Economic Commodity Debate: The Case of AI Self-Driving Cars


By Lance Eliot, the AI Trends Insider

Will AI self-driving cars be an economic commodity?

That’s the scuttlebutt these days that some are predicting. There seems to be a rumor circulating which asserts that once we actually achieve true AI self-driving cars, it will be a commoditized business. I’ve had both industry outsiders and insiders ask me about this matter, doing so at the conferences I’ve been speaking at, and even at local events and collegial dinners. The rumor appears to be gaining speed.

I’d like to tackle the argument that precipitated that kind of assertion.

Bottom-line: I don’t buy into the commoditization logic and nor rhetoric.

In my view, those that seem to be fostering such an indication are regrettably somewhat misguided in their logic and have perhaps fallen into some logic-traps that would hasten their belief in the notion. There are others that are going along with the notion since it sounds alluring and the surface-level rationale appears to be viable at a quick glance. My herein assessment at a deeper level suggests that the commoditized-camp and its converts might want to reconsider their stance.

Of course, we don’t any of us know for sure what the future will bear out, but I’d be willing to wager that AI self-driving cars are not likely to become commodities.

First, let’s start the discussion by considering what an economic commodity is. We need to all somewhat agree on the definition of an economic commodity so that we can then try to figure out whether true AI self-driving cars will become one. If we cannot agree to a definition, we’ll be arguing past each other and never see eye-to-eye.

Generally, an economic commodity is considered any economic good or service for which there is an abundance of fungibility, which means that instances of the good or service are essentially indistinguishable from each other and buyers do not perceive any material distinctions about the good or service. Notice that I’ve mentioned that it can be either a good or service, though by-and-large it is usually a good or product that gets into the commoditization category rather than a service.

Why not services, you might ask? When offering a service, there is frequently a greater likelihood that the service being offered is differentiated, including how and who offers it, and thus they will be distinguishable in the minds of buyers.

I might need a plumber to come over to my house to deal with my leaking faucet. The Ace Plumbers service provides so-called smell-good plumbers that will show-up on-time and do a good job for a reasonable price. The Zany Plumbers service has plumbers that are typically late to arrive, they often fail to fix the plumbing problem, and they are dirty and gruff to deal with and charge an arm and a leg to do the plumbing work. It is relatively apparent that we could distinguish these plumbing service providers from each other, therefore we might say that the plumbing service is not commoditized.

In terms of goods, I’m sure that you likely know that often times products such as gasoline, sugar, grains, are referred to as commodities. They are called commodities because presumably the buyers of those products do not see any material factors that differentiates them (by this I mean that gasoline is not differentiated from other gasolines, sugar is not differentiated from other sugars, etc.). These commodities are often sold in bulk to those in the supply chain that will ultimately get them to the consumer. This tends to squeeze out profit margins because there is no particular factor that makes any better or worse than the other.

Given that they are otherwise considered equals, the providers of those commodities are fully at the whim of the marketplace in terms of pricing and they cannot try to adjust their price higher based on their product per se. If I’m selling a thousand gallons of gasoline in bulk, and someone else is selling a thousand gallons of gasoline in bulk, and if there is no differentiation between our gasoline product, it then comes down to price. I cannot try to have a higher price in this circumstance since the buyer of the bulk gasoline would be unwise to buy from me at a higher price when they can get the same thing from someone with a lower price.

Dealing in commodities can be a rough business. Dog eat dog, so to speak. I am reminded of the famous line by Karl Marx, in which he stated: “From the taste of wheat, it is not possible to tell who produced it, a Russian serf, a French peasant or an English capitalist” (for this quote and other economic theories, see the famous book “Capital: The Process of Circulation of Capital,” published in 1885 by Fridrich Engels and with Karl Marx’s notes, which provides the core elements of Karl Marx’s economic theories).

In short, commodities are indistinguishable in the minds of buyers. This implies that the product or service can be offered by multiple sellers and yet the buyer does not recognize any differences about the product or service being sold by those sellers. The buyer can presumably choose any of the sellers and will be buying the same product or service. If that’s the case, it becomes a war on pricing among the sellers. Whomever can sell the product or service at the lowest feasible price would garner the buyers to buy from them.

Keep in mind that there are several caveats about this notion of commoditization.

The buyer has to be savvy enough to realize that the product or service is indeed a commodity. Often times, a buyer might not know this and will therefore proceed to pay a higher price for something that in theory they have overpaid on price.

Experiencing Ace Plumbers vs Zany Plumbers

The faucet in my house is running wildly and water is flowing out everywhere. Yikes, I’m in a panic. I need to get a plumber pronto. I do a quick online search and up pops the Ace Plumbers and the Zany Plumbers. As far as I know, they are both equal. I can get the same plumbing services from either one. I look quickly to see which ones has a lower price. The pricing posted by Zany Plumbers is lower than the Ace Plumbers and so I contact them and arrange for them to come to my house.

Little did I realize that the Zany Plumbers are in fact differentiable from the Ace Plumbers, recall my earlier description of the two plumbing services in which I depicted that the Ace Plumbers were more professional and the Zany Plumbers much less so. The plumber that shows up from Zany Plumbers is a miserable, grumpy, ill-prepared dolt and fails to fix my faucet properly. I did not realize beforehand that the two plumbing services were not commoditized.

