The Tesla Self-Driving Regulatory Myth

| About: Tesla, Inc. (TSLA)
This article is now exclusive for PRO subscribers.


Tesla is fertile ground for myth after myth.

A myth that remains unchallenged is the one about the need for regulatory approval when it comes to self-driving features. It's easy to see why it's a myth.

There are other developments as well.

To show how freakishly unreal things can get with Tesla, this car tried to kill its owner. While on Autopilot it provoked a massive accident which could have easily killed the driver, but didn't. The driver is grateful that the car "saved him." Again, the car provoked the accident. Is this the Stockholm Syndrome or something? It's like being stabbed by someone and being grateful to the attacker because he missed the heart or something.

One of the most amazing things about Tesla (NASDAQ:TSLA) is how Tesla fans manage to hold thoughts which are so patently false, against all types of evidence.

For instance, they widely perceive Tesla as being a leader in self driving, even though actual testing data on self-driving cars shows it to lag far behind. They also widely perceive the AP2 evolution to be reliant mainly on getting more data from customers driving their cars around, in spite of the fact that things like automatic wipers also are part of the missing features (implying actual hard programming is what's missing).

Another such example, which is particularly relevant, is Tesla's preemptive usage of a red herring to make up excuses for delays. This particular red herring is the supposed need for regulatory approval before fielding self-driving features. Here's how Tesla presents the red herring:

Source: - Order page

Why is this a red herring, might you ask? For a very simple and obvious reason:

  • Because all those self-driving features which are said to require regulatory approval, can be fielded simply as driver assistance features. That is, they might provide "self-driving," but they just need to come with the disclaimer "the driver needs to pay attention at all times." Indeed, Tesla itself says as much when it says, regarding Enhanced Autopilot, "That said, Enhanced Autopilot should still be considered a driver's assistance feature with the drive responsible for remaining in control of the car at all times."

So, given that Tesla just recently (January) fielded totally unusable lane keeping features (which are now being improved upon), who actually believes that much more advanced (even if faulty) self-driving features, if billed as driver assistance, would be opposed by regulatory bodies?

Makes no possible sense at all. If inferior driver assistance features are allowed, there are no grounds to block much more advanced - but still driver assistance - features.

As a result, if ever someone thinks Tesla has technology in this regard which it's not fielding in its cars because of regulatory opposition, that person believes in an obvious myth.

This is a similar reasoning to what I put forth in my Intel article titled "Intel Is Making A Mistake." To put it simply, today's existing self-driving tech is many orders of magnitude more advanced than today's fielded driver assistance technology. Players other than Tesla, which are much more advanced than Tesla in developing self-driving tech, cannot put it in customer cars until they get cheaper LIDAR units to go with it. However, Tesla can indeed put its self-driving tech - however incomplete - into customer's cars, just as long as it bills it "driver assistance." As a result, Tesla has no regulatory impediment for doing so, because however bad that self-driving tech might be at actually self-driving, it would still be years ahead of basic driver assistance of the kind Tesla already ships its cars with, at providing driver assistance.

As a result:

  • The only time Tesla would claim "regulatory approval needs" would be when it doesn't actually have the tech to deploy even in a driver assistance role. Tesla has fielded obviously-incomplete features, so no logical reason to withhold better features for no reason.
  • Tesla's driver assistance features will actually be a good proxy for how advanced Tesla's self-driving effort is, since there is no reason to keep self-driving capabilities out of costumer cars. The only requirement is to just label those capabilities "driver assistance" while they're not capable of good enough self driving.
  • And finally, in light of this logic, the initial Tesla self-driving demos, while wildly inferior to what other automakers are already testing, are still highly suspect. That is, there is reason to believe those self-driving capabilities didn't even exist at the time the demos were filmed, other than under very specific conditions (probably, going over and over the same path to teach the car to replicate that same path). This (or similar) would be a reason for the supposedly displayed capabilities not to be fielded yet. Of notice, in the second video there isn't a single steering avoidance action - the car just interrupts the journey each time there's a possible obstacle (for instance, notice when it goes past two joggers at 1:31). Remember, any capabilities which can be generalized (like keeping to the correct lane and road the way it does in that video) would instantly be available to be billed as "driver assistance." Yet for months the AP2 version available to customers was - and still is - much worse than what's displayed there, to the point of being straight-up dangerous and unreliable in similar roads.

