Uber is Toast

Do you know what's going to be one of the biggest disruptors here very shortly is the fact that the number one occupation in the United States is "driver", and because we're going to have automated driving, trucks, taxis, I will tell you it's going to be a gigantic disruptor moving forward.
-- Gary Johnson, 2016 Third Party Presidential Debate, Part 1

The other day, my chip nerd kid sent me the video below about Tesla's AutoPilot. It's a four hour mix of marketing propaganda and dense technical information. For those of you who are technically inclined, it is worth watching, but in the name of all that is holy, please jump to 1:09:12 and skip the filler they put in because Elon was an hour late...

For the less technically inclined, I will describe in a bit what they are doing in layman's terms, but here is the short version:

TL;DR Tesla has figured out how to use the fleet of cars belonging to their customers to train their self-driving AI on real-world data. By comparison, their competitors like Uber are stuck using artificial simulations and test drivers. Tesla's system is already twice as good as real drivers and they intend to set up an automated "RoboTaxi" service next year where owners can earn money providing driverless ride-sharing services when they are not using their car. The payback is 5x the purchase price for the owners and Tesla's cut is 35%. Ride-share gig drivers are, of course screwed, just not by the people they expected to do the screwing.

I'm not trying to be an Elon fanboi here, but I am familiar with many of the technologies being described and this is real. Musk has a tendency to exaggerate, but I think this is going to happen a lot sooner than I had expected. AlphaGo caught me by surprise, and I think this is in the same vein. The discussion below is intended to support this contention.

The presentation has four sections: chips, machine learning, infrastructure scaling and RoboTaxis. The infrastructure part was not very interesting, but I will try to summarise the other sections.

[video:https://www.youtube.com/watch?v=Ucp0TTmvqOE&t=1h9m12s]

Chip Design

Musk hired an entire chip design team to build the AutoPilot processor. I'm not a chip designer and I don't know the guy, but but I do work "close to the metal", and everything he said made sense to me. They had one customer with one task so they could design a custom chip that used as little power as possible (50W) to run the neural net and everything else. It looks like they also licensed the ARM design (used in iPhones) to have general computing capabilities and there are a few other specialised bits of silicon for doing image processing and such. The result was something that can process about 2100 frames per second. By comparison, a repurposed nVidia video card's hardware can do about 300fps. (For reference, celluloid movies are about 30fps.)

The power consumption is so low that the 50W budget actually powers two duplicate chips. The chips compare their answers and indicate a failure of they disagree. I'd be happier with three chips that can vote (like the space shuttle systems), but that may be in the next revision which they hinted at and uses smaller transistor sizes (10nm vs 13nm). The packaging was designed for easy replacement, so no one will be stuck with the old version.

Machine Learning

The chip stuff was interesting, but fairly straightforward in some ways: I don't doubt that another experienced team could do something similar. What they did for the training part is both powerful - and probably impossible to replicate.

There are two kinds of machine learning: Supervised and Unsupervised. Unsupervised learning means that the system can figure things out on its own. This is great (not to mention cheap...) but it is only suitable for highly restricted problems like playing a game. The way Google trained AlphaGo was unsupervised: They just had to give it the rules and let it play itself.

The rules for driving are a lot more complex than Go, so you really need to use supervised learning. The way this typically works is you give the system some data along with the right answer for each piece of data (like: given these camera feeds, change lanes to the left). You split the data into "training" and "test" data, train the network with the training data and then check the results with the test data (if you ever read about something where they used the training data for testing, ignore it - the results are worthless.) The tricky (and expensive) part is attaching the correct answer to each piece of data.

All of the other people working on this problem are using simulators to generate the test data. The trouble with this is that you are limited by the imagination of the simulation authors. Think about all the problems that have come up with poorly trained algorithms (like the African woman researcher at MIT who found the standard face recognition package didn't even recognise her as human...)

What Tesla did was start with some basic driving situations and had humans label the correct responses. They then sent the model out to all the Tesla's on the road had them run the model (without actually controlling the car) and compare the output with reality. If the model was wrong, it sent all the data back to Tesla to add to the training set. The key here is that this data was already labeled. The car should have done this but it did that. We expected the other driver to do this but they did that.

This means that they can refine their supervised learning system pretty much continuously. They get new data, add it to the training set and then the new model out for testing all in a matter of days if they want. Add in the half a million Teslas on the road, and the amount of data they can generate, use and test is overpowering. Uber, Waymo and others in this space are stuck with a few prototype cars on a test track - or in an unsuspecting Arizona town. The scales and systems are just not comparable.