Thus, the buyer needs to be astute enough to know when a product or service is indeed commoditized:

If the buyer fails to realize something is a commodity, the buyer might be fooled or lulled into paying a higher price when there is presumably no justification in doing so.

If the buyer believes something is a commodity when in fact it is not (it is non-commoditized), the buyer might buy at a price that seems equal to the other sellers pricing, but in the end get ripped-off by having gotten a lesser product or service than they could have gotten from another seller.

In a marketplace of goods and services, it can be difficult for a buyer to know what a commodity is and what is not a commodity.

There are sellers that might try to have the buyer believe there is a commodity in that good or service, when there actually is not. The Zany Plumbers probably wants you to think that all plumbing services are the same, hoping that you will pick them, and they are either going to try to confuse you into believing that all plumbing services are the same or they anticipate you won’t research well enough to know.

This reminds me of the bottled water industry. I knew a top-level executive at one of the largest bottled water firms. He told me over dinner one night that he could not believe that people think that there is any difference between the bottled water products. It is clean water in a plastic bottle. He claimed that the water is really about the same no matter which bottled water you pick. The core product, if you consider it to be the water, he expressed was really a commodity.

To try and get buyers to believe it is a non-commodity, the bottled water company shaped the bottle in a distinctive manner and slapped flashy graphics and colors onto the exterior of the bottle. They also boldly indicated that their water had only a miniscule amount of a certain kind of iron in it, but he told me that all of the bottled waters have about the same amount. Furthermore, he indicated that the iron content really didn’t matter because it wasn’t a health issue and had nothing to do with any kind of difference in taste.

In that sense, they were trying to make a commodity seem like a non-commodity. Why? Because it would allow them to convince buyers into buying their product over the other sellers, and possibly also pay a higher price when doing so. If the bottled waters are really a commodity, the buyers should be looking only at price and ignore the shape of the bottle and its flashy exterior.

For those of you that are in the bottled water business, please don’t write me to complain that I’ve suggested that the bottled water business is a commodity industry. There are a lot of businesses that the same argument can be made that the core product or service is really in a sense a commodity, and yet the sellers in that industry try to package or portray their product or service as a non-commodity, inducing buyers to believe that the products or services are not commoditized. I am not singling out bottled water and nor am I claiming that bottled water is a commodity – I hope that will calm the nerves of those of you in that industry.

Let’s take the perspective of the seller for a moment.

If you can do so, as a seller of a product or service, you’d likely prefer to offer a non-commodity product or service. This would allow you to then charge perhaps more for your product or service than others, since buyers might be willing to pay more, assuming that they perceive your product or service differences as valued.

What can be even more galling is to be a seller in a non-commoditized product or service wherein the buyers don’t realize that it is not a commodity. Imagine the plight of the Ace Plumbers. They know that they go to the trouble to have nice smelling plumbers that are professional. Perhaps this costs them more to arrange. They charge the same price as say Zany Plumbers. But, buyers that don’t know are fooled or lulled into assuming that Ace Plumbers and Zany Plumbers are indistinguishable in terms of the plumbing service, so buyers go to Zany Plumbers at times when they might have been better off to go to Ace Plumbers.

Suppose indeed that Zany Plumbers pricing is the same as Ace Plumbers. This implies that the buyer is overpaying for what they could get from Ace Plumbers. Unfortunately, if buyers don’t realize this, it is possible that Ace Plumbers is losing business to Zany Plumbers, and as a double whammy the Ace Plumbers has higher costs of offering their plumbing service. It’s not fair! Those darned buyers ought to be more differentiating in their buying.

Here’s the mess we have then about buyers and sellers:

  •         Buyers can be ill-informed and therefore believe a commodity is a non-commodity, or they can believe a non-commodity is a commodity.
  •         Sellers might be tempted to make a commodity seem like a non-commodity, or make a non-commodity seem like a commodity.

I mention these aspects about buyer and seller perceptions and psychology because it is vital for the notion of commodities. In theory, one might argue that all buyers are always all knowing, and all sellers are always all knowing, but that’s not necessarily the case. In economics, there is an entire ongoing area of debate and research about the rational and irrational behavior of humans, which plays into how the real-world works in terms of the selling and buying of goods and services.

For my discussion about rational versus irrational human behavior, see my article:

For the role of greed in human behavior, see my article:

For how curiosity impacts human behavior, see my article:

For how group behavior impacts human judgement, see my article:

What does this have to do with AI self-driving cars?

At the Cybernetic AI Self-Driving Car Institute, we are developing AI software for self-driving cars. As mentioned earlier, there are some rumors floating around that inevitably the industry of AI self-driving cars will be a commodity.

In essence, it has been postulated that AI self-driving cars will be commoditized and so the buyers of AI self-driving cars will presumably consider any AI self-driving car to be indistinguishable from any other AI self-driving cars. There will be no brand or feature or other facet or factor that would cause the buyers of AI self-driving cars to consider one to be better than or worse than any other AI self-driving car.

Why would that matter?

