Facebook, Uber, And Tesla
Summary
- The Facebook "Data Breach" suggests that more government regulation in general will be focused on "emerging" technologies and services.
- An UBER "self-driving" vehicle being tested in Arizona not recognizing both a pedestrian and bicycle may significantly delay the development, testing, development of "autonomous driving" vehicles.
- Tesla's stock valuation is significantly supported by believing that all of Elon Musk's proclamations will come true.
- Significant delays in being able to deploy "self-driving" functions could further detract from the "Tesla Story" that keeps the stock levitated.
Introduction
A prominent news item this week is what is being called the "Facebook Data Breach." Although I am not intending to either defend Facebook (NASDAQ:FB) or comment in length on that entire situation, in my opinion, what happened was what unfortunately now seems to be the "new norm" of "anything goes" in terms of how many organizations now do business.
The Facebook situation was not an explicit "data breach" where external hackers accessed Facebook's systems to steal information such as what happened with Equifax (NYSE:EFX) but actually a company-enabled event based on the company's operating policies at that time. In the very complicated way that many things can now work, Facebook's policies were that anyone who had a "connection" with a Facebook profile also then had access to data from the Facebook profiles of all of the "friends" of the initial profile.
In the current situation, Facebook's policies, although probably not completely understood by users, were apparently very well understood by external entities who would like to use as much user data that they could figure out how to gain access to. As such, an external consultant working with Cambridge Analytica posted some sort of "survey" on Facebook that 300,000 apparently bored people with Facebook profiles found interesting to respond to but those users apparently had, on average, 160 "friends" each. And so, 300,000 times 160 equals the headline number of "50 million users" whose data had been "scraped" from Facebook's systems.
As an aside, I actually find a lot of this both comical and dysfunctional. Although I am a very active and enthusiastic user of technology and have immensely benefited from the growth of the internet and all of the information that is now so accessible, I have not been an active "social media" participant. I have a very active life and prefer interacting with my real friends instead of passively "following" whatever is supposedly going on with multitudes of "online friends" that someone had some connection to in the past. As such, I also question how much is even "real" at this point about all of the virtual lives that people have constructed for themselves online or - in the case of the activity and interests data scraped from 50 million Facebook users' profiles - how much of that was really even real or useful itself.
Where I am going with this is that we seem to have entered a new environment where the intersection of the real world and the virtual world seems to have created a sort of fuzzy blurred space where real things and real issues are no longer sharply in focus. Effectively, this is a modern version of Descarte's proclamation of "I think, therefore I am" but somewhat changed to "if I can think about it, then it might be true."
As the human mind can think about many things but actually getting things done on the face of the earth is far more difficult for many reasons, there is then a disconnect - which often can be quite dangerous - between things that people think they either want to do or think that might be possible and what is actually achievable.
Such a disconnect was highlighted by someone connected to the Cambridge Analytica situation (Aleksander Kogen - the creator of the "survey" put on Facebook for data gathering purposes) who was quoted in a recent Reuters article as dismissing Cambridge Analytica's claims that the data was "incredibly accurate." In an earlier version of the article, which is unfortunately now not accessible, Kogen also added that there was a belief that the data "tells you everything there is to tell about you, but the reality is it's not that."
Such a disconnect between what I currently consider very "fast and loose" practices in many activities with the additional blurring of reality and potential risks is what now brings me to the very unfortunate recent incident with a "self-driving" vehicle being tested by Uber (Private:UBER) in Arizona.
Recent UBER "self-driving" vehicle incident
I am sure that the investigation into the recent UBER incident in Arizona will take a long time to complete, and so it may be too soon to make any definitive statements, but, on the surface, what is known so far is pretty shocking, in my opinion.
What is known is that a "self-driving" Volvo X90 (OTCPK:VOLAF) vehicle equipped with cameras, radar, and LIDAR, while driving at night, did not apparently recognize or react to a pedestrian who was walking across a road while also holding a bicycle next to them while walking. So, in this case, the systems of the test vehicle apparently did not recognize either the pedestrian or bicycle.
But here are some things that we know which are that:
- both radar and LIDAR can see at night
- and a properly developed and tested system should have been able to recognize both the pedestrian and the bicycle that was also alongside them
But somehow that didn't happen, and how could that be?
