Showing results for 
Search instead for 
Did you mean: 
Hagerty Employee

Avoidable Contact #66: The autonomous grift, and how you’re going to pay the price for it

I can’t tell you what the price of oil, bitcoin, or Tesla stock will be 24 hours from now, but I feel absolutely confident in making the following longer-term prediction: You are never going to share the road with a significant number of autonomous vehicles. “Never,” in this case, is a fancy way to say never. Not in the lifetime of anyone reading this column. I am stone-cold certain about this. If you want to contact me and make some sort of bet, I’m your huckleberry—but give me a chance to convince you on the matter first.


Read the full article on

Intermediate Driver

I don't want or need an autonomous car. I wouldn't buy one if they made one. I like driving, which is one reason I like cars. Now, where is my flying car?


Full on automotive automation only works one of two ways:


1) An integrated network where cars, roads, traffic control, and anything else that touches a roadway all communicates with each other.  It needs to be implemented overnight, because a mix of communicating and non-communicating entities will be a disaster (we're already getting a taste with Tesla Auto Pilot and similar systems).  This solution is simpler to implement than...


2) Automotive appliances that are so intelligent they can deal with any situation.  Not only does the AI need to be incredibly good, but the vehicle must be studded with numerous sensors that can withstand rain, snow, sleet, hail, dust, grime, etc. and continue to function perfectly.


Contemplate the cost, commitment, and cooperation required to achieve either goal and the odds don't look good.


P.S. - Jack: How does your son feel about you betting his inheritance, and what do you think a played out spec Miata will be worth in 40 years?


I'm not going to tell him --- and I think a Spec Miata will be worth quite a bit in the future, because virtually all of them will have been totaled at the regional level!


Honestly, these arguments have been heard since "forever" and of course history is littered with a lot more people who declare something would be impossible to see it happen than it is littered with people who died expecting something many people thought would happen soon.  Or, there are both types to be sure, but a lot more of the former.


So how do I take your money with a bet?   Of course, I will win the bet easily because people plan to "cheat" because that's what you do to make a commercial service.  We have your robots that will go to the restaurant and get your order and bring them to you.  We do many thousands of deliveries for paying customers in several towns every day.   No, we don't take your spoken order, because that's stupid, you order online like any 21st century human would.   I don't think Burger King is a partner but they could be.  This is not something from the future.  It's been running for over a year.   So I guess that would not be a very fair bet, would it?


But happy to take other bets.    Arguments about how long things took to develop are nice as analogies but don't actually inform.  Real developers pay no attention to the "levels" like "level 4" but they do indeed have no intention of operating outside a constrained zone because they don't need to to be useful.   They certainly don't need it to be always connected for any safety function -- they are not crazy.


How much of your money can I have?


You're right, of course, but you're also wrong. The system you describe offloads a tremendous amount of the difficult processing onto humans -- not least the effort it takes to start with a food order and rationalize it down into a multiple-choice online order that can be stored as a NoSQL record.  Similarly, a computer can drive a car as well as it can fly an Airbus A380 --- and under similar, easy-to-quantify situations. Natural language processing is a hassle. Context awareness is a hassle. And making decisions on-the-fly with little to no information is a hassle.


Excellent points. I’ve been saying for years that I’ll believe self-driving cars won’t be happen until we first have self-filling gasoline stations. It’s not too complicated for a person to open a gas flap, spin a cap off, pay, pick up the pump, turn the pump on, place the nozzle in the opening, and then the reverse. Trust a robot to do all of this - with different vehicles that operate differently, at least in specifics? I don’t think so.

Advanced Driver

Your prediction is still accurate from my perspective.