If such a prediction is correct, it implies that ultimately the AI self-driving car marketplace will be solely determined by price. Buyers of AI self-driving cars will merely look at the price to ascertain which AI self-driving car to buy. This also means that the auto makers and tech firms that are selling AI self-driving cars are competing solely on price. As I mentioned before, price alone is a dog eat dog world.

Within the auto makers and tech firms, if they were in a commoditized market space, the only thing they can presumably do is try to cut their costs as low as possible, which then becomes their only means of competition. I say this under the assumption that if commoditized, they are unable to differentiate their AI self-driving cars via any other means, since by definition we’re saying that there is no difference, i.e., they all have the same features and capabilities.

One would also assume that if commoditized, the price of AI self-driving cars will all end-up around the same price. There is no value to a buyer that pays a higher price for an AI self-driving car from one auto maker versus another. To sell their AI self-driving cars, an auto maker or tech firm would need to maintain a price close to the rest of the marketplace. Presumably, the marketplace will reach a point of a pricing that provides some economic incentive for the auto makers and tech firms to stay in the market, but only such that the price is kept under pressure because all the other auto makers and tech firms are vying on price too.

Varying Levels of Self-Driving Cars

Before I jump into the fray and explain why the above scenario about the commoditizing of AI self-driving cars is highly unlikely, I’d like to provide some helpful and instrumental background about AI self-driving cars that is essential to this discussion.

I’d like to first clarify and introduce the notion that there are varying levels of AI self-driving cars. The topmost level is considered Level 5. A Level 5 self-driving car is one that is being driven by the AI and there is no human driver involved. For the design of Level 5 self-driving cars, the auto makers are even removing the gas pedal, brake pedal, and steering wheel, since those are contraptions used by human drivers. The Level 5 self-driving car is not being driven by a human and nor is there an expectation that a human driver will be present in the self-driving car. It’s all on the shoulders of the AI to drive the car.

For self-driving cars less than a Level 5, there must be a human driver present in the car. The human driver is currently considered the responsible party for the acts of the car. The AI and the human driver are co-sharing the driving task. In spite of this co-sharing, the human is supposed to remain fully immersed into the driving task and be ready at all times to perform the driving task. I’ve repeatedly warned about the dangers of this co-sharing arrangement and predicted it will produce many untoward results.

For my overall framework about AI self-driving cars, see my article:

For the levels of self-driving cars, see my article:

For why AI Level 5 self-driving cars are like a moonshot, see my article:

For the dangers of co-sharing the driving task, see my article:

Let’s focus herein on the true Level 5 self-driving car. Much of the comments apply to the less than Level 5 self-driving cars too, but the fully autonomous AI self-driving car will receive the most attention in this discussion.

Here’s the usual steps involved in the AI driving task:

  • Sensor data collection and interpretation
  • Sensor fusion
  • Virtual world model updating
  • AI action planning
  • Car controls command issuance

Another key aspect of AI self-driving cars is that they will be driving on our roadways in the midst of human driven cars too. There are some pundits of AI self-driving cars that continually refer to a Utopian world in which there are only AI self-driving cars on the public roads. Currently there are about 250+ million conventional cars in the United States alone, and those cars are not going to magically disappear or become true Level 5 AI self-driving cars overnight.

Indeed, the use of human driven cars will last for many years, likely many decades, and the advent of AI self-driving cars will occur while there are still human driven cars on the roads. This is a crucial point since this means that the AI of self-driving cars needs to be able to contend with not just other AI self-driving cars, but also contend with human driven cars. It is easy to envision a simplistic and rather unrealistic world in which all AI self-driving cars are politely interacting with each other and being civil about roadway interactions. That’s not what is going to be happening for the foreseeable future. AI self-driving cars and human driven cars will need to be able to cope with each other.

For my article about the grand convergence that has led us to this moment in time, see:

See my article about the ethical dilemmas facing AI self-driving cars:

For potential regulations about AI self-driving cars, see my article:

For my predictions about AI self-driving cars for the 2020s, 2030s, and 2040s, see my article:

Returning to the discussion about whether or not AI self-driving cars will become a commodity, I’d now like to tackle the question head-on.

Monolith Myth

I’ll start with the monolith myth.

You might be aware of Arthur C. Clarke’s famous quote or law: “Any sufficiently advanced technology is indistinguishable from magic.”

This quote is contained in Clarke’s famous essay “Hazards of Prophecy: The Failure of Imagination” published in 1962 and included his so-called three laws, which by the way he jokingly kept to three laws since Sir Isaac Newton had three laws and it seemed that if Newton only needed three then Clarke figured he needed only three.

In any case, the reason I offer Clark’s quote is that if you take the viewpoint that all AI is essentially the same, and you don’t differentiate any differences between AI elements, you are apt to make the assertion that ultimately AI self-driving cars will be a commodity.

Consider toasters for a moment. I might claim that all toasters are the same. Magically, they are able to make toast for me. I don’t necessarily know how it works. All I know is that a toaster makes me toast. As such, are we to assume that toasters are a commodity? Are all toasters indistinguishable from each other?

If you say that yes, toasters are a commodity, I’d like to point out that it is by some accounts a billion dollar global market with at least 20 major players and they seem to believe that they are selling a non-commodity. Pricing ranges quite a bit. Features range quite a bit.

