Whether you’re right or wrong, this doesn’t affect Nvidia, which is used for levels 2-5
I agree completely Graham.
I think of the AI/Autonomous Driving market a bit more broadly. The article focuses on level 5 which is fully autonomous and it compares that to the internet bubble of the late 1990s and early 2000. Specifically, it calls out NVDA when making the comparisons. I agree that autonomous driving will happen in stages, but I disagree that level 5 is required for anyone to make money. I believe that level 2 “autonomous” can come rather quickly, perhaps becoming a standard in most cars within 5 years. That opportunity alone is enormous. In 2017 there were more than 17 million cars sold in the US (more than 75 million globally). If a level 2 AI chip at a price of $100 becomes the standard in cars that’s a $1.7B market in the US and a $7.5B market globally assuming that car sales don’t grow at all. Car sales are still growing and will likely be over 80 million in five years. Just yesterday I learned that Toyota (the largest car maker in the world until recently when VW over took them) will make level 2 autonomous driving standard in all their cars sold in the US and Japan by 2020. That’s more than 8.5M cars which doesn’t seem like a small number to me.
Let’s think about what a car really is. It is a box that carries people around from point A to point B. When I say that I think about AI/Autonomous Driving more broadly, I mean that there are A LOT of other uses for boxes that move around transporting things from point A to point B. Think of a warehouse, a shipping dock/port, a farm, etc. There are millions of items that need to be moved around. Automating the movement of crop harvesting, cargo sorting at ports and in warehouses can lead to massive cost savings, quality improvements, safety advantages, and automatic tracking, quality control, and inventory management benefits. These are a few examples of AI/autonomous movements that can be developed using essentially the same technology that would be used for autonomous driving. Work on these uses cases is already underway. At the GPU Technology Conference in May, there were 245 “automotive” companies(of which only 20 were car makers and 35 were vehicle makers) looking at ways NVDA’s technology can be used to automate motion and the movement of objects from point A to point B. John Deere was there to explore automation is farming. 244 other companies were there to learn how computing and machine learning can be used to make their use case a reality.
So when that author of the “great article” on autonomous driving says that autonomous driving contains a lot of hype, he’s missing the point. This technology revolution is not just about moving a person from point A to point B with zero human intervention. He makes it seem like that’s what’s required for NVDA to make any money. Well, NVDA already made $140M in revenue last quarter. The number of partners working with them in the automotive sector almost tripled from Q4 2017 to Q1 2018.
28 Likes
So when that author of the “great article” on autonomous driving says that autonomous driving contains a lot of hype, he’s missing the point. This technology revolution is not just about moving a person from point A to point B with zero human intervention. He makes it seem like that’s what’s required for NVDA to make any money.
I didn’t take it like that Chris. In fact I’ve been adding a little to my NVDA position almost every day, including today. I just agreed with him that there are many, many individual people who won’t accept Level 5 autonomous cars. A few weeks ago someone posted a study here saying that by 2025 all cars would be Level 5. No way.
I did like his quote “Technology isn’t adopted when it’s ready. People change slower than technology does - and the faster technology changes, the greater the gap becomes.” I thought there was a lot of wisdom in it.
Saul
6 Likes
I did like his quote “Technology isn’t adopted when it’s ready. People change slower than technology does - and the faster technology changes, the greater the gap becomes.” I thought there was a lot of wisdom in it.
I think the above quote is only partially true. The rate of technology change is accelerating. I think the gap is shrinking. Just look backward in time. People are changing how they do things at an accelerating rate.
What’s most relevant to this board is the rate of adoption of a new technology. The barriers to adopting new technology is dropping. Let’s take communication as an example.
You want to communicate to another person.
In the distant past, you would need to speak with them directly.
Then someone invented paper and writing instruments. There were a lot of barriers to communicating via paper and writing instruments. You needed to acquire these instruments, you needed to be literate, you needed to know where the other person would be and when, you needed a means of delivery. Lots of issues here so adoption was very, very slow. It took centuries to get mass adoption.