RoboTaxi

Once you have a self-driving car, you can make it into a self-driving taxi. Most cars sit idle most of the time, which is a terrible waste of resources, both in terms of manufacturing costs and limited public services (like parking and road surfaces). Tesla intends to set up a ride hailing service like Uber but with no drivers. Because there are no drivers, they can undercut ride sharing pricing quite significantly, and Tesla would take a 35% cut (Musk compared this to the Apple Store.) The owner's share would be enough to pay back the cost of the car five times over.

The other piece here is that because the cars are electric and require little maintenance, they can build them with longer lifespans. The current production run has an estimated lifespan of a million miles - which sounds like a lot if you think about your own car, but apparently that is the design lifespan of semis, so it is at least plausible.

I have some reservations about this system, but they are details. I'm not sure that some people would like to share their car with strangers - many of us think about our car interiors as our "personal space". I'm also not convinced that the 50x figure covers peak times, but there was a graph that looked like it was at least in the ball park.

The Good

That was the technical summary, but what are the implications, especially the political implications?

First, the good things:

  • Reducing the number of cars in the world by a factor of 50 would reduce the manufacturing environmental impact of cars
  • Switching more people over to electric vehicles by sharing would accelerate electrification of transportation
  • Tesla's system appears to already be safer than the average (Tesla) driver

One point about the perceived safety of self-driving cars: My other kid pointed out that most people compare automated driving to perfect driving when they should really compare it to average human driving. Add in the fact that 80% of drivers think they are above average, and there is a big disconnect. There is a point to be made about algorithmic transparency here, but it is also true that we don't really understand how humans drive, and at some point we have to fall back on empirical data, and it sounds like this system is already safer than humans. Nevertheless, we need to check that data.

The other Good Thing is the title of the essay: There are gradations of evil and Uber is way up there. If nothing else, this system will kill them and redistribute their enablers' cash.

The Bad

This is a solution to a social problem based on private ownership. Only those with capital reserves can buy these cars and profit from being rentiers. And even if Tesla provided financing based on the expected return from rentals, there are a limited number cars that would actually be needed (and reducing the number of cars is an important ecological good) so only those who lucked out and got into the game early would benefit. The rest of society would end up paying through the nose because the rental cost is higher than the cost of owning your own vehicle (otherwise there would be no point in renting your car out if the damage was greater than the fare.)

It's not clear to me that Musk would actually care about Tesla being nationalised. He has stated in the past that the point of Tesla was to jump start the conversion to EVs, so I suspect he could be convinced. The investors, however, expect their pound of flesh and would have to be dealt with so we can share the benefits of a system like this more equitably.

The Ugly

If we don't deal with this problem equitably, then Gary Johnson is right: We are going to see massive disruption caused by the disappearance of the number one job category in the country. MLK observed, "a riot is the language of the unheard" and some here might welcome the disruption as a prelude to revolution.

I always hope that we can solve our problems without violence, which is why I am writing this essay. If we understand the nature of impending change, maybe we can harness it to be less destructive.

Share
up
0 users have voted.

Comments

mhagle's picture

Out here on gravel roads, etc.

Going electric faster is good. Fewer cars on the road is good. But I will never get in one!

up
0 users have voted.

Marilyn

"Make dirt, not war." eyo

Hawkfish's picture

@mhagle

There are people making (and hacking) all kinds of vehicles, including pickups. So yes, the rental model doesn’t work everywhere. But I bet you don’t have Uber drivers out there either!

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

Pricknick's picture

@mhagle
Don't buy one.
Everything you do is recorded and added to a database.
So far, the database works as well as a 737 max.
I'm still amazed that humans die putting their faith in a two and a half ton battery.
I'm all for getting off the grid as long as the grid don't bite back.

up
0 users have voted.

Regardless of the path in life I chose, I realize it's always forward, never straight.

Hawkfish's picture

I’ll try to check in regularly this evening but we have record heat today in Seattle and I’m going to be in the basement watching Netflix, drinking lager and trying not to move...

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

Pricknick's picture

That makes tesla carbon.
Please encourage Ellon to take the first trip to Mars.
Now the silly asshat is into mining.
https://www.telegraph.co.uk/technology/2019/06/11/tesla-take-new-mining-...
Rare earth metals now outweigh sending a few hundred hamsters to a distant planet where they'll die.
The man is insanely backed by our dollars.
Take a trip. Drop acid.

up
0 users have voted.

Regardless of the path in life I chose, I realize it's always forward, never straight.

