused by the GPU on the car. I thought inference could only be processed serially. It is a single algorithm in end-to-end systems like V12. I am not saying you are wrong that they are using the GPU, only I would like to know how they do it.
FRAU MARLISA AHL SA BERLIN AKO TANAN ANG GAHIMO OG MARKETING DAX TESLA TANAN NI MAHAL❤ ELON HANTOD NA SA MARS AKONG ANG MAGHIMO NAGPROMISE NAKO TANAN MAHAL❤ELON PANGGAON KA NAKO LAB LAB HONEY DARLING❤❤❤❤❤❤❤
There are other companies getting lots of data too Dimo. Is paying drivers to video driving. Due to Tesla giving out how to do it microsoft and others are working hard to catch up. Now it is a race.
You are not wrong that others are collecting data but the problem is you need lots of data. The others are years if not decades behind Tesla because Tesla gathered data for free from all their customers. That is the 250bn fleet Elon is referring to.
I believe not as much is required with the use of Ai. That is reason why Microsoft moved the driving ai department to Vancouver. Guess where the head office of Dimo is located.
Final piece of the puzzle? All it can do is drive if supervised at all times by a human driver, that’s literally it! All it needs is some dirt on the camera and the car would be absolutely stranded. Surely they should have considered providing a simple lens jet wash when designing the cars if this was their plan.
It's easy to get the first 90% done. The last 10% takes 90% of the development time. I don't see robotaxis being viable for at least 5 years. At least not mainstream. Maybe some trial rollouts.
Are you telling me that Elon Musk is a Genius of a unassessible level? Why is the stock so low then? Also very few are able to discern greatness. Well ? I do, that is the reason I keep piling up Tesla stock. Is that the reason that they say to never bet against Elon if you like to win?
I am old enough to remember 'experts' saying Amazon was just a bookseller who hasn't made a profit in 2001, up 16,000% since. And what is Amazon now? An online retailer? Most of their profit comes from Cloud Computing/AWS 😉
@@grahambrown42 And I am old enough to smell the finest bargain of my entire life in Tesla stock below 500, but taking all in consideration, like the finest ,safest, most efficient, and lowest driving costs ever cars that has ever existed, and we add its trillions $ pregnant pipeline full of exceptional high tech products, and I am 86 years old. One has to be a fool not to load up as much stock as possible for $150.00 and forget it for a few years until its true value will be reflected in their mammoth earning results that surely will materialize.
I am not a Robotaxi expert but a retired aerospace engineer. Robotaxi is dead - stick a fork in it. Why is it dead already? Software. Robotaxis require SAE Level-5 autonomous driving Tesla's FSD (Full Self Driving) is barely Level-3 and it can NEVER pass Level-4 because of horrible design, AI, and cheap cameras and NO LIDAR or RADAR If Elon were to scrap FSD and start over it would be an easy 10 years to get to SAE Level-4 and another 10 years to get to SAE Level-5 Robotaxi is just another futuristic topic Elon will talk about trying to distract you from the piss poor performance he has done with Tesla. With Elon its always an upcoming event that falls flat on its face.
@@Freakyguy666 OK what does that mean? and for you to say I have no credibility does that make you credible? If so please cite your credentials that make you so. You made the charge now back up the charge. I've never claimed anything but my opinion I have opinions on lots of things...
![Six Hundred Strike [EPIC: The Musical] Full Animatic](http://i.ytimg.com/vi/zov6NXIAuow/mqdefault.jpg)
People in the US have no idea how lucky we are to have Tesla. So much greedy hate. This will be as important as the Railroads back in the 1800's
he should move back to South Africa 🇿🇦
Why did you write '$250 TRILLION investment" in the thumbnail? What does that refer to? Just click bait?
Just imagine if Tesla were to add Robotaxi to their existing app on Aug. 8th to give Uber a run for their money! With Tesla Insurance and a ride share function added, it would allow any Tesla owner to participate in the Robotaxi adventure! How is that for a new advertising plan? Just like the Uber phone app, people would add the Tesla app to their phone which would absolutely raise the user's awareness of the Tesla brand. That would actually be the Final Piece to the Robotaxi Puzzle!
