Who Is The Juicero Of Self-Driving Cars?
Why do startups fail? It starts with the investors.
Remember Juicero? The guys with the juicer that required connectivity, a subscription and bags of raw fruits and vegetables you could press by hand? It cost $400. Then it was $200. Then they shut down. How could they lose $100M+ of investor money? It’s easy to say it was something no one wanted or needed, but that doesn’t tell the full story. No, you’ve got to start by looking at the underlying assumptions that led allegedly smart investors to lose their shirts. Deconstructing the Juicero pitch will answer the fun question:
Who is the Juicero of self-driving cars?
First, a look at how the Juicero pitch went down:
Juicero: It’s a juicer. It’s connected.
Investors: Connectivity is hot and getting hotter.
Juicero: It’s a subscription model.
Investors: Recurring revenues are good.
Juicero: Raw fruits are messy. Buying them is a hassle.
Investors: Customers hate work. How much do you need?
The rush to invest in solutions for problems that don’t exist is at the heart of startup failures, but it was the myopia behind each of the underlying assumptions that really torpedoed Juicero. Just because connectivity is good doesn’t make for a good connected device. Subscriptions? If subscriptions universally made sense, there would be a subscription gun startup, probably based in Texas, but there isn’t. Do people really hate raw fruits? Is buying them a really a hassle?
Real solutions start with needs, not methods. No assemblage of methods—regardless of how clever, timely or popular it is—can invent an enduring need. If the assumptions behind methods are flawed, you’re dead in the water.
Take connectivity and subscriptions, replace bags of juice with self-driving cars, and you have a recipe for disaster.
You also have a fun game.
It would have been too easy to make a list starting with Faraday and ending with companies you’ve never heard of. It wouldn’t have taken long, and trust me, I wanted to, but then I found an article on TechCrunch that saved me fifteen minutes of work. Just when I thought the stench of self-driving clickbait propaganda couldn't get any thicker, two famous venture capitalists delivered a hilarious display of catchphrase-filled nonsense written with the insight and grammatical errors of an intern.
The “authors” are Tom Alberg and Craig Mundle of Madrona Ventures. Their resumes are weighty, but very smart people are often just people who make bigger mistakes. We shall dissect their article line-by-line, and in doing so find our Juicero on wheels.
That’s a nice title. What does it mean? Nothing, and everything. Of course we should plan for it, but first we have to know what “it” is. What plan does Madrona want? One they can profit from. Does that make it good for anyone else? Better than what we have? I like plans, but there are good plans and bad plans. We have to dig deeper to understand what they’re trying to get at. We don’t have to dig far.
The arrival of autonomous vehicles bring the prospect of improved transportation systems without the capital costs, operating subsidies and construction delays of new highway lanes and fixed rail systems.
Nothing like an opinion stated as fact. Is there a prospect of improved systems without the costs, subsidies and delays they refer to? Only in the utopian dreams of urban planners and VCs trying to sell “solutions” based on anticipated growth rather than success in the real world. Tom Vanderbilt’s seminal “Traffic” explains it for laymen. Autonomy doesn’t necessarily eliminate friction or inefficiencies. It shifts them within the very transportation systems everyone wants to improve. Traffic rises to meet capacity. Study fluid dynamics. The same “experts” touting autonomy say we need massive V2V infrastructure for it to work. That means massive capital costs, construction delays and endless repairs to a universe of as-yet undeployed hardware in and on our streets. Cities can barely maintain our current infrastructure. Just wait until it’s connected. So much for savings.
Cities, states, and the Federal Government, need to revise their transportation planning accordingly.
That makes sense, in a Malcolm Gladwellian, conventional wisdom-as-philosophy kind of way. Also, water is wet, and you don’t need two spaces after a comma.
This might be a nice time to check out the three images attached to the article. What’s happening in this one?
Nothing of value. A car with a wifi symbol. A man not driving. A clock. A phone. An app.
And what about this cheap stock image?
Nothing to see here.
What about this one?
That one had to be free, right?
Let’s keep going.
Autonomous vehicles have gone from a Jetson-like dream to a clear reality in less than one decade.