Hawkfish's picture

@Pricknick

I was looking at a chart the other day showing a serious growth in demand over the last 2-3 years. So Musk is just reflecting that reality.

The real problem is that all of these mines are quite environmentally destructive. I gather that a lot of our lithium supply comes from South American mines that are pretty grim. Then there is all the cobalt that is the "ion" in these batteries, which mostly comes from the Congo and the conditions there are often horrific for the miners.

But if we are going to switch to renewables for energy, we are going to need utility-scale batteries, and that means a lot more of both elements. The supply is there (lithium is one of the most plentiful elements, both in the universe and on earth) but the process is pretty toxic. There are no easy - or even good - answers, but we have to solve these problems somehow or a lot of people are going to live in misery and then die in horrific ways.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

@Hawkfish @Hawkfish
that India is working on. India has a lot of Thorium and no Uranium deposits. According to Scientific American, thorium is less poisonous, less radioactive and much less likely to explode. They say impossible to explode, but I never discount human stupidity. Chernobyl was the results of the operators doing unauthorized experiments and Three Mile Island occurred because the operators kept overriding the safety controls that were trying to shut down the plant. If they had all gone to the parking lot and left the plant, there would have been no meltdown, no radioactive release, just a blackout.

EDIT:
Fukishima was the result of criminally negligent design, namely storing radioactive waste on top a a reactor. Indeed, criminal negligence in building a reactor in an earthquake zone.

up
0 users have voted.

I've seen lots of changes. What doesn't change is people. Same old hairless apes.

Hawkfish's picture

@The Voice In the Wilderness

But I do expect the math to work out. Right now modular Gen-IV reactors like the MSRs seem like they with give the most power with the lowest environmental footprint.

It's all very well to point out how cheap renewals are today when you don't have to account for production externalities (mining, destruction of habitat for solar and hydro) but when we start to run out of strategic minerals (cobalt, iridium) and suitable deployment space (offshore wind locations) the idea that money makes everything infinitely fungible will break down.

The other thing almost nobody is talking about is resilience of these systems in the face of changing weather patterns. Tornadoes are increasing in frequency and severity, which does not bode well for midwestern wind farms. Rainfall changes may make hydro less reliable. Etc.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

Elon Musk and the handful of tech billionaires like him who want to profit from the mass destruction of jobs this will cause. Their destruction as a controlling class and the economic cannabilization of their parasitic monopolies is the only ultimate good that can come of robotization of transport and other non-tech occupations.

up
0 users have voted.
snoopydawg's picture

The retail stores have all been gutted by hedge funds and people fired with very little compensation. Banks are also laying people off as are the car companies. And unknown until recently tens of thousands have lost their jobs at Facebook, Google and Apple. Young people are graduating college with huge debts and can't find jobs because the older people are either still working or have gone back to work. Guess some people can join the military for a job and training...

So yeah this is a huge problem that needs to be addressed. Anyone think that congress is even thinking about this issue?

Interesting essay and topic. It's too bad that most people will just move into the new society instead of telling them to go F'ck themselves.

up
0 users have voted.

Which AIPAC/MIC/pharma/bank bought politician are you going to vote for? Don’t be surprised when nothing changes.

A. Surprising
B. Non-obvious

"Tesla has figured out"? How could it be anything other than obvious that the way to train autopilots was to gather the limitless possible data from cars that already have the necessary sensing equipment? If I'd wanted to develop driverless cars, I would have started by paying people to equip their vehicles with the sensing gear. And in case you're wondering, this isn't one of those cases of, "yeah, sure, it seems obvious now that you've heard about it". I explained something like this to somebody a few years ago.

I'm still not sure I will ever feel comfortable on the road with such machines, given that their experience and knowledge is limited to stuff that has happened, somewhere, on a road over which a Tesla has driven. Human experience and knowledge extends well beyond that space -- which means, a human may be much better equipped to deal with a situation where some other space has improbably intersected with the "car driving on a road" space.

up
0 users have voted.

The earth is a multibillion-year-old sphere.
The Nazis killed millions of Jews.
On 9/11/01 a Boeing 757 (AA77) flew into the Pentagon.
AGCC is happening.
If you cannot accept these facts, I cannot fake an interest in any of your opinions.

Lily O Lady's picture

@UntimelyRippd

when to brake or when I’m veering off the road. However, my car sometimes tells me to brake when there is no vehicle anywhere in front of me. It thinks I’m veering off the road because of sealant lines on the road. Considering my experience with this technology, I have no intention to trust my safety to a self-driving vehicle.

up
0 users have voted.