The telegram removed some of these barriers but many remained.
The telephone was another innovation but this one took decades to get mass adoption. Decades is a lot less than centuries though.
The mobile phone though took a lot less time for mass adoption than the land line.
Email adoption was a lot faster than snail mail.
Facebook was even faster then the above.
The thing about technology is that innovation builds on all past innovation. You also have a lot of advancements in infrastructure. Email would not be possible if computers were not widely accessible.
Unless you look back in time, it’s not so evident that change is rapidly accelerating. People change quickly when it makes life easier or more convenient. The gap between people’s willingness to change is shrinking not because they are less willing to change than before but because the friction associated with change is decreasing.
14 Likes
My take is that Level 2, 3, and 4 will be accepted (as long as they are analogous to a gradually improved “cruise control” that you can escape from by putting your foot on the brake), but that people will be very, VERY, slow to adopt Level 5 (with no no human controls or steering wheel, and the car completely in control of the trip). Would YOU accept to buy a car like that?
I would buy one. I would like it to have a steering wheel so I could drive it when I want to drive it. I enjoy driving. I used to deliver pizzas when I was in college, and I couldn’t believe that I actually got paid to drive around. But there are times when I really don’t like to drive. I hate drive in traffic, for instance. If I could flip a switch and let the car drive in bumper to bumper traffic then I would be more than happy to let it take over. Or if I was in a city with difficult to find parking I would let the car drop me off and then go find parking or go earn me some money by driving paying customers around until I needed the car again. Driverless capability will only be allowed if it’s proven to be safe. It should be safer than human drivers and it will be eventually. Insurance companies will know this. If the average person drives 12,000 miles per year and pays $1200 per year for insurance the insurance rate is $0.05 per mile. Now, an insurance company would charge less if the car is safer. I could see insurance companies charing $0.01 per mile when autonomous driving is enabled because the chance of an at fault accident would dramatically decrease with autonomous enabled.
Cruise control has been available in cars for at least 20 years. I think many/most cars have this feature. I never use it. I don’t like it because I feel it is unsafe and because you still need to pay full attention. I pay less attention when the cruise control is on. I also tend to get more sleepy when it’s on. And I get bored. Now, if the car could drive and I could read or do something else productive while I’m commuting somewhere that would be a benefit.
When we look at adoption of a new technology we need to look at the benefits and where people would want those benefits. As I’ve pointed out above, there are many benefits. The objections that Saul raised (full AI with no human control possible) can easily be addressed by making the car fully autonomous by flipping a switch. I don’t see why that wouldn’t be done and why that would be incredibly easy to implement.
3 Likes
While I do agree that full autonomy will take a while to become mainstream, however who would not want the massively increased safety that autonomous systems provide?
Look at these situations where Tesla Autopilot gives you advanced warning of impending crashes https://youtu.be/DuIrjRAzNPQ
This is huge! With nearly 40 000 deaths on US roads in 2016, even without level 5 this technology is game changing.
Justin
Full benefit of “no steering wheel needed” will not come until nearly all cars are so equipped and can communicate with each other real time …“There is a bad pothole 0.82 miles ahead” So maybe 15 or 20 years away . That’s beyond my investment horizon.
We don’t need L5 autopilot to make money in stocks, L3 or L4 should be enough So that is what I am interested in now.
There was no huge demand for ABS or cruise control when first introduced but now everybody wants them These are a form of autonomous driving. Polls are worthless when it comes to technology benefits people have no experience with. Before iPhone were people demanding apps?
The projections of deployment for AI and autopilot are based on more definite evidence than was the spread of PC in the early 1980s. Or the spread of Amazon beyond books (unless you knew the genius of Bezos)
I have listened to scenarios difficult to handle by machines and realize that some of them will indeed be tough for AI. Today or over the next 3 or 4 years, Example,it happened to me, stuck on X way due to ice, saw another car cross median successfully, followed it. Ones who did not cross were stuck for over 8 hours. AI probably could not figure this out , it was driving on a forbidden area.
