Poll: Autonomous vehicle timeline

Investment Analysis Clubs Saul’s Investing Discussions
1

Okay. I am not expecting complete self-driving vehicle anytime sooner, since I expect we need beyond technology, like infrastructure changes and a significant number of cars currently on the road need to be retired etc.

In how many years, you will have 50% of your driving will be automated and 50% manual as standard. The automated 50% includes lane monitoring to lane changing, self-parking, self-driving in freeway (beyond cruise control), adjusting speed to the flow of traffic, etc.

  • 3 years
  • 5 years
  • 7 years
  • more than 7 years

0 voters

2

With cars being connected, and AI advancing, 5 years may be too long.

Today, we are thinking of the AI as one intelligence gathering information as it goes, but with connected cars and smart cities, an intelligence learns instantly all the intelligence from all the other cars and from the city as it goes.

In other words, while you as a driver have to depend on your own senses to avoid me when I am driving like a maniac, a self driving connected car learns from all the other cars that I am driving like a maniac minutes before I get near, even if my car is not connected. If my car is connected, as soon as I start the car it can inform all the other cars that there is a loose nut behind the wheel.

The pieces are falling into place quickly.

Cheers
Qazulight(Trying to avoid new car purchases until 2022)

3 Likes
3

As a poll question, the problem with this poll is the use of the word “you”. Any of us that are driving enthusiasts will respond negatively because we will be hanging on to our 6 speed gear shifts long after the average car is driving automated for a significant percentage of the population … as evidenced that we are still driving manual shift cars many years after most people were driving automatics. A better question would have been to ask when a certain percentage of the population would be driving autonomously … but even that is limited by vehicle replacement cycles, which I don’t know if we know a lot about … but then maybe we do and I just don’t. So, I think the question, re investing, is more about when autonomous driving capability good enough for 50% driving will be a determining factor in X% of new vehicle purchases.

2 Likes
4

With cars being connected, … 5 years may be too long

I am still driving my 1998 Honda Accord. So there are going to enough cars, in fact, the majority of cars on the road may not have any ability to connect. Secondly, when you talk about connectivity, you do realize there are multiple car vendors, and technology vendors for an autonomous vehicle, at least to begin with. How do you expect these vendors to come to an agreement to share data and connect with each other and within 5 years?

I am not expecting connectivity or AI to be the solution immediately. To begin with, you have sensors on the vehicle and the sensor data is used by the onboard computer to make some decisions.

On the AI front, I understand that technology can make a leap, but for AI and machine learning to make such a leap you need a huge amount of data. Are we at a point where the technology vendors or automobile vendors are gathering such data? I don’t think even Tesla is doing that today.

2 Likes
5

the problem with this poll is the use of the word “you”

Those who drive manual are a percentage of the population and their views only represent their share and cannot skew the poll results. Also, remember when there are lots of manual drivers on the road, the automated vehicle has to learn lots of inefficient or bad logic, because humans are inherently flawed or “individuals” on the road and AV has to factor their driving habits to avoid accidents, thus successful.

6

With cars being connected, … 5 years may be too long

I am still driving my 1998 Honda Accord. So there are going to enough cars, in fact, the majority of cars on the road may not have any ability to connect. Secondly, when you talk about connectivity, you do realize there are multiple car vendors, and technology vendors for an autonomous vehicle, at least to begin with. How do you expect these vendors to come to an agreement to share data and connect with each other and within 5 years?

I am not expecting connectivity or AI to be the solution immediately. To begin with, you have sensors on the vehicle and the sensor data is used by the onboard computer to make some decisions.

On the AI front, I understand that technology can make a leap, but for AI and machine learning to make such a leap you need a huge amount of data. Are we at a point where the technology vendors or automobile vendors are gathering such data? I don’t think even Tesla is doing that today.

Tesla is doing this. It is one of the biggest moats that Tesla has.

But AT&T connected car/smart cities incubator has addressed or is addressing all of these concerns. They are working on protocols and connection standards for IOT and cars, traffic lights, speed trap cameras, 911 calls, are all being connected.

Today, and for the past year, several cars are equipped as connected cars. More are on the way and I expect that most will be connected by 2019. It helps manufacturers with quality control and liability. At the same time, the marketing departments are getting customers to pay for it.

I expect that by 2022 there will be enough connected cars on the road to identify your 1998 Honda and even have a personality profile of the human operating it. Lest you think this far fetched, find someone that has never been on Facebook and have them get on. They and you will find that Facebook has already gleaned and built a complete profile on them even though they have never touched a computer.

Cheers
Qazulight (There is a reason Musk is anxious about AI)

9 Likes
7

Those who drive manual are a percentage of the population and their views only represent their share and cannot skew the poll results.