When I tried to buy a toaster the other day, I was nearly overwhelmed with my options. Toast two pieces of bread or three? Toast bread and also hotdog buns? Old-fashioned timer or modern-day IoT (Internet of Things) that connects with your smartphone or smartwatch? Anti-jam capability or no anti-jam? Easy to clean or do you have to take apart the toaster with a screw driver just to keep it clean?

Before I tried to buy a toaster, I assumed that it was a monolith in the sense that no toaster could be distinguishable from any other toaster. All toasters must be the same, I assumed.

I believe that some of those arguing that AI self-driving cars will be a commodity market are assuming that all AI self-driving cars will be the same.

This is an easy mental trap to fall into.

If you aren’t aware of what the AI is going to do for AI self-driving cars, you can readily make a false assumption that the AI will be the same for all AI self-driving cars.

If you aren’t considering the fact that the AI is being developed by different auto makers and tech firms, and therefore the nature of the AI and what it does is going to be is different across those auto makers and tech firms, you might falsely think that all of the AI self-driving cars will have the same features.

Let’s consider the nature of Level 5 AI self-driving cars. I am choosing those since they are the true AI self-driving cars in the sense of being presumably fully autonomous.

Anything less than a Level 5 AI self-driving car is essentially axiomatically a non-commodity. I don’t think anyone can especially debate that point. I say this because at the less than Level 5, the auto makers and tech firms can choose what kinds of AI features they want to include into their AI self-driving car. This is similar to today’s cars that have a variety of advanced auto driving features. One auto maker offers auto driving for the brakes, while another might not. One auto maker offers lane correction for their cruise control, another might not. And so on.

This will continue for Level 2, Level 3, and Level 4 of AI self-driving cars.

Indeed, for Level 4 self-driving cars, the auto makers and tech firms can define their own ODD (Operational Design Domains). One auto maker might indicate that their AI self-driving car works in the rain but not the snow, and so they define an ODD for that aspect. Another auto maker might indicate their AI self-driving works in the rain and the snow, so they define an ODD for that aspect.

You, the buyer of an AI self-driving car, are trying to decide which of the Level 4 AI self-driving cars to buy. One that has an ODD that includes rain but not snow, and the other that handles both rain and snow. Which do you pick?

Certainly, it is clear cut that you are differentiating the two offerings.

If you are the type of person that lives in a place that gets snow, I’d bet that you’d buy the AI self-driving car that has the ODD that includes snow driving. For someone like me that lives in Los Angeles (LA), hint we don’t get snow in downtown LA, I’d probably opt to get the AI self-driving car that doesn’t have the snow driving capability (under the assumption that the one being offered that cannot handle the snow is likely a lower price or that the price somehow encompasses the lack of snow driving in some manner or another).

Okay, so we need to then only concentrate on the Level 5 AI self-driving cars, since we are agreed (hopefully) that anything less than a Level 5 can be differentiated due to the aspect that the auto makers can choose to include whatever AI features they wish to include or have available to include (and furthermore that not all auto makers will have the same capabilities to include, with some having some features ready to go and others not).

Once we get to true AI self-driving cars, Level 5, will they all embody the same AI capabilities?

That’s an easy answer: No.

An auto maker or tech firm can have a true Level 5 AI self-driving car that has different AI features than those of some other Level 5 AI self-driving car made by another auto maker or tech firm.

Suppose I am selling an AI self-driving car that can indeed drive you around without any human driver needed. It is a Level 5 AI self-driving car. Fully meets the SAE (Society of Automotive Engineers) definition for a Level 5.

My AI self-driving car has a robust Natural Language Processing (NLP) feature that allows the AI to interact with the occupants of the self-driving car. It not only asks where you want to go, it also carries on various conversations. It is keeping track of where you drive to and is able to anticipate what you might want to do. Would you like me to drive you to the grocery, since it is Tuesday afternoon and you usually seem to go shopping on that day and time?

Meanwhile, another auto maker or tech firm is selling an AI self-driving car, but it does not have the same nifty advanced NLP capabilities. It merely finds out where you want to go. It does not particularly interact with you. It does not keep track of where you’ve been going to. Please realize that it is still in all key respects a Level 5 AI self-driving car. There is no stated requirement that it must have the advanced NLP that my AI self-driving car has.

If you were going to buy a Level 5 AI self-driving car, which would you buy – the one I am selling that has the advanced NLP or the one offered by another auto maker or tech firm that does not have the advanced NLP?

I’d dare say that you are making your decision based on a differentiation among the Level 5 AI self-driving cars capabilities. If that’s the case, you are buying a non-commodity.

For my article about NLP and AI self-driving cars, see:

For more details about the levels of AI self-driving cars, see my article:

For the socio-behavioral aspects of AI self-driving cars, see my article:

For explanation-based AI and self-driving cars, see my article:

In other words, there are going to be lots and lots of AI features involved in AI self-driving cars that are beyond those needed to achieve the baseline of a true Level 5 AI self-driving car. The auto makers and tech firms will readily be adding into their Level 5 AI self-driving cars a variety of such features, doing so to purposely differentiate their product from others in the marketplace.

