My take on Tesla's current position in the industry.

[Dan Detweiler]

Everyone has different wants, needs and expectations in what they are looking for in their transportation. This will always be the case and because of this, there will always be disagreements on everything. Here are my impressions and I would be interested in hearing yours.

I have gone through a lot of stages during the digestion of the "Model 3 reveal part 2" or whatever you want to call it this week. My initial reaction was one of disappointment due to it not including the things I had begun to expect out of Part 2. Interior design, performance specs, steering wheel, HUD, etc. None of these things were discussed...bummer. Then I began to think. These expectations I had were my own interpretations of reality, not based on anything factual. I set myself up for this disappointment. Once I realized that fact, I started to think about what WAS said in the reveal.

With the release of Autopilot 2.0, Tesla has announced that they are putting themselves in a place no other manufacturer has ever realistically even considered. They are throwing down a gauntlet to everyone else in the industry that there is no way anyone can match in at least the next 10 years in my opinion. They are opening up a level of utilitarian function we have never seen before in our cars, let alone the convenience and safety issues. When you really start to think about how this technology can be used it is mind boggling to me. How about the potential impact on the lives of our elderly, handicapped and otherwise incapable citizens that can gain back a level of independence that has never been possible for them up to this point. Intoxicated drivers become intoxicated passengers when the car safely takes them home after a long night of partying. People suffering from emergency medical problems are taken directly to the nearest hospital by their cars. The possibilities are endless.

After much thought, there is now no way I would consider anything other than a Tesla for me and my family. I think they have literally made every other car on the road obsolete in many many ways. Performance, safety, convenience, functionality, looks...you name it. No matter how I look at it no other car company can begin to match it. Sure, I understand many people will disagree and that is one reason I am posting this...to hear other people's thoughts. For me however, Tesla just knocked it out of the park with this announcement. The only thing that would keep me from buying a Tesla as my next car would be if the price got completely out of hand, which for some it may have already, I don't know.

Bottom line...I think the Model 3 with Autopilot 2.0 and full autonomy is an absolute game changer in the industry on par with Ford's development of the assembly line. Like the assemble line, if any other manufacturer can hope to compete, they are going to have to start adapting some of what Tesla is now making available in their cars. Lord, I sound like a Tesla commercial here! Not meaning to but the more I think about it the more excited I get.

What are your thoughts?

Dan

 

[JeffK]

After much thought, there is now no way I would consider anything other than a Tesla for me and my family. I think they have literally made every other car on the road obsolete in many many ways.

What are your thoughts?

That depends... the self driving features won't be fully enabled until mid to late 2017 at the earliest. Volvo plans to have a Drive PX 2 powered car in 2017. There are at least two chinese backed electric car companies who plan to release models comparable to the model S in 2017/2018 etc. We don't really know if they will try to undercut Tesla on pricing.

Tesla is first to market and that's important, but there is competition coming. It lags behind by a few years, but it is coming.

The Model 3 with all these features seems like everything I could ever ask for in a car. Proper execution of Tesla's ramp up plans is vital at this point if they want to stay ahead of the curve. We also don't know if they'll have something special up their sleeves for the Model 3 reveal part 3.

I definitely wouldn't advise them to start slacking off in the innovation department once the Model 3 is a success. By 2020, it's going to be a totally different world.

 

[ecarfan]

Dan, glad to read how excited you are about Tesla and about the announcement this week that every Tesla is being built with the hardware for full autonomous driving. I agree with everything you wrote. I think the disappoint some people expressed about the Tesla announcement is in part due to their having expectations driven by rumor and speculation. They don't appear to understand the significance of what Tesla has done. But this has been discussed endlessly in other threads and I have posted my opinion in detail there.

 

[3Victoria]

I think TM continues to be on track with the Model 3. Most of us had confused Part 2 with Part 3 ... but we had guessed at the naming scheme from one data point, so we can be excused by our enthusiasm. There will be a second big reveal to show the interior, my best guess is end of March.