In the fast and loose way that a lot of things now appear to be developed, I will now go back to something that I mentioned earlier which is that the human mind can think up a LOT more things than which can actually be either achieved at all - or any reasonable time frame. I think everyone will also acknowledge that the world around us is a very complicated place in many different respects either in the ever-changing characteristics of the physical world or the ever-changing actions of human behavior.
And so, while it is easy to conceive of "hey, let's create vision systems that can control vehicles", the implementation in a very complex world has conceivably an infinite number of possibilities to consider and process.
The processing of all of those possibilities is also in the realm of software as programmers have to somehow envision ALL possible scenarios and initially create programs which can infinitely branch into processing the data received by any "sensors" (cameras, radar, LIDAR, etc.) and then both correctly make decisions about what that data represents and how the underlying systems should respond and react to what is being received.
There is a fundamental problem with software development, however, which is that the more "what ifs" are programmed into each program and the more subroutines are then developed to handle the ever-growing lists of "what ifs" is that the development time can then end up on an "exponential growth curve."
Hmmm, that is an interesting term as haven't I seen or heard that before? Oh yes, isn't that one of the favorite sayings of Elon Musk as he describes things like "new product ramps?" (but which then never seem to occur on time!).
And so, that is the fundamental problem with the hopes, plans, expectations, and things that humans can think about - which is that attempting to implement and achieve such things usually takes far longer than anyone expects to both:
- do something well
- and, in ever-changing world, do those things in a timely way
With those thoughts, I will then move on to how such things apply to Tesla (NASDAQ:TSLA).
Elon Musk's grand visions...
Elon Musk has now promised us "Full Self Driving" (and has been charging for it since October of 2016!), but that is yet to appear. Elon Musk has now also promised us many things but with a very mixed record of either achieving them at all or on in any reasonable time frame as compared to what was originally described.
But, moving on to Musk's vision of "autonomous driving" which he has also now recently described as being available in Tesla vehicles by "the end of 2019," my opinion is that such things are far more difficult to implement at all and, as has now been shown by the UBER incident in Arizona, to be implemented with acceptable risks, than can just be merely thought about. Another comical aspect about all of this as it regards Tesla are also some of Musk's favorite statements implying that he is some unrivaled, all knowing, erudite genius as he talks about "the first principles of physics" (which is actually merely the name of an introductory physic textbook first published in 1912!) and "neural networks."
The whole "neural networks" thing really makes me laugh as although there are definitely very profound and meaningful developments from work on neural networks, the fundamental limitations of neural networks can be seen from this basic description (from this article):
In machine learning and cognitive science, artificial neural networks (ANNs) are a family of models inspired by biological neural networks (the central nervous systems of animals, in particular, the brain) and are used to estimate or approximate functions that can depend on a large number of inputs and are generally unknown.
So, think about that for just a little bit... and then start thinking about the possible issues with "neural network" development (although it always sounds cool when Elon mentions the phrase!). I will also highlight the following excerpt:
- are used to estimate or approximate functions that can depend on a large number of inputs and are generally unknown
That excerpt describes the fundamental issue with such development - which are the large number of inputs (infinitely large in the case of developing safe "autonomous driving" vehicles) and that fact that such inputs are "generally unknown."
And so, however, on the other side of all of this, you have Elon Musk and all of his grand promises. As the very unfortunate pedestrian, who also was accompanied by another object at the same time that also should have processed in the form of a bicycle, a team of programmers just hadn't gotten around to thinking about what additional branches of "what ifs" and subroutines would have to be developed to handle ensuring the safety of all possible persons and objects that a vehicle might encounter in their "self-driving" activities.
What is also really appalling to me about both what I have seen and based on my own programming experience is that images I have seen are things that should have been able to be anticipated in merely initial lab testing of such systems - before such an apparently inadequately developed and tested system had been deployed in the real world.
As such, how much confidence should we really have in Elon Musk at this point about safely and reliably developing systems that will not expose nearby persons and objects to risks, given past statements and actions such as:
- Saying that 100,000 to 200,000 Model 3s would be built in the second half of 2017?
- Making a grand announcement of introducing "AP 2.0" in September 2016, but that system still can't even recognize raindrops on a windshield?