I may be wrong but my feeling is Tesla will springboard upwards value wise.
I agree. It’s only a matter of time.
Buddy, it’s gonna happen
I'm hoping it continues to stay low/go lower over the given months as I'll have some more money to drop on stock 😅
You're wrong. Robotaxi will never exist
@@ictrlaltdeleteu welcome bear !🐻
This also shows the issue with more models which means that the bucket of bits from each camera is fundamentally different when you have multiple camera positions and camera specifications..
Latest News ( copied from CNBC Web page ) about TESLA Auto pilot: Federal authorities say a “critical safety gap” in Tesla’s Autopilot system contributed to at least 467 collisions, 13 resulting in fatalities.
The findings come from a National Highway Traffic Safety Administration analysis of 956 crashes in which Tesla Autopilot was thought to have been in use.
Tesla’s Autopilot design has “led to foreseeable misuse and avoidable crashes,” the NHTSA report said.
I think Nvidia suggests that you simulate the environment using their Omniverse Cloud, so you can train on whatever environment you want with as many virtual cars you want, which means that you do not need a fleet of real cars to train the AI system. I think there is value in that idea, as Nvidia can produce quite realistic models of the world, and is of course in line with how you train for example chess engines, where they do not train on real chessboards but use a model of the chess board. Even Tesla uses simulated environments to train. Real world pictures and environment data is something that is of course useful but a) that data is collected by Nvidia and other companies too, not only Tesla, and b) you can in fact simulate most traffic by models.
Virtual training has it's place but it's still a simulation so nowhere near as useful as real world testing.
@@dwylhq874I believe simulation can be more useful. Chess engines didn’t learn watching video of people play the game. It’s the same for just about any other software program of this nature.
Head in butts…well stated (two of our gangsta’s to watch….
Thanks for the insights!
Thanks!
Hans, great content as always, but some constructive criticism from a video producer... Popping those half dozen text balloons up on the screen while someone is talking breaks concentration, and is inconsiderate of the speaker because the human brain can't read text and process verbal language at the same time, so either the text won't be read (not likely) or the text will be read and the viewer's attention will be pulled away from the speaker. Best to place those text messages on a screen by themselves at a stopping point within the video. This doesn't apply to very short written corrections to something that was said in error.
Is Cybertruck Hardware 4 or Hardware 3?
We have the best computers driving around on the streets, that are the best of the world! And Elong is probable the human with the most knowledge in manufacturing in the world! And his Digg is the most largest in the world, so I sugg it every day!
Used M3/Y prices will SKYROCKET. Jimmah's analysis NPV of autonomy is well OVER $100k to a fleet owner/personal owner in-network. So, all used will be bid up, up, up approaching the NPV by fleets such as Rental car companies & ride share like Uber. Price cuts by Tesla based upon demand will end...AND REVERT to increases in Price...due to NPV created by autonomy.
I presume most of the value increase would only occur if FSD came by default? Else, we may just see the price of FSD itself rise, or go to a pay per mile model
Awesome.
Lidar may not be necessary, but cameras are blind in fog. They need to add radar back in
FSD is extremely complicated and requires multimodal high quality sensors
Sadly, NONE of the Teslas on the road today is going to become a robotaxi. That's why Tesla's Robotaxi is yet to be revealed.
This is pure speculation
@@Sthielke Indeed: "Tesla will reveal its robotaxi product on Aug. 8, CEO Elon Musk said in a social media post on X."
@@arthdenton dedicated car, not product
@@Sthielke None of Teslas currently on the road - and I've bought 3 and currently own 2 - will ever be a Level 5 autonomous vehicle. Current Teslas lack not only the software but also the hardware.
Level 3... maybe. But Level 3 is far from full autonomy.
Lots of things are impossible in this world... until they're not.
You seem quick to assume that there's a lot of headroom left in the FSD 3 NNP, but given how it was designed back in 2018 when Elon believed it would be a vastly simpler problem, I feel like there's room for skepticism. How do you know the end-to-end NN FSD12 isn't already using a large amount of the NNP's resources, since the code is entirely rebuilt? It's not like the same line of code went from a slow CPU in FSD11 to a hyperfast NNP for v12 - for all we know the FSD12 inference based code could be infinitely more complex than in previous versions, and could require the full brunt of the NNP, with room for only code based optimization.