Achieving level 5 AP will be very difficult for technical and political reasons, but even the existing functionality is worth having, and any advance welcome. Tesla has pushed this much closer in time. If we get to level 5 anytime in the next 1-10 years, it will be a win. Predicting the changes to our personal lives a society is difficult. All change comes with pluses and minuses.

 

[msanborn08]

I think the M3 is the make it or break it point for Tesla. Even more so now with the release of L5 hardware / public disclosure of enabling software. I think it is so much so that if Tesla fails to get as many of these cars into customers hands with AP enabled in the next 2-3 years you'll end up seeing Tesla become the fall back go to company for any issues or regulation policy's once the Fed's catch up with the technology. Understandable, the mass public opinion needs to be overwhelmingly positive as well.

What I am getting at here is, at the moment the microscope is focused on Tesla, while the other auto manufactures sit back and watch how this bleeding edge technology is going to unfold. They get to watch not only how Tesla is solving the problem and develop a competing product in the shadows, but also watch the public / government response. One might speculate that the legacy auto manufactures position is to use Tesla as a burner company. Use them as a beta test, then pivot their legislative influence they have accumulated over the years to force them out of their own market.

 

[Dan Detweiler]

That depends... the self driving features won't be fully enabled until mid to late 2017 at the earliest. Volvo plans to have a Drive PX 2 powered car in 2017. There are at least two chinese backed electric car companies who plan to release models comparable to the model S in 2017/2018 etc. We don't really know if they will try to undercut Tesla on pricing.

Tesla is first to market and that's important, but there is competition coming. It lags behind by a few years, but it is coming.

The Model 3 with all these features seems like everything I could ever ask for in a car. Proper execution of Tesla's ramp up plans is vital at this point if they want to stay ahead of the curve. We also don't know if they'll have something special up their sleeves for the Model 3 reveal part 3.

I definitely wouldn't advise them to start slacking off in the innovation department once the Model 3 is a success. By 2020, it's going to be a totally different world.

Something else to keep in mind that gives Tesla a HUGE leg up on everyone else. When it comes to legislative regulation I assume it will be similar to the aviation industry. When it comes to aviation certification, everything is hardware, software and application specific. You can't just plunk a battery that is certified in a Cessna into a Piper. That same battery has to be certified in each application. So, my assumption is that they won't just say at some point "ok, full autonomy is now authorized and any manufacturer can do what ever they want". My guess is that each manufacturer will have to demonstrate (probably billions of miles of data) the viability of their system, even if it uses the same hardware as a certified system.

This would put Tesla years and years ahead of all the other systems out there due to its collective data network that will now include every single car they sell, whether it has autopilot enabled or not. Something to keep in mind.

Dan

 

[LectrikPower]

I hope Tesla has a more affordable SUV type car in the next 5 years or so because I really can't see myself ever buying another ICE car.

That is significant to me. I am a driving enthusiast and although I will terribly miss driving a standard transmission car I just can't see buying a dirty, smelly inefficient ICE. I really see electric catching on in a big way. It makes ICE seem ridiculous. I love that Tesla proved everyone wrong and I'm hoping that other car makers follow Tesla so we can quickly move to all electric transport.

 

[zenmaster]

They are certainly leading the industry in rapidly deploying what driving assist tech they do have, which is great. But I don't see how Tesla can be ahead in autonomous driving AI, which is the most important component for practical full autonomy.

 

[JeffK]

But I don't see how Tesla can be ahead in autonomous driving AI, which is the most important component for practical full autonomy.

Easy, AI and DNNs are all about training data. Tesla by far has collected the most training data of any auto manufacturer in the world.

Consider Tesla *five months ago* said they have over 780 million miles worth of data. Tesla reveals new details of its Autopilot program: 780M miles of data, 100M miles driven and more

Google by comparison has spent years and years collecting data and they have only recently passed 2 million miles. After two million miles, Google’s robot car now drives better than a 16-year-old

 

[melindav]

and Google's data is isolated to a few locations, whereas Tesla's data is being pulled from everywhere they have owners driving.
Is Google's "better than a 16 YO" with it in the bay area? how is it when driving in WDC?

 

[Jayc]

After much thought, there is now no way I would consider anything other than a Tesla for me and my family.