On a recent trip Google maps directed me incorrectly several times ,dangerously so a couple of times.
6 Likes
The objections that Saul raised (full AI with no human control possible) can easily be addressed by making the car fully autonomous by flipping a switch. I don’t see why that wouldn’t be done and why that would be incredibly easy to implement.
I agree that that would have to be the solution (or people would never accept it). It would probably add some extra cost as you’d have to have two redundant sets of controls, but that’s probably the way it’s going.
Before iPhone were people demanding apps?
There may be a big difference in how people see it. No one felt that having an iPhone, or any of the apps on it, was taking something away from them.
A lot of people would feel that no longer ever having the option to drive their own car was taking something away from them! Even if sometimes they’d be happy to have the car drive so they could do something else. As long as you have a switch you can flip so you can drive yourself, it will be okay. It’s not the car’s autonomy people will worry about, it’s their own autonomy.
Delivery vans in an industrial park will do fine without a steering wheel and human controls, but I don’t believe the majority of humans will accept personal automobiles without a steering wheel and human controls in my lifetime.
Saul
2 Likes
I think there is an age/generational factor in play here as well. I don’t think Millennials and the wave after will have any qualms whatsoever about driverless, and urban dwelling older people won’t have many either, I don’t think.
2 Likes
I did like his quote “Technology isn’t adopted when it’s ready. People change slower than technology does - and the faster technology changes, the greater the gap becomes.” I thought there was a lot of wisdom in it.
Saul:
I assume you are very familiar with the TALC:
https://en.m.wikipedia.org/wiki/Technology_adoption_life_cyc…
It has been a generally accepted process for adoption of new technology…I rather doubt that we will see any difference with autonomous driving…realizing there are MANY issues that need to be worked out (viruses, what makes decisions about life/death if one or another are inevitable, big brother tracking your every whereabouts, error rates, etc.).
But keep in mind that there are 77 MILLION new cars sold worldwide EACH year. That is a whole lot of room to sell autonomus technology even if the “late majority and laggards” delay a bit longer…say now nearly 40 million annually adopting pretty quickly.
What may be more difficult to calculate from an estimation perspective is what impact this will have on need for car ownership to begin with. Some are predicting by 2030, none of us will need to own a car…they will be just circulating around picking and dropping off passengers…the cost being less than ownership. That could certainly reduce the above revenue generation expectations for NVDA, etc.
But in the near term (10 years) the sales expectations for autonomous are likely to be pretty robust.
7 Likes
SaulR80683: Delivery vans in an industrial park will do fine without a steering wheel and human controls, but I don’t believe the majority of humans will accept personal automobiles without a steering wheel and human controls in my lifetime.
I tend to think the same, especially as I am someone who places an extreme value on my own self reliance and personal freedom. Except lately I wonder … is this just fear of the unknown? Fear of a radically different way of life I filled with benefits and drawbacks which will be far different from the fantasy floating around my head? Once I experience autonomous driving will find a different perspective?
1 Like
Think of this use for an autonomous car, for those who can work anywhere as long as you get a decent wifi or LTE signal.
I could get a beach home 8 or 9 hours a way, and spend every weekend there at the cost of gas and wear and tear. Which on the right car is not that much.
Why I could do this is if the car was driving itself I could easily have a work station in the car and get as much work done remotely, while the car drove, as I could in the office. So it would be like I got in the office, hit the go button. And then I step out of the office 8 or 9 hours later (of course with a few breaks en route) and got work done I would have to get done anyways, and I’d be at the beach home.
I’d have no wear (as I would have had to work anyways), no time wasted (as I’d have to have worked anyways), and I get to enjoy my time at the beach home without the burden of travel that is a major bummer for having vacation homes or just taking vacations.
Even if I took an airline (e.g. Atlanta to Minneapolis is one about a 2 hour flight, but it is a 1 hour drive, 1.5 hour in the airport and waiting to board and waiting for the plane to taxi and take off, taxi at the destination airport, get bags and rental car and drive to location (heck that is another 1.5 to 2 hours at least, and by this time the 2 hour flight is a 7 or 8 hour affair itself, and your exhausted).