"The object of persecution is persecution. The object of torture is torture. The object of power is power. Now do you begin to understand me?" ~Orwell, "1984"

Hawkfish's picture

@Lily O Lady

The systems in most cars are pretty stupid because they don't have a good training system. Tesla (whatever else you may think of them) is doing this right - and I bet UR agrees with that because he came up with the same idea.

The real problem is that everyone else is trying to "compete" when they aren't solving the problem the right way. That doesn't create better systems, it just creates dead bodies.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

Hawkfish's picture

@UntimelyRippd

My point in the essay was twofold: Help people understand the various issues and explain that some of them weren't obvious to the other players in this space. (It also wasn't obvious to me, but I didn't mean to imply that only the genius of Musk could have come up with it.)

It does raise the question of why only certain people get to benefit from such ideas.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

Pricknick's picture

@Hawkfish
It raises the question of why only certain people are daft enough to put faith in such ideas.
We're all techno in some way. It's when we start to believe that techno is the solution that we lose.
I prefer a human copilot. Jesus is not included.

up
0 users have voted.

Regardless of the path in life I chose, I realize it's always forward, never straight.

Hawkfish's picture

@Pricknick

I am a bike commuter and even on my short ride I see cars doing dangerous stuff all the time. As I said to a friend a few years back "Monkeys make terrible control systems for cars."

But I hear you on techno-worship. I prefer the Amish approach: They just ask whether this will make our lives better. Which is why they use horse-drawn buggies. Unfortunately that doesn't scale to the number of people we have right now, so we have to make some trade-offs.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

@Hawkfish
that it should be "obvious to anybody," but only that it should be obvious to anybody who works on this or comparable problems -- what the patent laws would refer to as a "skilled practitioner in the art" or something like that. If the much-better-paid-than-me AI teams at Google and other such places didn't recognize that the only sensible way to gather the "training data" for their system was to put cameras and other sensors on actual cars, and then ... gather the training data ... well, then, I guess the only good reason for them making 3 times what i make is that they work twice as hard. or something.

As far as who benefits ... yes, hmm. I have no good answers. Neither, really, does anybody else. There are some people who really aren't capable of doing work that can't be done better by machines. Rather a lot of such people, in fact.

up
0 users have voted.

The earth is a multibillion-year-old sphere.
The Nazis killed millions of Jews.
On 9/11/01 a Boeing 757 (AA77) flew into the Pentagon.
AGCC is happening.
If you cannot accept these facts, I cannot fake an interest in any of your opinions.

Hawkfish's picture

@UntimelyRippd

Probably because the Googs are all software people. As an old computer dinosaur, I am constantly amazed a how few people understand that hardware is part of the system.

(Of course the Uber people are just dumber than a bag of hammers, so no surprise there...)

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

@UntimelyRippd
with Illinois drivers. YMMV. And yes, I am aware of the pitfalls. I was in the embedded software business before the tech wreck.

The US Navy had an automated carrier landing system. Pilots didn;t trust it but eventually accepted it because there were fewer crashes. Not zero crashes, fewer crashes.

Recently there was a horrific accident when an 18 wheeler crashed into a state police car that had pulled another driver off the road causing a fatal multi-vehicle crash. Truck driver was playing with his smart phone and shot off the curve. You can find and fix software bugs, but there is no cure for human stupidity. Just yesterday another expressway crash where three people, one an infant were thrown from the vehicle. None had a seat belt and the baby was not in a carrier. All three dead.

up
0 users have voted.

I've seen lots of changes. What doesn't change is people. Same old hairless apes.

Tesla has yet to make a profit from selling cars. The company is burning through cash at a rate that cannot be sustained. Tesla may go bankrupt within a couple of years. Meanwhile, the problems of operating self-driving cars are not going to be solved any time soon. A main one is the problem of liability--when a self-driving vehicle kills someone, who is responsible?

There are other problems with the whole concept of driverless vehicles. The main thing we actually need in the transportation sector is better, more widespread public transportation. From the point of view of safety and environmental impact, public transportation is far superior to any system depending on individual cars, driverless or not. We need less cars and trucks and more subways, trolleys, trains, etc. That is where the real future of transportation lies.

up
0 users have voted.
Hawkfish's picture

@out of left field

in the presentation (some bankster asked about it in the Q&A). He said that Tesla would probably be liable.

If you accept his claim that AutoPilot today has half the accident rate of human drivers, that implies they could get insurance at half the rate of humans. The big difference is that Tesla is a big fat lawsuit target, so they had better get the terms right - not to mention some new legal protection.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg

k9disc's picture

in general.