Any new videos coming up any time soon?
imagine in the future police go them them with a time, date and coordinates and they search the vods and get 360 videos of any crime near a tesla lol
I didn’t even think of that! All the murder mysteries can be solve easier if there is a Tesla parked near a crime scene! Holy smokes! This is at a different level! It will change the WORLD Indeed !!!
i don't think its going to work that way... Tesla already says its very selective with the data it trains on, which means it probably only keeps data from interventions or other strange activity, because its would take WAYYYY to much digital storage to keep a copy of every Tesla driving at every moment...
Really? You want Teslas used for police surveillance?
What's the tech analysis on Waymo? Are they going to make their version of robotaxi work as well?
These people don’t say much about Waymo other than to say it’s ’too expensive’. They don’t acknowledge the cost of a Waymo is less than what they say a Tesla ‘robotaxi’ could generate. Doesn’t make much sense to me.
The cars need to read traffic signs and avoid potholes to be robotaxi
after watching elon at earnings it's clear that tesla is toyota, aramco, apple, nvidia and uber combined. fsd is pretty cool but is nothing compared to replacing human labor with humanoid robot labor. all the world's factories will be run on this stuff.
Hardware 3 might have headroom in the “processing” department, but the video QUALITY is Capped due to inferior resolution hardware…without that added resolution, no amount for “processing” will make HW3 able to compete with HW4 in terms of READING SIGNS which will be mandatory for Unsupervised FSD…
Ever heard of retrofitting?
@@joelarrarte8486 ya it’s called upgrading from HW3 to HW4…irrelevant to the point…
@Freakyguy666 how is it irrelevant?
@@joelarrarte8486 because the discussion was on HW3 hardware…of course if you upgrade to HW4 hardware then it would no longer be HW3….wild that you needed the explanation…
i guess these guys have never heard of george hotz, and comma ai
Hardware and data are overstated. As AI get better each year and semiconductor too.
What was very imposible and expensive two year ago is now cheap and easy.
Iphones were the firsts, but there is many brand of 100$ phone with similar capabilities.
No matter how much compute you have it’s not going to be able to come up with every edge case on its own.
I am skeptical. This is in the category of a solution looking for a problem. I really doubt that this will ever work in a blizzard in North Dakota.
WHAT HAPPENS IF THE PEDAL GETS STUCK TO THE FLOOR IN A F*****G ROBOTAXI 👀
The brakes will override the accelerator as needed. Just like how automatic emergency braking already works even if driver has accelerator pressed.
Hans u made a key mistake. The intense Matrix mul is used in the triaining of AI models. This is what Teslas Dojo chip does as well as GPUs.
The inference chip though is just an accelerstor chip that is loaded up with the end result of the training which give the best wieghts and the best model(s) ( which is comprised of the transfer function(s) of the and a graph.
The graph is just a list that connects up the NN instances together and the weights.
Each instance is a function call to a given tranfer function.
Training is an optimation process of the weights, given an NN archtecture. Before the trainig, u need to design the arctecture of the NN(s) to five u the estimared behaviour.
The arctecture is a combo of what kind of transfer function each layer in the NN will have. Each Neuron can have different transfer function ie a Relu function or any appropriate non-linear function. Each neuron will have a number of inputs each with a given weight.
Ie. Below is a simple Neuron with two kinds of input vectors InA(i) and inB(i), weight vector Wa(i) and Wb(i) and Neuron NN and output O.
i is from 0 to X.
InA(i)-----Wa(i) - - - N __o
InB(i)-----Wb(i) - - -N
You could have just put a link to the Wikipedia page. 😅
The issue with simulating cars is not physics, but the insane complexity of the world, and especially the countless other intelligent agents operating in it. Tesla uses a hybrid approach, their simulations are variations of interesting scenarios from the real world.