Yeah, an autonomous car opens up real possibilities. I could equal these possibilities by either having someone else drive me (and then they are wasting their time), or to pay a chauffeur to do it and then we are talking real big money.
So yes, autonomous driving will mean a lot of things to a lot of people. Things we may not even have thought about.
Tinker
4 Likes
I wonder if the author has taken a plane trip lately. People get on a plane with no reservation that the majority of time it is in autopilot or driving itself. Now the concept of open air space versus being on the road, weather, remote locations, security, etc. are all still issues.
One of my son’s just turned 16 and is taking driving classes and one of the first things he said was; if we had self-driving cars I wouldn’t have to be dealing with this. I think that is applicable to today’s youth - they would embrace it.
On people accepting technology slower - completely agree. I’ve been in Enterprise high tech sales for many, many years. One of my first “sales” jobs was selling something called Bank at Work in the late 80’s.
It was one of the hardest sells I have ever dealt with in my career. I was trying to “sell” them the concept of having their hard cold pay check directly deposited into a bank account. They wouldn’t have to wait in the teller line on Friday afternoon, mgmt. wouldn’t have to give extended lunches, you would get this ATM card and it would allow you to get money out any time you wanted it, etc.
No way, I want to see that cash in my pocket and I like saying hello to everyone at the bank.
Sox Nation
6 Likes
So yes, autonomous driving will mean a lot of things to a lot of people. Things we may not even have thought about.
Tinker,
How about a micro distillery and bud bar. You drink and toke here, we auto drive you home.
Sweet, but the potato chips are 5 dollars a bag in the car.
Cheers
Qazulight
1 Like
I have several issues with the original article
- The ‘cost’ of owning a car - any type - is a factor, but until there is a fairly large difference, it won’t shift most of the population. I could live in a smaller house and save money. Do I? I could not have a swimming pool but instead use the ‘community pool’. both are wet. Do I? I cook many meals at home. Many urban millennials eat out most of the time for the convenience of not having to make meals - or get ‘take out’ - or ‘delivery food’. Likely doesn’t save them money.
Just about everyone lives in a recently built dwelling with individual per unit climate control systems. Why? Heck, the buildings could have ‘central systems’ but now most don’t. You’ve got your own a/c.
the author also talked about 10:1 differences in cost. I really doubt it
- the cost of maintenance will not be that much cheaper. Oh, yeah, once a year $39 oil changes - but you’ll have once a year ‘car checkups’ anyway to check the electronic systems, check the steering components, the a/c, any air filters for the cabin, lube the door locks, etc.
Also…any EV or AEV is still loaded with the same dozen systems - steering and wheels and suspension to start…then a/c and compressors and heating systems. Then audio enterntainment (which is one of the biggest repair items on cars)… and seats and power windows and defrosters and windshield wipers… all of which are going to crap out long before anyone gets to 500,000 miles
You ‘rent out’ your AEV or EV, thinking magically it good for 500,000 miles…and you’ll be seriously disappointed.
My 1990 Honda went 17 years and 170,000 miles. typical 10,000 miles per year for most drivers. Did it crap out? No…but the seats were shot from lots of butt time. The carpets were a disaster area after 17 years of daily use. everything else worked - original a/c, suspension - although it needed new rubber bushing parts - it wasn’t as good as ‘new’ - power windows - all good.
But you still , even on a EV, have to replace tires and windshield wipers and likely fix power windows and audio systems and a/c and a whole bunch of other stuff.
Then add it complicated and high power electronics - battery cooling systems - high power charging systems.
Tesla just announced plans to open 300 MORE service centers and hundreds of mobile service centers to fix…EVs…
- Just because a technology is ‘there’ doesn’t mean folks will rush out and ‘buy it’. I’ve seen ten years of hype about ‘virtual reality’. How many of you have ‘virtual reality’ in your life? maybe your kids have a headset and play games with it. And the military uses it.