I answered the poll based on what I would do personally not what I think will happen. My next car will be electric for sure. I will probably also wait for autonomous before I buy another car.

1 Like
8

Tesla is doing this

Let us unpack this statement. Are you suggesting that Tesla vehicles (those which has an autopilot) are gathering data about the road conditions and the profile of driver as they drive around, and then they are transmitting those details back home to Tesla the organization? I know they are doing it but not anywhere near the kind of data required to develop a detailed profile.

Oh btw, Tesla barely sells 100K vehicles in a year and last year 17.6 million cars and trucks were sold.

Now, the self-driving car technology is developed by companies like Apple, Google, etc and they are not automobile manufacturers or auto component OEM manufacturers. They are software vendors, and you have component vendors like sensor and radar manufacturers. How these various kinds of vendors are going to come together, and create an eco-system, an eco-system which is ubiquitous and understood by all cars has to be seen. So the technology vendors are different from the automobile manufacturer and who gets to own data is a battle yet to take place. At least Tesla has an edge on this but they are 100K out of 17.6 M market and this doesn’t include other players like cyclists, distracted drivers like myself and drivers who believe others have to adjust to their style of driving like my wife. Now even if Apple and Google form an alliance with auto manufacturers, how do you handle Ubers’s of the world? They can collect data and don’t need GM’s of the world and competes with both GM’s and Google’s simultaneously.

Which technology, and which model (like uber’s of the world, vs zip car, Hertz of the world vs GM’s of the world) will eventually survive needs to be seen and in the mean time how many of these technology vendors will survive and what technology standards will emerge and widely accepted needs to be seen.

Today, and for the past year, several cars are equipped as connected cars.

I am not sure how you define connectedness but these autonomous vehicles are not going to be connected to the mothership, to determine the road conditions, a map of the streets or to make a determination of their location at given point in time, leave alone determine their location viz-a-viz other objects in their vicinity. That determination is going to be made by onboard computer which could get some meta-data from the mothership, but not going to rely on the mothership for determining the next course of action when it is moving at 65 Mph and other objects moving at anywhere between 60 to 80 mph’s and have different ability to communicate back.

they and you will find that Facebook has already gleaned and built a complete profile

Let us assume it is true for a moment, but do you realize the simple fact that it could be either me, my wife or my teenage son could be driving the car, and we could be in different moods and the profile doesn’t have an ability to determine my mental state at a given point in time? The sensor and radar technology will develop, but remember today we are still operating in 2D and humans have 3D skills and so called gut feeling (a machine learning process honed by centuries if not more).

9

I am still driving my 1998 Honda Accord.

excuse me. Is it okay if I worry more about meeting you on the highway head-on in your old Toyota than when I meet a 2021 NVida Robo Ruster XL?

The solution and adaption will come fast once Congress mandates insurance premiums must be raised by 75% for cars controlled by SAOFs (Slow-Adapter Old Farts). It’s win/win/win. Safer for drivers, insurance companies make more money - and we know they aren’t rich and powerful enough now - and CongressCritters get more campaign funding 'cause they need millions to make up for their actual performance. Done deal.

Better sell that old Toyota, and quick.

I’m such a cynic today. < sigh >

"What do you mean today?"

I heard that. Who said that? Gaucho, was that you?

Dan

2 Likes
10

The solution and adaption will come fast once Congress mandates insurance premiums must be raised by 75% for cars controlled by SAOFs

I am sure at some point it will happen. Do you think Congress can drive innovation and technology adoption? Remeber, Congress can create incentives but have not seen they are very successful with stick approach.

11

Do you think Congress can drive innovation and technology adoption?

I doubt it. But governments will be part of the adoption, kind of like all the cities and states that are fighting over high-tech companies’ headquarters. Even my little burg has a booming technology park. Something similar could happen with AI traffic, but I think it would be slower adoption, as most of the advantages in the change will be safety-related rather than economic and in my experience it seems not near as many government entities are willing to pay for safety as those willing to pay for jobs and economic “progress.” (Although, sadly, when was the last time your heard of a tax moratorium that actually paid economic returns to a community?)

Saving injuries and lives has to be worth something. But to whom? It’s not car manufacturers. Locales? Nope, no profit there. Governments, ditto. Us? Sure. But we can’t buy as many pols as corporations, so …

That leaves a few civic-minded cities (Portland, Jackson Hole, Redmond, WA, maybe?) or since they like controversy and are headed for bankruptcy anyway, maybe California. Some influential individuals will have to get involved, people like Elon Musk, Gates, Buffet, etc., or maybe companies like GOOG, AAPL, or FB will step up even more than they already have with non-profit-oriented incentives just because it’s the right thing to do. I really don’t know specific drivers that will be make-it-or-break-it to the trend, but that doesn’t mean they aren’t out there, only that I haven’t looked hard enough. Maybe all of those above and more - that would be my guess but that’s all it would be and worth … not much.