As for humanoid robots, for them simulation actually works well. Physics simulation is pretty advanced now, and there are even AI models that can do it with much less resources. The usual issue with simulation is that the AI tends to get very sensitive to the exact workings of the simulator, which trips them up in the real world (literally in case of walking). The solution for that has been demonstrated by OpenAI years ago. Adding small random noise to every part of the simulation prevents overfitting, and forces the AI to generalize instead.
And robots will work in far simpler environments, and can fail safely, at least at first, that's why simulation is adequate. For example factories are already fully simulated and they are designed to be highly ordered and predictable.
Physics is hard too. What is the friction coefficient of that section of road compared to the other section. What would they each look like on camera?
Until it works it doesn't work. Nasdaq will not honor anything until it works. :(
Until it’s monetized. They won’t even care if it works, only til it’s sold at good margin.
@@andyandreou i dont think thats true..... nobody is stupid enough to think you couldn't monetize FSD. what people think is that FSD will either never work or its far in the future..
@@sportbikeguy9875i'd like to believe this is true but i'm sure there will be people denying that cars can drive themselves even when robotaxis without steering wheels are driving right in front of them. I expect Wallstreet to be remarkably obtuse
Even if perfect FSD drops tomorrow good luck getting the required government approval quickly, also what insurance company will be the first to write a policy covering it!?
Years away.
@@timsmith6815 you havent been paying attention have you? It is already legal now, no laws need to be passed.
I’m just hoping today is a strong earnings
Robotaxis are a great idea and I'm sure Tesla will get there with the AI but, I don't want one and I don't know anybody that does. The market for this type of vehicle seems limited but I look forward to an Uber without the driver.
I remember all to well when computers just started to emerge in business use, my old boss (a seemingly very intelligent business man) said to me...
I cannot fathom why anyone would want a pc for home! 😂😂
TLDR It’s great idea, I don’t want it, but I want it.
Interesting discuss 😊🥰🤣
By the time the general public wakes up to the scope of what Tesla is doing, the stock will be through the roof. The analysts are asleep too, as are the OEM's that watched Tesla evolve over at least a decade. By the time they realize they should have taken action ages ago, they will be out of business.
It’s not happening, not because of the tech, but because of lobbyists representing taxi and Uber
Uber driver will buy Tesla.
Umm Florida says “hold my beer”…
If taxis were such a good idea to serve as public transport there would be far more of them. They would cripple the road systems with congestion
Your basic idea is not incorrect, but see where that thought experiment goes if you drop the price of taxis 10x.
@@michaelnurse9089 The roads won't cope, with empty taxis added to the mix. They fail miserably when dealing with rush hour as all the cars will be be in the wrong place at the wrong time and useless the rest of the working day. It's a fools errand to believe that they'll be any cheaper than uber. The driver is the cleaner
The cheaper transportation is the more people used it and the more traffic you have!Robotaxis wont solve anything and make the problem worse!
@@MegaWilderness The empty taxis will park in parking spaces. Rush hour traffic will be identical because the same number of people are travelling. AI can easily solve the wrong place problem. I suspect you are just giving a hot take and have not thought this through or done any investigations.
@@brunoheggli2888 You need to explain why you say this. In theory Robotaxi's are 10x more efficient than other methods of transport.
Tesla stock is undervalued at 20% below its average. There are things which have scared some stock investors away which is opportunity for us
really? the pe is 33. you need your head checked
Hilarious, said that when ChatGPT released and all the fanboys denied it
Hole in one….again!
Tesla is a JOKE. You realize major tech companies spend 5 to 6 TIMES more on R&D than Tesla? Lol. You realize the supercomputer running TikToks algorithm has 3x the procuring power of "Dojo"? Haha.
Tesla is a joke.
It's not the amount spent on R&D only but how efficiently you spend that money. Nobody is as frugal or hungry as Tesla was and still is in efficiencies of operations with R&D definitely at the forefront. But hey, you do you.
@@joelarrarte8486 lol. EFFICIENTLY??? Blowing a 12 year FSD lead, and still failing is efficient? 🤣🤣🤣🤡🤡🤣
Herbert please, simUlate
As an AI practitioner I am somewhat confused by the GPU on the car. I thought inference could only be processed serially. It is a single algorithm in end-to-end systems like V12. I am not saying you are wrong that they are using the GPU, only I would like to know how they do it.
