Tesla’s software lead is so big it should worry other automakers, AI expert says

[Hamartia Antidote]

https://electrek.co/2019/05/08/tesla-software-lead-big-worry-automakers-ai-expert/

According to Lex Fridman, a research scientist at MIT working on human-centered artificial intelligence and autonomous vehicles, Tesla’s software lead is so big it should worry other automakers and could even slow down the pace of innovation due to the lack of competition.

Fridman is known for his work studying Tesla’s Autopilot and other driver assist and autonomous driving systems.

In a new interview on the Joe Rogan Experience podcast, the AI expert explained that he supports a “software first” approach to developing self-driving systems.

He said:

“I am 100% for the software approach. I think we need to let good software engineers lead the way is the way to go. I wish other companies were competing with Tesla on this. On the software side, Tesla is far ahead of everyone else in the automotive sector and that’s one of the problems. Competition is good and I am worried that there’s people way too far behind to actually give Tesla new ideas and outcompete Tesla on software.”

He went on to explain how most cars can’t do over-the-air software updates other than to the infotainment system.

While it’s true that no other automaker has quite the software update capability of Tesla, several automakers, like GM and Porsche, have announced efforts to increasingly expand their capability to send software updates to vehicles.

The AI expert argued that the software is at the foundation of Tesla’s vehicles and it’s not the case for any other major manufacturers.

Electrek’s Take
I think he makes a good point. Tesla is definitely years ahead of any other automaker when it comes to software and yet, there are still learning and tweaking their approach when it comes to the technology.

Recently, Tesla released a new update that lets owners have more control over when they get new software updates pushed to their cars.

Tesla also introduced software update signatures to protect its vehicles against hacking not that long ago.

Those are all things that Tesla is just now figuring out after years of focusing on vehicle software as a fundamental part of its vehicles.

I understand why Fridman would think that other automakers have a lot to do in order to catch up, but hopefully, they are watching closely and they will learn a lot from Tesla.

But I agree that as long as they play catch up, they are not really pushing Tesla through competition. That said, I am not sure that Tesla really needs that to innovate since they are mostly mission driven.

 

[Falcon29]

Bmw, Cadillac are doing well in this department too. I don't think it is a big gap, but I am not a fan of the idea. Cadillac has a super cruise feature which will take over on freeway but still requires driver to have attention and take back control. Anything more than that would be worrying.

 

[HAIDER]

https://electrek.co/2019/05/08/tesla-software-lead-big-worry-automakers-ai-expert/

According to Lex Fridman, a research scientist at MIT working on human-centered artificial intelligence and autonomous vehicles, Tesla’s software lead is so big it should worry other automakers and could even slow down the pace of innovation due to the lack of competition.

Fridman is known for his work studying Tesla’s Autopilot and other driver assist and autonomous driving systems.

In a new interview on the Joe Rogan Experience podcast, the AI expert explained that he supports a “software first” approach to developing self-driving systems.

He said:

“I am 100% for the software approach. I think we need to let good software engineers lead the way is the way to go. I wish other companies were competing with Tesla on this. On the software side, Tesla is far ahead of everyone else in the automotive sector and that’s one of the problems. Competition is good and I am worried that there’s people way too far behind to actually give Tesla new ideas and outcompete Tesla on software.”

He went on to explain how most cars can’t do over-the-air software updates other than to the infotainment system.

While it’s true that no other automaker has quite the software update capability of Tesla, several automakers, like GM and Porsche, have announced efforts to increasingly expand their capability to send software updates to vehicles.

The AI expert argued that the software is at the foundation of Tesla’s vehicles and it’s not the case for any other major manufacturers.

Electrek’s Take
I think he makes a good point. Tesla is definitely years ahead of any other automaker when it comes to software and yet, there are still learning and tweaking their approach when it comes to the technology.

Recently, Tesla released a new update that lets owners have more control over when they get new software updates pushed to their cars.

Tesla also introduced software update signatures to protect its vehicles against hacking not that long ago.

Those are all things that Tesla is just now figuring out after years of focusing on vehicle software as a fundamental part of its vehicles.

I understand why Fridman would think that other automakers have a lot to do in order to catch up, but hopefully, they are watching closely and they will learn a lot from Tesla.

But I agree that as long as they play catch up, they are not really pushing Tesla through competition. That said, I am not sure that Tesla really needs that to innovate since they are mostly mission driven.

tesla need a smart competitor , he is right . otherwise ..... i don't know how this software react in future ... when machine turn into enemy ...

https://www.theverge.com/2014/1/28/5339246/war-machines-ethics-of-robots-on-the-battlefield

SHOULD A ROBOT DECIDE WHEN TO KILL?
The ethics of war machines

 

[fallstuff]

Always hearing Tesla is this , Tesla is that, Tesla can be driven on water, and maybe someday Tesla will cure cancer. However, the big question is Tesla selling enough cars now so it can pay its bills ?