It’s funny. “Everybody” agrees that autonomous vehicles are a done deal and yet “Nobody” has determined who is going to build the infrastructure.

But do I think it will happen? Hey, I may be a cynic, but “everybody” means everybody. Besides, thousands are being killed and maimed each year and that alone makes it the right thing to do.

Besides, autonomous vehicles would end the worry of legal ramifications of driving at 180 mph. Unless AI becomes so humanized that it gets addicted to the need for speed. Then again, the need for speed and AI are kind of locked in an infinite loop, wouldn’t you agree?

Does that mean we can send our speeding tickets to Nvidia? Just wondering.

Dan

12

Do you think Congress can drive innovation and technology adoption?

Greetings, Kingran,

I doubt it. But governments will be some part of the adoption, kind of like all the cities and states that are fighting over winning high-tech companies’ headquarters. Even my little burg has a booming technology park. Something similar could happen with AI traffic, but I think it would be slower adoption if mostly government-dependent, as most of the advantages in the change IMO will be safety-related rather than economic. In my experience it seems not near as many government entities are willing to pay for safety as those who are willing to pay for job increases and economic “progress.” (Although, sadly, when was the last time your heard of a tax moratorium that actually paid economic returns to a community?)

Saving injuries and lives has to be worth something. But to whom? It’s not car manufacturers, unless you think the Teslas and Toyotas will wipe out the Chevies and Fords. Locales? Nope, no profit whatsoever there. Governments, ditto. Public? Citizens? Us? Absolutely. But we can’t buy as many pols as corporations.

That leaves locales, individuals and corporations. A few civic-minded cities (Portland, Jackson Hole, Redmond, WA, maybe?) or since they like controversy and are headed for bankruptcy anyway, maybe California will lead the way. Some influential individuals will have to get involved even more than they have. People like Elon Musk, Gates, maybe Buffet, etc. Or maybe companies like GOOG, AAPL, or FB will have to step up even more than they already have with non-profit-oriented incentives just because it’s the right thing to do.

I really don’t know specific drivers that will be make-it-or-break-it to the trend, but that doesn’t mean they aren’t out there, only that I haven’t looked hard enough. Maybe all of those above and more - that would be my guess but that’s all it would be and worth … not much.

It’s funny. “Everybody” agrees that autonomous vehicles are a done deal and yet “Nobody” has determined who is going to build the infrastructure. But do I think it will happen? Hey, I may be a cynic, but “everybody” means everybody. Besides, high-multiple thousands are being killed and maimed each year and that alone makes it the right thing to do. Besides, maybe autonomous vehicles would end the worry of legal ramifications of driving at 180 mph. Unless AI becomes so humanized that it gets addicted to the need for speed. Then again, the need for speed and AI are kind of locked in an infinite loop, wouldn’t you agree?

Just wondering. Will this all mean we will just send any tickets for driving infractions to Nvidia? “Your Honor, I can’t be responsible for something I can not control.” (Dilemma #317 of 1438)

Dan

13

Tesla is doing this

Let us unpack this statement. Are you suggesting that Tesla vehicles (those which has an autopilot) are gathering data about the road conditions and the profile of driver as they drive around, and then they are transmitting those details back home to Tesla the organization? I know they are doing it but not anywhere near the kind of data required to develop a detailed profile.

Oh btw, Tesla barely sells 100K vehicles in a year and last year 17.6 million cars and trucks were sold.

Now, the self-driving car technology is developed by companies like Apple, Google, etc and they are not automobile manufacturers or auto component OEM manufacturers. They are software vendors, and you have component vendors like sensor and radar manufacturers. How these various kinds of vendors are going to come together, and create an eco-system, an eco-system which is ubiquitous and understood by all cars has to be seen. So the technology vendors are different from the automobile manufacturer and who gets to own data is a battle yet to take place. At least Tesla has an edge on this but they are 100K out of 17.6 M market and this doesn’t include other players like cyclists, distracted drivers like myself and drivers who believe others have to adjust to their style of driving like my wife. Now even if Apple and Google form an alliance with auto manufacturers, how do you handle Ubers’s of the world? They can collect data and don’t need GM’s of the world and competes with both GM’s and Google’s simultaneously.

Which technology, and which model (like uber’s of the world, vs zip car, Hertz of the world vs GM’s of the world) will eventually survive needs to be seen and in the mean time how many of these technology vendors will survive and what technology standards will emerge and widely accepted needs to be seen.

Today, and for the past year, several cars are equipped as connected cars.