Dan

I agree with most of what you say. However, in the Model 3, Tesla will also have to demonstrate either that they can do a fairly reliable car to start with or that they will deal with any issues promptly. For me that seals the deal as well as guarantees continued loyalty to the brand.

 

[Dan Detweiler]

I agree with most of what you say. However, in the Model 3, Tesla will also have to demonstrate either that they can do a fairly reliable car to start with or that they will deal with any issues promptly. For me that seals the deal as well as guarantees continued loyalty to the brand.

Yes, that is true. Perhaps I am naive but I really think Tesla learned a very difficult and yet priceless lesson with the Model X. I think they are making it a policy to only make public, information that has been known to them secretly for a long time. I think they realize what's at stake moving forward.

Dan

 

[zenmaster]

Easy, AI and DNNs are all about training data. Tesla by far has collected the most training data of any auto manufacturer in the world.

Consider Tesla *five months ago* said they have over 780 million miles worth of data. Tesla reveals new details of its Autopilot program: 780M miles of data, 100M miles driven and more

Google by comparison has spent years and years collecting data and they have only recently passed 2 million miles. After two million miles, Google’s robot car now drives better than a 16-year-old

Training data is helpful, but I'd argue the AI algorithms are the essential tech and what would ultimately differentiate the industry "leaders".

 

[Dan Detweiler]

Training data is helpful, but I'd argue the AI algorithms are the essential tech and what would ultimately differentiate the industry "leaders".

Not as far as government certification is concerned. This is all uncharted waters and as such is all speculation but if the government handles this at all like FAA certifications of aviation technology, manufacturers will have to demonstrate millions, maybe billions of miles of driving data, as well as system redundency. There is no other company poised to provide that much data any time soon...except Tesla.

Of course, this is just my opinion. Only time will tell how this will all play out.

Dan

 

[zenmaster]

Not as far as government certification is concerned. This is all uncharted waters and as such is all speculation but if the government handles this at all like FAA certifications of aviation technology, manufacturers will have to demonstrate millions, maybe billions of miles of driving data, as well as system redundency. There is no other company poised to provide that much data any time soon...except Tesla.

Providing sufficient driving data is important. But again the most important thing for industry leadership, as far as full autonomy, would be the actual capability of the car which is dependent on the functionality provided by algorithms.

 

[woof]

They are throwing down a gauntlet to everyone else in the industry that there is no way anyone can match in at least the next 10 years in my opinion.

These Tesla is "10 years" or sometimes "5 years" ahead of everyone else is, in my opinion, easily refutable.

1. It was less than 5 years ago that Tesla started concentrating on autonomous operation, so there's one data point that it can be done from 0 to something in less than 5 years. Chances are the next player isn't starting from 0.

2. Once someone has shown something working is a great motivator to others who previously said "It can't be done." See Bob Lutz and Chevy Volt for such an example.

3. If Tesla patents how they do it, it should enable others to learn from the patents...and they won't even have to license them due to the "all our patent are belong to you" statement.

4. Tesla's "smart guys" are most likely no smarter than other's "smart guys" (and those same guys have probably already changed jobs, in the silicon valley tradition). The difference comes down to will and funding. It's assumed the big 3 have neither the will nor the funding. That may not be correct, and the Europeans certainly have the funding and research, but never seem to have the will. That may change now that Tesla has gone first. Also suddenly there are a plethora of new players in the market (see statement 2) such as Faraday Future, NextEV, Lucid Motors, etc. who appear to be just as willing as Tesla to push the boundaries.

5. "Tesla has all this data!" Yep, they do. And assuming they won't share, (I think they will, but that's another story) how hard will it be for Toyota or GM to slap some data collection hardware into one of their best sellers and collect their own data 1000 times faster than Tesla? AP hardware was only introduced 3 years ago. There are less than 100K Tesla's with that hardware driving around collecting data over the last 3 years. Assume Toyota outfits the 2017.5 Camry with the required sensors...at 400K units per year they can make up that ground really fast if they want to.

 

[JeffK]

Providing sufficient driving data is important. But again the most important thing for industry leadership, as far as full autonomy, would be the actual capability of the car which is dependent on the functionality provided by algorithms.