In case you want to try and argue that the NLP example is maybe “inconsequential” as a differentiating feature, I’ll be happy to take you on about that. I think it is nearly self-evident that we see today how much people like Alexa and Siri and that it is a differentiating feature in terms of devices that can talk versus not talk. Furthermore, those that use Alexa and Siri can readily indicate to you that one of them is superior in terms of NLP over the other.

I’d say that if all else is equal about two Level 5 AI self-driving cars other than one has the advanced NLP and one does not, there would be buyers that would value the advanced NLP. In that sense, I don’t think you can make the case for the NLP being inconsequential. I’d say it is at least consequential, and will cause some buyers to differentiate between those Level 5 AI self-driving cars that have it or do not have it. I’m not saying that all buyers will think it consequential, and indeed there are bound to be some buyers that don’t care about whether there is an advanced NLP or not.

If you are now leaning toward agreeing that maybe the advanced NLP might be a differentiator, but perhaps you are now wanting to argue that it is just one factor and I am hanging onto a thin thread, well, in that case, let’s discuss other such differing capabilities.

Suppose we have one Level 5 AI self-driving car that has AI features that provide a smooth ride. When the road surface is detected as rough, the AI is able to maneuver the self-driving car in a manner that offers a less bumpy ride than you might otherwise have gotten. This involves the AI detecting that a rough road exists, it involves doing AI action planning to go slowly and steer the car around potholes and rough patches. Etc.

If you were faced with buying one AI self-driving car of a Level 5 versus another, and you were informed that one had the AI-enabled smooth ride feature and the other did not, would this make a difference? Maybe you don’t care about having a smooth ride and so you might say it isn’t important to you. I’d bet though that a lot of other people would relish the added AI feature that offers a smoother ride. This would be a differentiator.

One of the Level 5 AI self-driving cars offers 5G connectivity, another one does not. The use of 5G makes the time needed to do an OTA (Over-the-Air) connection with the cloud of the auto maker or tech firm much faster, allowing you to readily download the latest patches and updates into your AI self-driving car. Other auto makers and tech firms aren’t including 5G. Which would you want, the faster connecting self-driving car or the slower transmitting one? Again, a differentiator.

I can go on nearly endlessly with such capabilities.

For more about OTA, see my article:

For my article about 5G, see:

For my article about edge problems of AI self-driving cars, see:

For different kinds of AI self-driving car driving styles, see my article:

Some of the features will be directly related to the driving task, such as the smooth ride AI-enabled feature. Some of the features will be indirectly related to the driving task, such as the 5G connectivity. You might argue that the advanced NLP is not strictly speaking related to the driving task per se, though it does have to do with the overall utility of the AI self-driving car. Anyway, I’m certainly willing to say that there are some AI-related features that perhaps have little to do with the driving task and yet will nonetheless provide a differentiating aspect of a Level 5 AI self-driving car from another one.

This then addresses the monolith myth.

In recap, it will not be the case that all true Level 5 AI self-driving cars will be the same. They will differ on features that are involved in the driving task, and other features that are related to the driving task but not directly so, and even other features that are tangentially related to the driving task but nonetheless likely valued by buyers of AI self-driving cars.

That’s a non-commodity.

Spontaneous Features Myth

So far, I’ve tried to make the case for the aspect that even once we reach the vaunted Level 5 AI self-driving cars, they will be differentiable by their features, and thus it is not the case that all AI self-driving cars are the same and indifferentiable.

When I get to that point in my discussion, I sometimes will have someone stop me and offer the thought that Okay, it might be the case that there will be some differences initially, but eventually things will flatten out and all AI self-driving cars will be the same.

This is the spontaneous features myth.

Let’s get back to my toaster conundrum. The first electric bread toaster is credited with being invented by Alan MacMasters in 1893. You can look it up.

Suppose we go back in time to the 1890’s and early 1900s. Toasters are starting to be sold. They initially had a difficult time trying to perfect the heating elements. This meant that your toast might get overheated or burnt. In 1905, Albert Marsh came up with a new kind of heating element that was better than what existed for toasters at the time. Less chances of burning your toast!

Some toasters included Marsh’s version of the heating element, other toasters did not. Was the better heating element by Marsh a differentiator between various toasters? Yes, definitely.

In 1921, Charles Strite invented a pop-up mechanism that would automatically get your toast to be pushed up and it was easier for you to get your toasted bread. Was this a differentiator between various toasters? Yes, definitely.

Why am I telling you about these toaster advancements? Because this went on and on. Starting in 1893, we have had improvements in toasters over the years. It is a continual cat-and-mouse game. Some new feature is invented, and some toasters have it and some do not.  At any given point in time, there will be features that some toasters have, and others do not. It is of course likely that over time some features will creep into the core aspects, such as the automatic pop-up contraption. Meanwhile, something else new will be invented.

This has been going on with toasters since 1893 — that’s nearly one hundred and thirty years long!

In terms of true Level 5 AI self-driving cars, do you really think that it will come a day when they all have the same exact features? Really?

If toasters have kept changing over the 100+ years, do you think it makes any sense to suggest that AI self-driving cars, which will be one of the most advanced forms of technology ever made by mankind, they will simply all end-up with the same features?

