From the moment I received my driver’s license until I went away to college, I drove a new car every three months. That included everything from compacts to luxury models. I recall cruising in a brand-new 1990 Mercedes 500 SL convertible one Saturday. The traffic had cars in both directions inching along. A guy in a Ferrari ragtop, moving slowly enough and close enough in the opposite direction that I could have given him a high five, nodded my way and said, “Nice car.” (If you saw what I drive now, you’d know that this sort of exchange will never again happen in my lifetime.)

As absurd as having a new car every 90 days sounds, it was not unusual for those of us who grew up in an automotive household, as I did. My father worked for GM for 31 years. I’m not useful under the hood, but with all that metal coming in and out of our garage, I came to know a few things about the business.

That’s why I feel confident saying that while technology will change many things about the way the industry works, there are cornerstones of this economy that not even Google or Apple will be able to change, at least not in the next couple of decades. With the likes of Benedict Evans pondering the auto industry’s future, here are the dynamics that car guys and gals with a Silicon Valley orientation are missing.

1. Urban driving encompasses a relatively small and unique market: Technologists tend to reside in densely populated urban areas. That can’t help but color their thinking about the problems technology should solve and the opportunity to extend those solutions outside of the city. Overall, though, theirs is a relatively small market. For example, according to 2014 census and DMV data, New York City contains 43% of the state’s population but only 23.5% of its registered vehicles. That means the volume and the money isn’t in the city, but in the suburbs and rural areas.

2. Electric vehicles have fatal shortcomings: People like Evans base their tech-forward assumptions on the electric car’s inevitability. Yes, they may contain fewer moving parts, which in theory should reduce costs, but their limited range requires that the government subsidize their purchase. An all-electric Tesla, for example, tops out at about 295 miles on a single charge, yet it costs up to $105,000. By comparison, a VW TDI clean diesel will go between 594 and 814 miles on a single tank and starts at about $18,000. Yet taxpayers subsidize the Tesla, a car only the super-rich can afford, to the tune of thousands of dollars. This sort of income transfer to the wealthiest is exactly the sort of policy that conservatives and progressives can agree makes for lousy policy. It’s unlikely that the electric car industry can survive their subsidies being wiped out, yet that could happen at any time.

3. Drivers still love to drive: Evans and others also assume that self-driving cars will dominate someday because of the convenience they’ll bring. By leaving the driving to the car, every passenger would be free to watch movies, shop, read, or work. That’s alluring to people who, like many urban dwellers, find the idea of driving so hideous. If I had to regularly endure city traffic, I’d probably feel the same way. But for the majority of drivers who, for the most part, can avoid constant snarls, driving remains something many people enjoy. A self-driving car is going to be a solution in search of a problem, i.e. a non-starter for these consumers.

4. Car dealers are really powerful: Many tech-based automotive futurists believe that we’ll all move to a Dell-like vision of the future. That means they’ll simply go on a manufacturer’s website, order the model they want, and buy or lease without ever browsing a massive showroom or inventory parked on a lot. Embracing the Dell model, though, requires dismantling arguably the most powerful state-wide interest group outside of public sector unions: car dealership associations. They simply won’t allow direct sales at scale.

5. A driverless car inevitably will kill someone: If people still won’t adopt Apple Pay because they fear their information will be compromised, can you imagine how mainstream adoption will suffer once a driverless car inflicts its first casualty?

A small target market, backward economics, an entrenched industry lobby, and consumer preferences all point to one thing: incremental rather than fundamental changes.