Mobileye CEO sayd FSD companies will need to consolidate

[diplomat33]

"Shashua believes that the autonomous car industry will need to consolidate because it's too difficult for separate companies to cooperate in developing self-driving vehicles, with each competing to develop their own proprietary technology, Reuters reports.

The CEO of Mobileye said the strategy of having individual companies each focusing on one component, like, cameras, sensors or HD maps, is not sustainable because it's an "end-to-end system" that cannot be broken down.

In addition to cameras, future autonomous vehicles will rely on AI, machine learning and sensor fusion to navigate. For a vehicle capable of autonomous driving, all of these systems need to be integrated.

Shashua had said in the past that in order for autonomous vehicles to make the best driving decisions, they will need to process data from a combination of cameras, high-definition maps, radar and lidar. Right now many of these technologies are being developed separately.

"It's a formidable task, and there are going to be very very few actors who can go from silicon (chips) to self-driving systems," Shashua said during an online conference by the Israel government-backed EcoMotion.

"Therefore, what we see in the industry and what will continue in the industry is a great consolidation," he added."

CEO of Intel-Owned Mobileye Expects a Big Consolidation of Autonomous Driving Technology - FutureCar.com - via @FutureCar_Media

I feel like this goes along with my prediction that eventually, we will see a consolidation into just 2-3 big providers of autonomous driving.

 

[Silicon Desert]

Interesting CEO comments. So can I interpret that to mean his company can't do it alone, so he thinks other companies can't do it either? Maybe I'm reading too much into that statement

 

[Tam]

...too difficult...

He sounds like L-5 is very far away so he needs "great consolidation".

I think if everyone is thinking and doing the same thing, then consolidation is great.

However, doing so might discourage that there are potentially different ways of solving the same problem and the current ways might be wrong or inefficient.

 

[diplomat33]

Interesting CEO comments. So can I interpret that to mean his company can't do it alone, so he thinks other companies can't do it either? Maybe I'm reading too much into that statement

I am not sure he is saying that Mobileye can't do autonomous driving. But Mobileye is not an automaker and does not manufacture sensors. So even if Mobileye achieves the software for full autonomous driving, they will still need to partner with someone to deploy autonomous cars.

His main argument is that autonomous driving requires an end-to-end solution that combines many different technologies. Therefore, companies that are focused on different components, will need to join forces in order to deliver viable autonomous driving. Only companies that are focused on that end-to-end solution, will be in a good position to deliver full autonomous driving.

 

[ItsNotAboutTheMoney]

"Shashua believes that the autonomous car industry will need to consolidate because it's too difficult for separate companies to cooperate in developing self-driving vehicles, with each competing to develop their own proprietary technology, Reuters reports.

The CEO of Mobileye said the strategy of having individual companies each focusing on one component, like, cameras, sensors or HD maps, is not sustainable because it's an "end-to-end system" that cannot be broken down.

In addition to cameras, future autonomous vehicles will rely on AI, machine learning and sensor fusion to navigate. For a vehicle capable of autonomous driving, all of these systems need to be integrated.

Shashua had said in the past that in order for autonomous vehicles to make the best driving decisions, they will need to process data from a combination of cameras, high-definition maps, radar and lidar. Right now many of these technologies are being developed separately.

"It's a formidable task, and there are going to be very very few actors who can go from silicon (chips) to self-driving systems," Shashua said during an online conference by the Israel government-backed EcoMotion.

"Therefore, what we see in the industry and what will continue in the industry is a great consolidation," he added."

CEO of Intel-Owned Mobileye Expects a Big Consolidation of Autonomous Driving Technology - FutureCar.com - via @FutureCar_Media

I feel like this goes along with my prediction that eventually, we will see a consolidation into just 2-3 big providers of autonomous driving.

Predicting consolidation isn't difficult, given that it's the default for any mechanical or technological market.
But, I think protectionism will win, so there will be more, although few per market.

 

[Mo City]

Agree with this completely. In the end, there will be 4-5 global entities (excluding China) that will dominate the industry. Tesla's top-to-bottom integration is forcing this change decades sooner than it would have happened naturally.

Tesla'ls business model may be the ONLY model that can succeed. Every company currently manufacturing autos, creating software and sensor tech used in autos, involved in ride-sharing, etc must navigate a route to the this destination or face extinction.

In particular, I feel Uber has a difficult future.

 

[mikes_fsd]

His main argument is that autonomous driving requires an end-to-end solution that combines many different technologies. T

Man, if only there was some company that did this already...
Sounds to me like the man is validating the Tesla approach! (End-to-End + vertically integrated).

 

[diplomat33]

Man, if only there was some company that did this already...
Sounds to me like the man is validating the Tesla approach! (End-to-End + vertically integrated).

And there is a company doing end-to-end and vertical integration of both hardware and software for FSD: Waymo. They custom built their sensors in-house and do the software.

He is validating the end-to-end and vertical approach part of Tesla's FSD plan but not the specifics of Tesla's approach. Sashua says "in order for autonomous vehicles to make the best driving decisions, they will need to process data from a combination of cameras, high-definition maps, radar and lidar". So he is not agreeing with Tesla's sensor choice for FSD. Also, remember that Mobileye and Tesla split because they did not agree on the right path to FSD.

 

[mikes_fsd]

He is validating the end-to-end and vertical approach part of Tesla's FSD plan but not the specifics of Tesla's approach. ...

Thank you captain obvious.... for that specific reason I listed out what he was validating in parenthesis!

 

[diplomat33]

Thank you captain obvious.... for that specific reason I listed out what he was validating in parenthesis!

Yes, but there is more to Tesla's approach that end-to-end and vertical integration.

 

[heltok]

I remember when I asked George Hotz about Amnons statement regarding end2end learning a few years ago. This is what he said:

Me:
Amnon Shashua wrote a paper claiming that End-End Deep learning for SDC is not feasible. On the Sample Complexity of End-to-end Training vs. Semantic...

Any comments?

George Hotz:
In other news, horse and buggy maker talks about why removing the horse from the carriage is just not feasible.
Some things are not possible, like FTL communication. Other things, like this, will just take time and cleverness.

 

[TLLMRRJ]

Very good article that doesn't just call autonomous vehicles a pipe dream, like I do, but shows how the industry is realizing this fact on their own, and how companies like Waymo are a farce. They are doing what anyone does when they are working on a doomed project, and redefining what constitutes a success.

But rest assured, your hopes of using your current Tesla as a robotaxi to drive you or others around without a human at the controls are completely dashed at this point.

Self-driving industry takes to the highway after robotaxi failure

 

[qdeathstar]

Interesting CEO comments. So can I interpret that to mean his company can't do it alone, so he thinks other companies can't do it either? Maybe I'm reading too much into that statement

Looking for a buyer