Are autonomous cars a clear reality today? Nope. NHTSA/DOT policy is a mess. The Automation level charts are vague and stupid. No one has gotten a Level 4 car cross-country without disengagements, and no one knows when that will happen. There are multiple cars coming out classified as Level 3, and their behavior is radically different. Low cost solid state Lidar still hasn’t hit the market. By Madrona’s clock, the reality of self-driving cars will hit in 2020. Automakers are saying 2025 at the earliest, and that’s for highly geofenced Level 4. Have the Madrona boys spent much time in anything better than a demo car on a closed course?
I’m all for autonomy, but this is nonsense.
In 2010, when Google first started developing autonomous vehicles, people asked, “Why are they wasting money on this? That’s never going to work.”
You could say that of a lot of things, and it would remain true. You know what they say about broken clocks? For every technology that works despite naysayers, there are dozens that never do. Hello, flying taxis. I’m bullish, but that Jetsons future remains decades away
Today, we have not only seen public pilots of autonomous vehicles from companies like Uber and recent announcement by automakers such as Audi that it plans to begin selling, in 2018, a production car with Level 3 autonomy (meaning it requires no human attention to the road at speeds under 37 miles per hour), we have also begun to see striking data on the benefits of autonomous vehicles.
WTF are they talking about? THAT IS NOT THE DEFINITION OF LEVEL 3. Here’s the actual definition, courtesy of SAE, also known as the Society of Automotive Engineers:
“L3 Conditional Automation: The driving mode-specific performance by an Automated Driving System of all aspects of the dynamic driving task with the expectation that the human driver will respond appropriately to a request to intervene.”
Dear Madrona: fire your intern. Also, there’s a thing called Google. Audi’s system is but one variant within Level 3, and it is a semi-autonomous system, which does not make it a self-driving car. Your intern should have known better. Ive read your resumes. So should you.
As for that striking data on the benefits of autonomous vehicles...
For example, after the National Highway Traffic Safety Administration investigated a Tesla Autopilot crash, they found a 40% decrease in traffic accidents when Tesla’s Autopilot feature was enabled in cars.
I am a strong supporter of any efforts to improve safety, but we still don’t have clarity as to which component of Tesla Autopilot may have led to the 40% figure. Since Autopilot is only a semi-autonomous technology, ascribing any safety benefit to autonomy in general is disingenuous. As someone who has spent a mountain of time using Autopilot, I can tell you that figure is likely derived from the Automatic Emergency Braking system, which is a sub-component of ADAS, which stands for Advanced Driver Assistance System, all of which falls around Level 2. When it comes to Level 4 autonomy, no matter how many bricks you stack, you will not get there without gluing together the component technologies. Where's the glue? Maybe Madrona should invest in that.
When investing tens or hundreds of millions of dollars, know your terminology.
In addition to significant reductions in accidents, the benefits of autonomous vehicles will also include less congestion, reduced emissions, reclaimed productive time, fewer new roads, reclaimed parking space, lower transportation costs for all and improved mobility of the elderly and disabled.
The bigger the problem, the harder it is to solve. We can solve a lot of things, but there is one thing we cannot solve for, and that is human nature. People will buy EVs without autonomy. Build real rail in California and traffic on the 5 and 101 will go down. Are there use cases for self-driving cars anytime soon? Absolutely, but all the near term use cases combined do not come anywhere close to reaching a a tipping point. Massive, multi-decade cultural changes need to take place before we see real autonomy on a mass scale. Well intentioned platitudes can’t and won’t get us there.
As traffic planners across the nation wrestle with the issues of moving people and goods within and between cities, there are a variety of transportation options to consider with everything from Hyperloop, to highways, bridges, buses, and light rail.
Nice SAT writing here. I don’t see anything about driver education, which would solve most of the overwhelming majority of the safety issue, saving hundreds of billions in autonomous R&D. Not very profitable for VCs, though.
One thing that these projects all have in common is large-scale infrastructure projects which take a long time (often multiple decades) and require large up-front capital investments. Examples include the Big Dig in Boston, which cost nearly $15B, began in 1982 and was not completed until 2007; and in Seattle, the Alaskan Way Viaduct replacement, which began in 2001 and is not expected to be completed until 2021, at a cost of more than $4B.
Incorporating autonomous vehicles into our transportation systems, on the other hand, is a very low cost way to drastically improve the flow of goods and people within a region as long as we begin now to make the policy changes that will allow the benefits to be achieved as the technology comes to market.