I’ve seen 30 years of hype about the ‘smart home’. Oh, call home and turn on the living room lights, adjust the blinds, turn on the oven for dinner, let the dog out … How many people here have a ‘smart home’ with blinds they can adjust from anywhere? or even turn and off a light in the house? Open doors remotely to let in the Maytag repairman service guy?
After 10 years of hype, about the only thing folks have is an electronic thermostat…and most of them not ‘net controlled’.
- Yes, I’m sure it is coming - but I’m not sure how fast will rush out to buy fully A/V vehicles. Long before, I’m sure folks will get ‘autopilot’ type ones…but it will be interesting to see how they do in ice and snow…and through puddles. And potholes. And construction zones.
Wow…7 seconds of warning…duh! You’re going 65 mph…and suddenly the car is in a construction zone at 55, you haven’t been paying attention, you’re reading or watching a video, and suddenly your Tesla gives an alarm - it’s confused - it takes you a few seconds to look out the window - it’s raining - it’s night - and you’re confused - suddenly the lane ends, you go off the concrete on the bridge 150 feet up and your Tesla suddenly is airborne… oops, it missed the ‘lane ends’ in 500 feet sign - merge left… try it some day as a passenger at night - on a rainy night - and tell me how many seconds it takes to take in where you are at, what the situation is, and how to respond…especially if there is an 18 wheeler just to your left preventing emergency lane change - for which you aren’t prepared for either…
–
As to sitting in your car going to the beach home 8 hours away…it’s still sitting in the car for 8 hours. I can’t go 8 hours without a pee break. Neither can your EV…it’s got 200 mile range, so you’d likely have to stop at least 3 times…for 15 minutes a time or more…and maybe you’re 3rd in line at the ‘pump’…45 minute wait - times 3…and of course, you’re car is smart enough to know that, but you still sit in the same seat in the same car…because - it might get confused… and then you have the return trip at the end of the weekend. Dream on. But let us assume that’s the norm…so everyone is letting their EV take them to the beach on weekends…100,000 of them…hmmm…traffic jams? maybe 20 cars waiting at the ‘pump’ in front of you? 15-20 minute recharge time?
Volvo made a big deal about EVs…I don’t think it’s delivered one of them yet, and if you read the ‘footnote’ that EV deal includes hybrids.
I suspect a lot of hybrids will get made first. Millions of them. Just so folks can go to the beach…or to MT for vacation…in the boondocks where there are no EV charge points…and they’ll have ‘smart systems’ like collision avoidance and better ‘cruise control’ and forward collision radars/lidars…
Right now, people can ‘save’ bucks on gasoline by buying hybrids. Maybe 1000 a year. How are hybrid sales? 30% of the market? Gimme a break…
- Another factor is that half the people in the country have no way to charge an EV. Live in NYC or urban area and park on the street? out of luck. Live in apartment building with 300 car lot in front? out of luck. Live in an old home with 40 amp entrance service? out of luck… live in a rental unit? Likely out of luck…Live in a mobile home? Maybe out of luck. That’s about half the people in the country.
Others with 3 or 4 cars and 2 car garages - houses with teens…same problem…park on street…all over suburbia…
Until the 15 minute charge time arrives and you can get a charge at the corner ‘gas station’ …half your market is gone. Same for anyone wanting to travel.
No, I don’t think it will be ‘exponential’ as in doubling and doubling and doubling from 5 to 10 to 20 to 40 to 80% in 8 years…
but I could be wrong…but I’m still watching a 5 year old TV set…living in a 27 year old house…with original carpets and paint and wallpaper…(new furnace and a/c though)…27 year old refrigerator. 10 year old car with 50K miles - hybrid. 43 mpg around town.