There seems to be some confusion... there are very few, if any, hand coded algorithms in a DNN. As a matter of fact hand coded algorithms often make it less robust according to Andrew Ng, who is an expert in the field.

For example, if programming image recognition using something like Alexnet, you'd only need to provide it sample data for training and validation and the machine does the learning on it's own. Once you can recognize everything in an images then you can do pathfinding etc.

It's not a whole bunch of if/then statements and you don't have to program every single case for it to be able to figure out what to do in a situation it has never encountered unless it's wildly outside the bounds of normal, in which case you'd need to train on it.

 

[zenmaster]

There seems to be some confusion... there are very few, if any, hand coded algorithms in a DNN. As a matter of fact hand coded algorithms often make it less robust according to Andrew Ng, who is an expert in the field.

For example, if programming image recognition using something like Alexnet, you'd only need to provide it sample data for training and validation and the machine does the learning on it's own. Once you can recognize everything in an images then you can do pathfinding etc.

It's not a whole bunch of if/then statements and you don't have to program every single case for it to be able to figure out what to do in a situation it has never encountered unless it's wildly outside the bounds of normal, in which case you'd need to train on it.

Not sure what you are arguing. There are inherent limitations in any given problem approach, regardless of your data set size. In other words, sure the number of miles driven improves the NN training but only within a limited scope determined from the start by how those NNs were used by the system. The driving logic can and should be thought of as a set of algorithms with the NNs being tools utilized by those algorithms.

 

[EinSV]

Dan,

I agree with much of your post. Tesla already had a significant lead with AP 1.0 and IMO they appear to have left the competition in the rear view mirror with the introduction of AP 2.0.

I do think the end game is somewhat complex, however. It appears likely that Tesla will achieve Level 4 functionality quickly in some locations. That by itself could be a game changer. But especially with the possibility of a patchwork quilt of regulations (both within the U.S. and internationally) fully autonomous driving could be delayed in many areas, which could create a window for competitors to try to catch up. As @woof noted, Tesla does not have a monopoly on the concept of fleet learning, although it does have a huge head start. So I think it is too early to declare game over, but I do think Tesla is in an enviable position.

This is undoubtedly an industry changing event (with the potential to be a societal changing event in some ways). And it certainly appears likely that it will start very, very quickly, or as Elon said in August "a hell of a lot faster" than people think.

 

[Red Sage]

What are your thoughts?

Essentially, that I agree with you. The revelation you experienced upon actually thinking about it, is just the sort of thing that I expected. The difference is that far too many will be too busy complaining about what was not said, presented, or answered, and will not take the time to reflect upon what was actually said. Denial is their first, last, and only impulse.

That process is similar to two years ago, when so many news organizations sought to downplay the introduction of the Model S P85D at the Tesla 'D' Event. They basically claimed that Tesla had wasted everyone's time by unveiling what 'everyone else' already had -- an all wheel-drive car with traffic aware cruise control -- and thereby 'nothing special' at all. They wanted to show that Elon Musk was naught but a hype leader, a ringmaster, a showman, and that there was no substance to what was presented at all. That is certainly one way to look at it, and I understand how some might come to that conclusion. But it is the wrong one nonetheless.

The same is true for those who want to downplay this latest announcement from Tesla. Again, no one seems to quite ready to accept the importance or consider the relative magnitude of what is happening here. Tesla motors has gone from being the first mass market vehicle manufacturer to offer 100% fully electric, to be the first to offer a 100% self driving capable fleet of new vehicles.

There are still those that want to refer to the 'D' as a 'one trick pony', only noting its acceleration. While they simultaneously ignore the improvement in range and efficiency or in safety and handling it affords. That same group of people will not let go there criticism that Tesla is somehow 'beta testing the world' and that they are moving 'too fast, too soon' or whatever. And the exact same people will never admit to talking out of both sides of their mouth when they simultaneously claim that Tesla is 'taking too long' and 'missing deadlines'. Tesla's methodology for everything under the sun will always be criticised, and mostly unfairly. Such is life and stuff.