I suppose investing in autonomous vehicles may be cheaper than the infrastructure they refer to, but it’s still way more expensive than improving driver's ed, which isn’t sexy. Ubiquity of autonomous vehicles is also decades away, and a mixed environment of human and self-driving cars is likely to exacerbate traffic, not improve it.
Here in Seattle, we are proposing to convert Interstate-5 between Seattle and Vancouver, over a number of years, into an autonomous vehicle-only highway. Seattle and Vancouver are vibrant cities, but the transportation between them is tedious and impedes valuable economic partnerships.
That sounds fun, but dedicated a lane does not an autonomous Utopia make. Not even close.
Our proposal, which could be applied to many main Interstate highways and local limited-access thoroughfares, is to begin turning carpool lanes into autonomous vehicle lanes as early as next year.
Too bad there won’t be any Level 4 or 5 cars that can use it for a long time. There will be tens of millions of human drivers who can’t however, increasing traffic in the remaining lanes. Unless you add lanes, which gets you back to square one re: construction costs. Have fun with that.
Recall that HOV lanes were created as an incentive for commuters to change their behaviors. It is time to use that incentive to accelerate the move to the fututre architecture of transportation. As the number of autonomous vehicles grows over the coming decade, we could gradually dedicate entire lanes exclusively to AVs (and perhaps fit three AV lanes into the space of two traditional car lanes). Eventually, the entire highway would become autonomous-only.
Great ideas, but no Level 4 or 5 cars on the horizon.
While some stakeholders believe it is still too early to begin planning for autonomous vehicles on public roads, at the current rate of technological progress and with the early data suggesting drastic improvements in traffic safety, we believe we will reach a major tipping point during the coming decade, well within the time frame of a major transportation project.
Malcolm Gladwell must be popular at the Madrona office. Vanderbilt? Not so much.
The very same computing advances that allow your cell phones, computers, tablets and other gadgets to recognize you and respond to your voice now underpin the enabling of cars to achieve more and more autonomous operation. In the computing world these advances proceed quickly, with society adopting them en masse.
Yes, computing speeds are increasing, but computing speeds don’t equate to building the hardware and software stacks necessary for the autonomy Madrona is talking about. Data is not information, and processing it requires machine learning/AI that still isn’t ready.
The only thing that will delay the arrival of autonomous vehicles in the U.S.A. will be timidity on the part of regulators and legislatures to give the public the chance to take them up at a similar rate.
The platitudes just keep on coming.
There are still many questions in introducing autonomous vehicles into our current transportation systems– what level of autonomony will be required for inattentive driving under various conditions, will most people continue to own cars, how will street parking change -- but dealing with these issues is clearly less expensive than the massive investment of building new highways, bridges, or rail lines.
Hard to take this seriously when the author(s) didn’t spell check “autonomony”.
Autonomous vehicles offer a plethora of benefits to cities and their tax payers -- fast to market, low investment requirements, and many societal benefits ranging from fewer accidents to broader access to low-cost transportation.
“Tax payers” is usually one word, but that’s not as egregious as saying autonomous vehicles offer the benefit of being “fast to market”. If I was investing in a VC fund in this sector, it wouldn't be Madrona. Does anyone there read Reilly Brennan's Future of Transportation newsletter?
Transit planners across the country should accelerate their consideration of the importance of autonomous vehicles as they map out their plans for the future of city and interstate transportation. These changes will be as important as the construction of the Interstate Highways System in the last century. It is time to get going…
NHTSA/DOT and Congress have already given up making any serious effort to regulate the relevant technologies. Without subsidizing autonomy itself, there is literally nothing they could do to speed it up.
So, what is Madrona's plan? Invest in anything in sector, it seems. VCs can afford to lose, on the assumption a handful of investments have big exits. Based on their article, they're jumping on the bandwagon.
Who is the Juicero of self-driving cars?
It’s whomever in sector Madrona invests in next.
Alex Roy is Editor-at-Large for The Drive, Host of The Autonocast, co-host of /DRIVE on NBC Sports and author of The Driver, has set numerous endurance driving records in Europe & the USA in the internal combustion, EV, 3-wheeler & Semi-Autonomous Classes, including the infamous Cannonball Run record. You can follow him on Facebook, Twitter and Instagram.