Another Myth, By Baird

Hardly a day goes by where something else that's patently outrageous (pun intended, as we'll see) is said about Tesla. The latest is Baird's take on Tesla's innovation, as if some kind of Tesla advantage:

We believe TSLA will continue to apply for new patents as the company is constantly innovating to improve the quality, safety, and features of its vehicles.

Well, for sure Tesla applies for patents. But people are so mystical about this that:

  • Even Seeking Alpha pointed toward a patent search for "Tesla" which would have brought up more than 12,000 entries. Seems nice, doesn't it? At least until we consider that "Tesla" is also a magnetism unit, and thus nearly all of those patents aren't actually Tesla's but someone else's.
  • What's the true number of actual Tesla patents? Try an advanced search for "AANM/Tesla" (applicant name containing Tesla). How many of those are there, since 1976? 83 (that's "Eighty Three").

Now, let's consider the competition.

Someone has actually done the work for us here. Just taking a sample of five different automakers, and just for 2016, we have:

  • Ford (NYSE:F) at 1,524.
  • Toyota (NYSE:TM) at 1,417.
  • General Motors (NYSE:GM) at 1,123.
  • Hyunday at 1,035.
  • Honda at 922.

Remember, this is just for 2016. These companies file about these many patents every year. This is versus the 83 for Tesla's entire past history.

The thing with Tesla is that creating myths around it is so easy. Even in the most extreme of circumstances, people believe these myths. They believe in a self-driving lead against all evidence. They also will believe in some kind of innovation lead here, too. And backed by patents, too, for sure. Baird seems to have had no trouble believing it.

Latest Development In Countergate, The Sequel

I've covered the Countergate scandal in two of my articles, "A New Tesla Scandal Is Brewing" and "Tesla's Countergate II - The Sequel." This scandal was unearthed by Tesla's customers, particularly the most technologically adept and curious among them.

While at first it seemed that Tesla would cave in and remove the performance limits it had imposed, reality turned very different. Instead of removing performance limits imposed "a posteriori" on Performance model customers, Tesla actually widened the limits. Basically, Tesla made every Performance car with Ludicrous mode be performance-constrained except under very specific situations (the use of Max Battery at the same time as Launch Mode, which basically excluded everyday driving).

Now, many customers could find out these constraints by using apps and websites such as TeslaFi and other apps relied on Tesla's undocumented API, which got data directly from the car.

So in light of the Countergate scandal, what has Tesla done? It removed the battery_current response from the RESTful API. How is this relevant? Well, battery current, along with voltage, can be used to directly calculate the power being produced by the car. Thus, it was an easy path to check what power was available now, vs. in the past. Moreover, the performance limitations imposed on customer cars were apparently implemented by limiting the maximum allowed battery current. Think what you will.


The main conclusion to be drawn from this article is that Tesla's progress on self-driving will be trackable through seeing how far along AP2 is. Since any self-driving abilities can be passed along as driver assistance features, there's no reason for Tesla - which has shown itself able to field wildly inept lane keeping assistance - to keep much better features from customer cars.

These features, thus fielded in customers' cars, will seem well ahead of what competitors have. Yet, such isn't the case - these features will basically be direct competition to what others are testing and which can handle heavy and complex city traffic for hundreds/thousands of miles between needed human interventions.

Moreover, if Tesla at some point argues that the lack of regulatory approval is keeping it from fielding self-driving features will basically be a red herring. It's perfectly obvious that regulators didn't keep Tesla from deploying wildly imperfect lane keeping, so there's no reason to believe any actual self-driving feature would be barred from deployment as long as it required the driver to keep attention.

Self-driving ability and supposed leadership is a large part of Tesla's current market value. Indeed, Morgan Stanley even attributed Tesla's entire upside (which the stock has already fully realized) to its Tesla Mobility inexistent business. Tesla Mobility would require fully working Level 5 self-driving to even exist (since the car would have to drive itself with no one inside, to go pick up customers).

Even in this article, I show that this isn't the only myth. Lauding Tesla's patents when its competitors each file 10-20x more patents per year than it did in its entire history is particularly egregious.

Disclosure: I am/we are short TSLA.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.