I am not sure how you define connectedness but these autonomous vehicles are not going to be connected to the mothership, to determine the road conditions, a map of the streets or to make a determination of their location at given point in time, leave alone determine their location viz-a-viz other objects in their vicinity. That determination is going to be made by onboard computer which could get some meta-data from the mothership, but not going to rely on the mothership for determining the next course of action when it is moving at 65 Mph and other objects moving at anywhere between 60 to 80 mph’s and have different ability to communicate back.

they and you will find that Facebook has already gleaned and built a complete profile

Let us assume it is true for a moment, but do you realize the simple fact that it could be either me, my wife or my teenage son could be driving the car, and we could be in different moods and the profile doesn’t have an ability to determine my mental state at a given point in time? The sensor and radar technology will develop, but remember today we are still operating in 2D and humans have 3D skills and so called gut feeling (a machine learning process honed by centuries if not more).

I find your lack of faith . . .disturbing.

Yes Tesla is sending the data back. This is a known mantra through out the industry. “Data is the new oil.”

None gets discarded. Ever.

Cheers
Qazulight

3 Likes
14

Manual is only one piece of the equation. There are many reasons people hang on to the car they have and, not having a new car, it won’t be autonomous.

15

Yes Tesla is sending the data back. This is a known mantra through out the industry. “Data is the new oil.”

I know Tesla is sending data back on it “auto pilot” models, but that data is hardly significant in terms of determining the behavior or record of other participants to create a profile or even mapping of the streets.

Currently, Tesla is getting meta-data but it is all related to the performance of their vehicles, rather than building profiles of other participants.

None gets discarded. Ever.
I hope you do realize in a 30-mile journey the amount of data the sensors and radar can create, and most of it, if not all are often discarded.

In data science, more data is not always necessarily a panacea. In fact, more data is often leading to noise and not necessarily value.

I find your lack of faith . . .disturbing.

That’s okay. I believe technology will create solutions for many problems in future and may not necessarily at the speed that we think, and often they may end up creating some problems along the way.

For example, the kind of unemployment that autonomous vehicles will create is not a problem that technology will be able to solve. It can only exasperate the problem.

16

Saving injuries and lives has to be worth something. But to whom?

To begin with, the argument in favor of autonomous vehicles is to reduce accidents, that means the manufacturers (that includes technology vendors and the automobile manufacturers) have a great incentive and will sacrifice features in favor of safety. This could potentially create a situation where those who want the experience of “driving” will not be liking boring vehicles, hence will not adopt and the early adopters are all safety drivers (who are more than happy to give up the stress of driving) rather than “aggresive” drivers.

We will go through a period of chaose where some grand standing on the safety will play out. I think safety will be in the middle of the conversation whether for right or wrong reasons.

It’s funny. “Everybody” agrees that autonomous vehicles are a done deal

Agreed. Not just the physical infrastructure, even things like road-rules, insurance rules, various legislations regarding liaility etc need to be hashed out. It may take sometime and there are attempts in the margin are being made now. How soon they will happen needs to be seen.

Just wondering. Will this all mean we will just send any tickets for driving infractions to Nvidia?

A long time, very long time. To begin with these autonomous vehicles will have enough manual control, a new generation of drivers who have never touched wheels has to emerge for the vehicles to be completely self-driving and the manual controls are hidden like “airbag”. Even then I think the liabilities will be borne by the individual owner and may not entirely by the manufacturer.

1 Like
17

AT&T is pulling so much data that it is refining it close to the source before sending it to the data lakes.

My comment about faith was a hint about the dark side of AI.

Cheers
Qazulight

18

AT&T is pulling so much data that

Pulling from where?

19

“Every step you take, every move you make, I’ll be watching you.”

AT&T is required by the 911 laws to know the position of every phone all the time.

AT&T is required by physics to know the signal strength being received by and received from every phone all the time.

In order provide different services, I.E. streaming video, video phone calls, gaming, email, texting, web surfing. AT&T must establish a separate bearer within the cell phones data stream. Each bearer gets its own QOS bit. (Quality of service bit) To do this AT&T must know what type of service is desired so the proper QOS is set for the bearer carrying the service.

AT&T anonymizes this data, refines it, it is too large to transport, and stores it in what is called a data lake.

Cheers
Qazulight

3 Likes
20

AT&T is required…

First of all why you think what ATT is doing is relevant? Is there anyone developing any technology with ATT for AV’s?

AT&T is required by physics to know the signal strength being received by and received from every phone all the time

The cellular network is the definition of highly available but unreliable network. It is available at most places in USA but the calls drop, signals fade all the time. If you use that technology the accident rate is going to be so high, AV’s as a concept will be dead, uber drivers can rest easy because their jobs are safe.

The AV’s are not going to rely on GPS, or cellular network, etc. Rather they are going to rely on sensors, the radar’s installed on the vehicle and use that to determine their position viz-a-viz to other vehicles, pedestrians, objects etc.