I don’t see how any reasonable person could believe that the auto makers and tech firms will not be continually battling over coming up with new AI features for self-driving cars.

Here’s something for you as an example. I’ve mentioned in my writing and in my speeches about the dangers of humans being able to potentially prank AI self-driving cars. This consists of tricking an AI self-driving car into doing something because you know how the AI self-driving car will react. For example, if an AI self-driving car is setup to come to a stop the moment that it detects a pedestrian stepping into the street, you can pretty much guess that some pedestrians will purposely step into the street since they know the AI will force the self-driving car to come to a halt.

Today, we as humans have to gauge whether another human is going to let us step off the curb. It is the pedestrian eyeballing the human driver. There are risks involved in this dangerous game. Maybe the human driver will stop, maybe not. In the future, with AI self-driving cars, the human pedestrian will be able to potentially trick the AI into halting, simply by the human making a motion as though they are going to step into the street, whether they really are intending to go across the street or not. Maybe they really do want to go across the street, or maybe that just find joy in tricking the AI into stopping the self-driving car.

For my article about the pranking of AI self-driving cars, see:

For the burnout of AI developers, see my article:

For the idealism that some have about AI self-driving cars, see my article:

For the egocentric views of some AI developers, see my article:

A true Level 5 AI self-driving car does not have any stated requirements about dealing with pranking. Indeed, I’d say that most AI developers and auto makers and tech firms are not even considering what to do about pranking. They are so emmeshed in just getting to a Level 5 AI self-driving car that the pranking matter is not on their radar.

I mention the pranking because we’ve been working on an anti-pranking solution. This is a feature that I predict will ultimately be of value to AI self-driving cars. Right now, no one really cares about it. Once we have actual AI self-driving cars on our roads, and once people figure out how to prank those AI self-driving cars, I’d bet that there will be an uproar about what to do about it.

My point is that this is a feature that probably won’t enter into the mosaic of AI self-driving cars features for a very long time. When it starts to get introduced, I’d predict that it will begin with some simple kinds of anti-pranking capabilities. Over time, they will become more robust.

The key is that it won’t be as though all AI self-driving cars suddenly are outfitted with an anti-pranking feature. Furthermore, those that have such a feature might have a simple version, while other AI self-driving cars have a more complex one.

I know some that have said that with OTA, all AI self-driving cars can spontaneously all have the same features.

Magically, someone is somehow going to get all of the differing auto makers and tech firms to agree to a certain kind of feature, and get all of them to adopt it into their own proprietary AI systems of their self-driving cars, and then get them to all have it downloaded into their fleet of AI self-driving cars, doing so for all AI self-driving cars of all brands and makes and models, across the globe.


How did we go from a marketplace of firms that are fiercely competing with each and each is wanting to differentiate their product from everyone else’s, and now they are suddenly all cooperating and fully agreeing to have the same exact features and to implement those features in the same exact way on the same day and time?

I can only imagine this occurring if somehow the government stepped into the AI self-driving car marketplace in a really heavy-handed way and somehow mandated this kind of all-features, all-the-same, for everyone at the same time, kind of approach. Doesn’t seem very realistic to me. This would be an incredibly altering force to our economy and to a rather massive industry. Seems unlikely.

For aspects about federal regulations and AI self-driving cars, see my article:

For state and local municipality related regulations, see my article:

For product liability aspects of AI self-driving cars, see my article:

For the crossing of the Rubicon about AI self-driving cars, see my article:

For more about the coopetition among AI self-driving car makers, see my article:

Homogeneous Cars Myth

I believe that I have now generally refuted the notion of AI self-driving cars possibly becoming a commodity, which I’ve tried to make the case they are a non-commodity by pointing out that they will not magically converge onto the same set of AI features and that instead there will be an ongoing and at times bruising form of competition to continually come out with new features.

Let’s suppose that somehow the AI features all do converge in the sense that all AI self-driving cars have exactly the same AI features and do so at the same points in time. This is utterly unimaginable in any reasonable stretch of the imagination.

But, suppose as a thought experiment we entertain this rather farfetched notion and do so to see if we can help the commodity believers.

Would we not differentiate the AI self-driving cars on other factors beyond the AI features?

Today, the car industry markets us their cars based on the color of the car, the shape of the car, the overall look of the car, and so on. Why do people buy one kind of car over another?

Many of today’s cars have pretty much the same driving capabilities.

You would be hard pressed when comparing some models of cars to make the case that the reason a buyer buys one versus the other is due to the driving capabilities. Instead, it is more like the feel of the seats, the interior upholstery, the cup holders, and other non-driving related features.

Indeed, the automotive industry spends millions upon millions of dollars to study the psyche of car buyers and then tries to design the shape and look of their cars to attract those buyers. There is nearly as much effort put into the actual driving mechanisms as there is in the look-and-feel of the car.

I would assert that an AI self-driving car is still a car. By this, I mean that people will still buy a car based what a car looks like and what the interior is like, and so on. Set aside the actual driving capabilities and focus for a moment on all of the non-driving aspects.

Some commodity believers seem to think that we are going to end-up with homogenous cars.