I do have a cell phone and smart phone and remember when they cost $3000 for a bag phone and $200 a month - heck, I designed systems…but now you can get service for $10/month - or $25 a month for more than most people talk…
when EVs cost $2500 each - let me know…
and of course, for business and families - cellphones save backtracking, add $$ to the bottom line real quick.l…and convenience.
for most people, the main factor in buying a new car is the initial price. Right now gas operating costs is not that big a deal. Heck, 200 miles a week - 30-40 mpg - at 2 bucks a gallon…not that big a deal. Maybe only 100 miles a week. five bucks a week. Once a year oil changes - big deal.
Now you buy an EV…but it doesn’t cost another $2000 for a charger for a hybrid or gas engine car…right? and that charger will likely be obsolete in 10 years. If it lasts that long in the 117 degrees heat during the Phoenix summers for 10 years in your garage… You still use the charger for your cellphone from 10 years ago? or from your PC from 12 years ago?
t.
10 Likes
Thanks telegraph, you put a lot of my concerns and skepticism very elegantly.
Saul
Telegraph,
Some good points but a few things you overlooked.
One biggie is how an electric car enables more space in the interior of the car. The engine is tiny, about where the transmission box is. There is nothing in the hood nor the trunk and this allows for much more room in the interior. The Tesla S for example has the same cargo space and interior room as an SUV.
Second, autonomous driving alllows for thisblarger interior space to be designed for comfort or utility and not as much for driving. Enabling one to have a comfortable lounge space. Heck, I do most of my work at my at home bar.
Third, YES there will be stops. Do you think an ADD quack like me w a bladder the size of a thimble is going to work 8 hours straight? Heck no, never.
Fourth, not clear on the traffic problem but not a big d al as we have traffic now that the autonomous vehicle can now deal with.
But long a short Level 5 is still some time away, but if it were here today the long weekend vacations is an enticing reality. Yes I’d work there and back while my vehicle did the work.
Tinker
Yes, I’m sure it is coming - but I’m not sure how fast will rush out to buy fully A/V vehicles…
The link below is to an interesting On Point podcast that discussed the adoption issue. There was a thread that centered on when autonomous vehicles will have to be adopted due to the significant cost benefits that will be provided by doing so, putting those that do not adopt at such a disadvantage that they will be economically forced to adopt faster.
http://www.wbur.org/onpoint/2017/06/23/pulling-out-of-car-ow…
This is probably a stronger influence on adoption than choice.
DJ
the cost of maintenance will not be that much cheaper.
I’m only one person, but that’s not my experience. I have a 2008 Honda Odyssey. It’s great utilitarian car that is very reliable – we love it – but it does need regular maintenance. It just turned over 100K so I had to have:
$135 Power Steering Flush
$ 63 Brake Fluid Flush
$107 Transmission Flush
$1134 Water Pump & Timing Belt replacement
+tax = 1582.
This is on the high side, but there is something to be done most years.
Conversely, I also have a 2013 Nissan Leaf SL. I bought it used 1.5 years ago. It’s amazing in every way except it can only go 85 miles on a charge. Which works great for my 20 mile each way commute, and errands. I charge it at night from a regular outlet with the trickle charger that was included with the car. If I need to go farther I take the Odyssey.
I just went to the Nissan Dealer for yearly maintenance on the Leaf. They performed a bunch of inspections, and charged me $20 to rotate the tires. That’s it. To tie it back to stocks, I think dealers are going to be in bad shape in a few years, like the Maytag repair man.
Now the Odyssey is 5 years older so it’s not a perfect comparison, I get that, and some maintenance doesn’t go away just because it’s an EV, I get that, and there is also risk that I’ll have to replace the battery at some point (though I’m still at 12/12 bars so it’s likely a long ways off), but for the moment the difference between the two is quite large, setting aside the savings on gas which are substantial compared to electric where I live.
Not having cylinders/fuel injectors/exhaust/timing belts etc seems to be a lot simpler, mechanically. Brake discs/pads get used a lot less due to regenerative braking (my 2004 Prius had stock brake pads at 100K when I sold it).
Overall, time will tell, but my money is that EVs cost a lot less and will prove more reliable/less expensive due to fewer moving parts/fluids.
4 Likes