We have old vehicles. The computer runs fuel and that's about it. I feel good about that.

I'm terrified of self driving cars in the dystopic Orwellian hellscape we're creating for ourselves. Little assassination machines, IMO - Minority Report meets 1984 meets I Robot.

Musk moving into mining makes sense. He needs assets and experience mining on Earth so he can do it in space.

If you look at Musk's businesses, EVERYTHING is aligned towards colonizing Mars. Boring Company, Space X, Tesla cars and energy, seems like mining fits the production model and could easily provide funding and R & D for future Mars projects.

I'm not sure about him. He's not accepted by the big money people. His move to open source Tesla car tech makes me think he's pushing against the current hegemony, but I'm still not sure.

Interesting piece, Hawkfish. Thanks for writing it. I follow Tesla pretty close and this was full of novel details. Good show.

up
0 users have voted.

“Tactics without strategy is the noise before defeat.” ~ Sun Tzu

ggersh's picture

@k9disc only thing that makes sense, if your comment
is correct. He has to have CIA approval for all
of this, we are all his guinea pigs.

If you look at Musk's businesses, EVERYTHING is aligned towards colonizing Mars. Boring Company, Space X, Tesla cars and energy, seems like mining fits the production model and could easily provide funding and R & D for future Mars projects.

up
0 users have voted.

I never knew that the term "Never Again" only pertained to
those born Jewish

"Antisemite used to be someone who didn't like Jews
now it's someone who Jews don't like"

Heard from Margaret Kimberley

k9disc's picture

from Established institutions. He also gets a lot of support, so I'm not able to call it so cleanly as your comment.

Musk is doing the public front end at best, his highly public tech is behind that of DARPA and the Deep State, I'm sure. Liquid fuel rockets are not of the future. I could see TPTB allowing him to tinker on while yanking his chain every once in a while.

I'm starting to wonder who runs the CIA, which is a VERY interesting question, and it's most likely related to von Braun. I'm starting to think there are some non-state actors that have major geopolitical power and military juice and they pull the trigger on or pile on to Deep State activities.

@ggersh

up
0 users have voted.

“Tactics without strategy is the noise before defeat.” ~ Sun Tzu

ggersh's picture

@k9disc that needs clarification, my thinking was that the Bushes
were very much in charge, could it be the clintons have taken on
that role?

I'm starting to wonder who runs the CIA, which is a VERY interesting question, and it's most likely related to von Braun. I'm starting to think there are some non-state actors that have major geopolitical power and military juice and they pull the trigger on or pile on to Deep State activities.

up
0 users have voted.

I never knew that the term "Never Again" only pertained to
those born Jewish

"Antisemite used to be someone who didn't like Jews
now it's someone who Jews don't like"

Heard from Margaret Kimberley

k9disc's picture

I don't think State actors are in charge. They are hired help. Not sure who is in charge, but I'm pretty sure they're not government entities.
@ggersh

up
0 users have voted.

“Tactics without strategy is the noise before defeat.” ~ Sun Tzu

ownership for a long time. I would assume it is in their blue print somewhere. Uber will love it. Dump the drivers and keeps the whole price of the ticket.

up
0 users have voted.

"Religion is what keeps the poor from murdering the rich."--Napoleon

WoodsDweller's picture

individuals succeeding in sending their car out to earn them money while they're at work.
Some capitalist will raise some money, figure out the most profitable vehicle to use, give it a logo, get the correct insurance, and run his fleet 24/7. He'll be able to undercut individuals. Probably bribe politicians to require cab licenses or something to create a barrier to entry.
So many people, so many miles travelled per day (mostly at peak commute times), means only so many robo-cabs would be needed.

Johnny Cab:
[video:https://www.youtube.com/watch?v=rCUYMM6pABw]

up
0 users have voted.

"The greatest shortcoming of the human race is our inability to understand the exponential function." -- Albert Bartlett
"A species that is hurtling toward extinction has no business promoting slow incremental change." -- Caitlin Johnstone

Hawkfish's picture

@WoodsDweller

Owners will sign up with Tesla's system. But your question was asked in the final Q&A session and I think the answer Musk gave was fuzzy, but I think his reaction was basically either a) we have more cars or b) they can't control the cars remotely. a) is hard to match scale wise as Tesla will have access to ~1M cars at $30K a pop so you would need $30B to compete. b) may be prevented by the public key cryptographic signatures on the software packages the cars will run.

up
0 users have voted.

We can’t save the world by playing by the rules, because the rules have to be changed.
- Greta Thunberg