The problem is even an electric car with 300 miles is still not practical for average American family living in the cold weather states as the heater eats up to 30-35% of the driving range depending on the temperature. Additionally, Tesla recommends not to drain the battery below 20% as this will reduce the battery expected life. Unfortunately this leaves only a 50% usable charge.

It maybe fine for driving around town or to and from work in extreme cold weather, but any trip of 100 miles or more will take careful planning as this maybe a life or death situation being out of charge in extreme cold weather.

 

[fallstuff]

More competition coming in the next few years. Cars like Audi Etron, Electric BMW X3 and one from MB. I wonder how Tesla will hold up against these premium car makers.

 

[Beast]

Always hearing Tesla is this , Tesla is that, Tesla can be driven on water, and maybe someday Tesla will cure cancer. However, the big question is Tesla selling enough cars now so it can pay its bills ?

The problem is even an electric car with 300 miles is still not practical for average American family living in the cold weather states as the heater eats up to 30-35% of the driving range depending on the temperature. Additionally, Tesla recommends not to drain the battery below 20% as this will reduce the battery expected life. Unfortunately this leaves only a 50% usable charge.

It maybe fine for driving around town or to and from work in extreme cold weather, but any trip of 100 miles or more will take careful planning as this maybe a life or death situation being out of charge in extreme cold weather.

Technology will improve. We may even reach a stage charging 300miles range EV will be even faster than filling up a gasoline tank of similar range.

 

[fallstuff]

Technology will improve. We may even reach a stage charging 300miles range EV will be even faster than filling up a gasoline tank of similar range.

Only possible with solid state batteries which are still at research level with some semi-solid state making into the market.

 

[Beast]

Only possible with solid state batteries which are still at research level with some semi-solid state making into the market.

Technology improved very fast especially with large number of funds poured in EV. It's possible in near future such tech will be available.

 

[Hamartia Antidote]

Tesla Top Speed On Navigate on Autopilot & Lane Change (Germany Autobahn)

 

[Menthol]

I haven't test it for myself.

But seeing a youtube video of Tesla self-driving, I think it will still take time before we can have a fully automated driving car.

From the video, the driver still need to guard if the AI made a mistake, especially if the AI misidentified the object on the street. Like misread the traffic lights, unusual road shape, etc.

I think for a truly automated driving, we can't just depend on the camera and AI alone, there should be assistant on the road, some kind of guide sign or signal for AI so they don't miss read it.

And we also need a communication between cars on the street, so AI can adjust each other position toward other cars, or arrange every car position to maximize traffic smoothness.

After that, we can have a truly full automated driving car.

 

[Cookie Monster]

Only possible with solid state batteries which are still at research level with some semi-solid state making into the market.

Tesla has already demonstrated a "battery replacement" ability of sorts. Basically a car in need of a charge drives up on to a platform, an automated system switches out the car's battery with a fully charged one. All of this takes the same amount of time as a person completely filling up his almost empty gas tank.

While it's just a concept, it can easily be implemented on the many charging stations Tesla has already opened(with more planned). In addition technology improves fast. Just look back at how electric cars were 10-15 years ago...now look at where they r now. In the next 10-15 years they would sure improve even more to the point where worrying about range and charging time wouldn't be an issue.

 

[Hamartia Antidote]

I haven't test it for myself.

But seeing a youtube video of Tesla self-driving, I think it will still take time before we can have a fully automated driving car.

From the video, the driver still need to guard if the AI made a mistake, especially if the AI misidentified the object on the street. Like misread the traffic lights, unusual road shape, etc.

I think for a truly automated driving, we can't just depend on the camera and AI alone, there should be assistant on the road, some kind of guide sign or signal for AI so they don't miss read it.

And we also need a communication between cars on the street, so AI can adjust each other position toward other cars, or arrange every car position to maximize traffic smoothness.

After that, we can have a truly full automated driving car.

There's already self-driving cars on the road...but by Waymo.

Obviously not every situation has been thought up but it can easily handle many situations. Until they are confident they do keep a safety driver behind the wheel for the general public (to avoid a bad PR incident)...but none with people who basically signed their life away in the "early rider" program.

general public with safety driver

people in the early rider program with no driver

 

[Hamartia Antidote]

 

[nahtanbob]

Tesla Tells Employees It’s Preparing to Increase Production
https://www.bloomberg.com/news/arti...t-s-preparing-to-lift-output-at-fremont-plant

 

[Menthol]

I wonder what is the progress of Google self driving ai?

If Google succeed, it will isolate Tesla, just like how Google Android against Apple iOS.