- Announcing in January 2017 that Tesla's "Gigafactory" was shooting out batteries faster than a machine gun, but that over a year later, the company discloses that battery packs for the Model 3 are limiting production?
As for software systems and development, however, we are moving far beyond the realm of just whether a company is "meeting its targets" given the risks to both people and objects from inadequately developed and tested systems. Additionally, in the "fast and loose" environment that now seems to be the "new norm" at most companies, such inadequately developed functions and systems are then rushed to market.
I do not want to sensationalize the recent UBER incident by providing any links to information about that, but for anyone who finds any of the video that I have unfortunately now seen, I think a lot of people will conclude that the world needs to turn back from the headlong development and deployment of anything and everything into a more thoughtful and safer approach to only introducing and deploying things when there is certainty and confidence that things will work in a safe and intended way.
Mark Zuckerberg has also now said essentially the same thing after the recent issues concerning Facebook have been more fully disclosed. He has said that "some regulation is probably appropriate", but he also said that he wasn't sure yet what that would be. From my own perspective, now that I have seen a "self-driving car" run over two objects that it should have recognized and reacted to, I believe that the visceral reaction to such an event will be "whoa, we really need to slow this down and establish rigorous testing and deployment requirements."
Possible implications for Tesla's stock
Tesla's current stock valuation is a mystery to many people, and so I don't think it is useful to get into a "valuation" discussion in general. But I do thoroughly believe that a significant part of the stock's valuation is based on unquestioning acceptance about how Tesla will dominate every product segment it claims to be addressing whether "self-driving cars" or "ride hailing networks" or "$35,000 Mass Market vehicles" and so forth.
That unquestioning acceptance doesn't seem to ever recognize that all of these things take much longer to implement and are much more difficult to implement than Elon Musk is ever willing to tell them (although he still provides many ongoing examples of the reality of that!). From the recent incident in Arizona, in addition to Tesla never seeming to be able to implement things in any reasonable time frame relative to Musk's ongoing statements, there will now probably be increasing regulatory constraints to such developments and implementation of such systems.
From my own software development and testing experience, in my opinion, it will be at least three years before "self-driving" systems can be thoroughly developed and tested for incorporating all of the possibilities and inputs that such systems would have to handle in a very dynamic and complex world. There are two problems with such a time-frame for Tesla, however. The first is that such a time frame is well-behind Musk's public statements (whose unquestioning acceptance helps prop up the stock!). The second is that LOTS of other entities are also developing such systems!
The fact that lots of other entities are also developing such systems, which will ultimately be able to be purchased "off the shelf" from component and systems suppliers such as Bosch, Continental (OTCPK:CTTAF), Delphi (NYSE:DLPH), and Magna (NYSE:MGA), also effectively invalidates the "Tesla has a lead on everything" narrative that is so frequently quoted like a mantra by Tesla supporters.
Conclusion
As I have described, I think the recent UBER incident in Arizona will significantly slow the development and deployment of "self-driving systems" by all industry participants. Not that I am really a fan of "government regulation" but I also think that the visceral reaction to seeing a pedestrian mowed over by a self-driving vehicle will result in some sort of more restrictive development and testing environment that will significantly slow the development of such functions.
But I would guess that at least 20 percent of Tesla's current valuation is based on the assumption that Tesla both has the lead in and will dominate such functions. There was also a truly crazy research study (in my opinion!) cited recently that assumed that "trillions of dollars of market capitalization" would be created by "only five companies" that will dominate such functions!
From someone with both software development and investment research experience, I have a much more conservative perspective about all of this. From such experience, I believe that such developments will take far longer than anyone anticipates (not a surprise with any of Elon Musk's announcements) and will turn out very differently than the unquestioning assumption "that everything Elon says will come true." Both of those things occurring will create significant risks for Tesla shareholders, in my opinion.
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.
Additional disclosure: This article expresses the author's opinions and perspectives about various investment related topics. Since all statements in the article are represented as opinions, rather than facts, such opinions are not a recommendation to buy or sell a security. My own investment position described in the disclosures is not intended to provide investment advice or a recommendation of a specific investment strategy but is a required disclosure item by Seeking Alpha. My own investment position may have been initiated at very different price levels than current prices levels and so that is also why my disclosed position is definitely not intended as an investment recommendation. All investors should also do their own research before making any investment decision.