In this theory, all cars, and let’s say that we end-up ultimately with all and only AI self-driving cars, will be the same look, the same shape, have the same interiors, etc.


If you look at the futuristic designs by many of the auto makers, you can readily see they are diverging already in terms of what the size of their future cars might be and what the interiors might be like. Why would the auto makers decide to merely copycat each other in terms of the look-and-feel of their cars? It doesn’t seem to hold water. It is without any reasonable logic.

I know that there are sketches made by some futurists and animators that try to portray a world in which we all are riding in the same identical looking AI self-driving cars. I think this is the same as the idea that we will all dress alike in the future. We will all wear the same jump suit, universally available, which simplifies life. It also wipes out the fashion industry.

When I see such homogenous futuristic how-we-will-live predictions, it reminds me of the Dr. Seuss story that I used to read to my children when they were young, telling the tale of the Sneetches. Do you know the stars upon thars story? Please take a look at it. The point of my bringing it up is that even if we all were issued the same looking jump suits in some kind of strange Big Brother kind of future, I would bet that some people would want to add something to their jump suit to make it look different. It is human nature, I claim.

Are we going to end-up with AI self-driving cars that have the same identical look-and-feel, and for which from the outside they look exactly the same, and while inside they are exactly the same in terms of the layout and interior capability?

Hardly likely. I’d say straight out no.

There will be a variety of differing interiors. I’ve predicted, and others have too that we’ll have swivel seats inside these AI self-driving cars and be able to swivel around and face our fellow passengers. Some might have seats that can become beds, allowing people to sleep while in their AI self-driving car. There will be a rich variety of ways in which the interiors will be designed and provided.

I’d argue that same is true about the exteriors of AI self-driving cars.

I know that some commodity believers would say that people won’t care about the exteriors, since they will only have in mind that the self-driving car is intended to get them from point A to point B. They presumably won’t care what the exterior looks like. I’m not so convinced that we are going to give up our imagery of the status of different kinds of cars, even in a world that becomes saturated with AI self-driving cars.

In any case, I doubt that you can quite make the same case about the interiors. Certainly, most would agree that the interior will become important, perhaps even more so than with today’s cars. If you are going to be riding in self-driving cars, perhaps doing so a lot more than you ride in conventional cars today, I’d bet that you would care quite a bit about the interior of the self-driving car. In our shift and transformation to a mobile economy, wherein we will readily have available AI self-driving cars on a non-stop 24×7 basis, we might get finicky about the interiors.

If you are going to counter-argue that all interiors will end-up the same, I don’t buy into that suggestion. In a sense, if that were true, wouldn’t all interiors today be the same? They are not. People like different kinds of interiors and there is no one universally accepted only-way to arrange the interior of a car. The same will be true with the advent of AI self-driving cars.

As per my earlier points about the AI features, the interior features of AI self-driving cars will come and go, some features appearing and being clamored for by buyers, while other interior design features will tend to die out. Some interior features will be considered core, while other features will be discretionary. There will be an ongoing battle waged by the auto makers and tech firms about the nature of the interiors of their cars. Stars upon thars will rule.

The key takeaway is that we ought to not only consider the AI features as distinguishable, I’d assert that the interiors will be distinguishable, and perhaps the exteriors too (though I’m willing to concede a tad on that aspect, but not much!).

Recall that we earlier defined that a commodity is a product that is fungible and there is nothing that enables the buyer from distinguishing one such product offering from another. I claim that for true Level 5 AI self-driving cars, the AI features will be distinguishable, and furthermore the interior, and possibly too the exterior.

You will readily be able to distinguish one AI self-driving car from another. Deciding which one to buy will involve consideration of those factors.

Riding in one will also be somewhat a choice made about the factors, because I’d anticipate that people will want to ride in an AI self-driving car that best suits their needs.

I might feel safer being in the AI self-driving car that has the anti-prank features, so I purposely specify this when I request a ridesharing service to provide me with an AI self-driving car to get me to work. Or, maybe I like the AI self-driving car that has the all-bed interior and I want to use it to get me to my relatives across the country, but when I am in downtown I want the AI self-driving car that has the seats that allow me to get my work done.

That’s a full suite of factors, therefore a true AI self-driving car would be a non-commodity.

For the marketing of AI self-driving cars, see my article:

For family trips in AI self-driving cars, see my article:

For the future of ridesharing, see my article:

For the non-stop 24×7 of AI self-driving cars, see my article:

No-Services Needed Myth

Let’s next consider a twist that some don’t consider about the future of AI self-driving cars, the services side. I’d like to share with you the “no-services needed” myth that some have about AI self-driving cars.

So far, I’ve focused entirely on the product side of AI self-driving cars in this discussion. It is indeed a product, just like a toaster. If you are buying a toaster, you primarily look solely at the product features. You are less likely to think about the services aspects related to the toaster. Toaster are inexpensive, and you can discard it if the thing breaks. Also, you tend to assume a toaster is going to last a long time before it breaks or falters.

For an AI self-driving car, I’d suggest that there will be a services element that can further potentially differentiate one AI self-driving car from another. In many ways, the services and the product tie together. A car is unlike a toaster. Cars are relatively expensive in comparison to a toaster. There is a lot more “moving parts” for a car and a lot more chance of things going awry with a car.