Clickbait. James is not part of this discussion. Unsubscribe
FRAU MARLISA AHL SA BERLIN AKO TANAN ANG GAHIMO OG MARKETING DAX TESLA TANAN NI MAHAL❤ ELON HANTOD NA SA MARS AKONG ANG MAGHIMO NAGPROMISE NAKO TANAN MAHAL❤ELON PANGGAON KA NAKO LAB LAB HONEY DARLING❤❤❤❤❤❤❤
Tesla used car market will explode!
Not 10 billion ea yr, cumulative. Look at r&d annual.
There are other companies getting lots of data too Dimo. Is paying drivers to video driving. Due to Tesla giving out how to do it microsoft and others are working hard to catch up. Now it is a race.
You are not wrong that others are collecting data but the problem is you need lots of data. The others are years if not decades behind Tesla because Tesla gathered data for free from all their customers. That is the 250bn fleet Elon is referring to.
I believe not as much is required with the use of Ai. That is reason why Microsoft moved the driving ai department to Vancouver. Guess where the head office of Dimo is located.
Robotaxi is not happening. It’s Elon distracting from no model 2. What a bunch of fools Tesla investors have become
Why are you sooooooo smart? We all have so much to learn from you.
Easier to make Model 2 than Robotaxi…yet Tesla decided to shift focus to robotaxi…the question is “why?”
Final piece of the puzzle? All it can do is drive if supervised at all times by a human driver, that’s literally it!
All it needs is some dirt on the camera and the car would be absolutely stranded.
Surely they should have considered providing a simple lens jet wash when designing the cars if this was their plan.
It's easy to get the first 90% done. The last 10% takes 90% of the development time. I don't see robotaxis being viable for at least 5 years. At least not mainstream. Maybe some trial rollouts.
Are you telling me that Elon Musk is a Genius of a unassessible level? Why is the stock so low then? Also very few are able to discern greatness. Well ? I do, that is the reason I keep piling up Tesla stock. Is that the reason that they say to never bet against Elon if you like to win?
I am old enough to remember 'experts' saying Amazon was just a bookseller who hasn't made a profit in 2001, up 16,000% since.
And what is Amazon now? An online retailer?
Most of their profit comes from Cloud Computing/AWS 😉
@@grahambrown42 And I am old enough to smell the finest bargain of my entire life in Tesla stock below 500, but taking all in consideration, like the finest ,safest, most efficient, and lowest driving costs ever cars that has ever existed, and we add its trillions $ pregnant pipeline full of exceptional high tech products, and I am 86 years old. One has to be a fool not to load up as much stock as possible for $150.00 and forget it for a few years until its true value will be reflected in their mammoth earning results that surely will materialize.
@@sat7755Blimey I'm retired but just a youngster 😂
I am not a Robotaxi expert but a retired aerospace engineer.
Robotaxi is dead - stick a fork in it.
Why is it dead already?
Software.
Robotaxis require SAE Level-5 autonomous driving
Tesla's FSD (Full Self Driving) is barely Level-3
and it can NEVER pass Level-4
because of horrible design, AI, and cheap cameras and NO LIDAR or RADAR
If Elon were to scrap FSD and start over it would be an easy 10 years to get to SAE Level-4
and another 10 years to get to SAE Level-5
Robotaxi is just another futuristic topic Elon will talk about
trying to distract you from the piss poor performance he has done with Tesla.
With Elon its always an upcoming event that falls flat on its face.
That’s what you said about the Model y and it ended up being the most sold car on the planet…so you have no credibility…
@@Freakyguy666 Hello - Earth to Freaky Guy - read my 1st sentence.
Good grief..
@@noleftturns read this: you have no credibility.
@@Freakyguy666 OK what does that mean?
and for you to say I have no credibility does that make you credible?
If so please cite your credentials that make you so.
You made the charge now back up the charge.
I've never claimed anything but my opinion
I have opinions on lots of things...
/ˌsimyəˈlāSH(ə)n/ 🤷♂