When you buy a car today, you know that inevitably you will need to do maintenance on that car. It is going to happen. You might think ahead and say to yourself, should I buy a car from auto maker X if they are known for doing a lousy job of service? Maybe you would have second thoughts about buying that car. Perhaps some other auto maker provides a car of a similar nature, but they are known for their incredible service.

I continue to have to remind some of the pundits about AI self-driving cars that a self-driving car is still a car. There are pundits that seem to live in a perfect world in which you get an AI self-driving car and it never breaks down. The car parts never wear out. How would this magically happen?

Let’s be honest about cars. They have all kinds of problems. Sometimes they are a lemon from day one. Sometimes they have issues due to wear and tear. Realistically, there is maintenance needed on cars. There will be maintenance needed on AI self-driving cars. There will be recalls of AI self-driving cars. Etc. I know this is shocking to those that believe in a future in which cars never breakdown and never have any problems. It flies in the face of reality.

For more about recalls of AI self-driving cars, see my article:

For what happens in accidents involving AI self-driving cars, see my article:

For towing of AI self-driving cars, see my article:

For the notion of maybe starting over with AI self-driving cars, see my article:

Suppose that I decide to start my own ridesharing service. I approach several auto makers that have their own brands of AI self-driving cars. Let’s pretend the AI self-driving cars are all about the same (though I’ve mentioned repeatedly herein that I don’t believe that will be the case).

One of those auto makers has a great maintenance program that takes care of the AI self-driving car bumper-to-bumper and they are well-known for keeping their AI self-driving cars in top shape. I’ll call it the Ace Auto company. There’s another auto maker, the Zany Auto company, known for doing a lousy job of maintaining their AI self-driving cars. Their AI self-driving cars are often in the auto shop getting fixed up.

I might say to myself, well, both of their AI self-driving cars are about the same, from a product perspective, but one of them offers a much better service in terms of maintaining the AI self-driving car. I want to keep my ridesharing fleet of AI self-driving cars going as non-stop as possible, since I only make money when those AI self-driving cars are giving paying passengers a ride. I will go with the Ace Auto company.

Aha! I just made a decision based on a differentiation related to the services side of AI self-driving cars.

We now have two ways to differentiate true Level 5 AI self-driving cars, one that is based on the AI self-driving car as a product, and the other based on the AI self-driving car and its associated services aspects.

I believe this further bolsters the non-commodity assertion, namely, an AI self-driving car is not a commodity.


We’ve walked through the reasons why I’m claiming that it seems unlikely that AI self-driving cars will become a commodity.  There are those rumors out there about AI self-driving cars becoming commodities and I appreciate that those having such a belief might well have their own basis or logic for making such an assertion.

I think that many of those commodity believers are unaware of the myths aspects that I’ve tried to delineate herein, which are:

  •         Monolith myth
  •         Spontaneous Features myth
  •         Homogeneous Cars myth
  •         No-Services Needed myth

Is there any reason to worry about whether the future of AI self-driving cars might indeed be one of becoming a commodity?

You might be tempted to wonder if this is all a tempest in a teapot. Maybe it really doesn’t matter if sometime in the future we see that AI self-driving cars have become a commodity.

I believe it does matter.

If those commodity believers are to be believed today, it can dampen the existing effort toward advancing AI and its use for AI self-driving cars. There might be enterprising AI developers that become disillusioned if they think that all AI self-driving cars are going to merely converge onto the same set of features and be a blur of the same repeated things.

Why fight hard now to be innovative if the end result is going to be that nothing seems any different from the other?

I’ve said many times that AI is helping to inspire and advance progress on self-driving cars, and likewise that self-driving cars are helping to inspire and push forward on AI technologies and techniques. There is an important synergy between the advances in AI and the advances related to self-driving cars. The CEO of Apple had likened the effort to produce a true Level 5 AI self-driving car as a moonshot, which I agree with his sentiment. Recall that when we were trying to get to the moon, it also pushed forward advances in technology that I doubt we’d have today or that it would have taken many more years to reach fruition of those technologies (if ever discovered at all).

For the moonshot aspects of AI self-driving cars, see my article:

For the Turing test as it relates to AI self-driving cars, see my article:

For the potential of the so-called singularity, see my article:

For start-ups in the AI self-driving car industry, see my article:

The push toward AI self-driving cars is providing that same kind of impetus.

By spreading the word that AI self-driving cars will ultimately be a commodity, it could regrettably (IMHO) very well cause some auto makers or tech firms to figure it isn’t worth the resources and attention they’ve been putting toward it. In the end, if AI self-driving cars are entirely indistinguishable, presumably the only thing that will matter is price. That’s not much of a motivator for anyone involved in this journey.

I earnestly do not see any reasonable path of having AI self-driving cars becoming the same as say sugar or wheat. Since I earlier opted to quote Karl Marx, let’s end this discussion on his quote that capitalism “brings forth living offspring, or, at the least, lays golden eggs.”

Succinctly stated, AI self-driving cars will be non-commodity of golden eggs.

Copyright 2018 Dr. Lance Eliot

This content is originally posted on AI Trends.