We’re delighted to announce our partnership with Bell Trucks America to be their exclusive provider of autonomous and ADS solutions.

With this partnership, Bell trucks will be the first commercially available off-road trucks featuring AI-powered “Level 4” autonomy, unlocking an opportunity for mines and quarries of all sizes to experience the savings, optimization, and safety realizations, that autonomy can bring.

We started Pronto out of a desire to build real applications for autonomous technology today, and our partnership with Bell Trucks America is confirmation that our technology and strategy are providing valuable solutions to real world problems.

In the coming months, all BTA dealerships and service centers in the US and Mexico will be able to offer the Pronto autonomy stack to customers of articulated dump trucks (ADTs). To learn more about this offering, send an email to sales@pronto.ai

Onward!

Announcing Our Partnership with Bell was originally published in Pronto on Medium, where people are continuing the conversation by highlighting and responding to this story.

I’ve worked on autonomous vehicles (AVs) for almost 20 years. Throughout my career, I have seen the industry make massive strides in the pursuit of AV and I am grateful to have contributed to some of the key milestones. Over the years, there have been a lot of predictions made, billions of dollars spent, and a huge amount of hype.

The global AV market is estimated to be worth more than $800 billion and AV companies continue to raise many billions every year. However, since I last published my thoughts on the state of the industry in 2018, not much has advanced in terms of AV deployments. The simple truth of the matter is that we are still years away from having driverless vehicles at any meaningful scale on public highways or city streets, and commercially viable AV deployments will take much longer.

Don’t get me wrong, I am still a true believer in AVs and am more excited about the ultimate vision than I’ve ever been. The vision for our roads is compelling: fewer traffic jams, more efficient transport of goods, safer journeys and millions of hours and lives saved. If you imagine the internet as our planet’s nervous system, then autonomous vehicles would be the circulatory system. There’s no doubt in my mind that all of that will still happen. We just need to be smarter about how we can actually get there.

Returning as Pronto’s CEO

In order to bring autonomous vehicles onto streets everywhere, where every car knows where it is going and can get there safely without any human intervention, AV companies face a complex set of challenges. Solving these challenges is both time-consuming and expensive — and with the tech currently available, it is simply not possible. The truth of the matter is we are years away from having driverless vehicles on the roads everywhere.

When we started Pronto our goal was to deliver real solutions to real customers as soon as possible in order to unlock a realistic, evolutionary path to commercially viable autonomous deployments. The last three years have validated many of our assumptions and I am as confident as ever that the approach we’ve taken is the right one. I’m proud to say that we’re still very much on that path and that I’m returning as the company’s CEO.

The Evolutionary Approach to Self Driving

I believe that we can make a real difference by safely integrating new technology with driving in a constrained environment: off road haulage. In 2018, Pronto became the first — and is still the only — company to drive coast-to-coast in the United States without a single driver input to the vehicle’s controls. It was an incredible milestone for the industry and it helped us advance our technology even further. We are now applying that experience and more than 15 years’ worth of other technological advances to A2B by Pronto, the world’s most advanced autonomous haulage system (AHS).

While I believe AV is not ready for driving on public roads, it is ready for off-road. And the off-road industry is ready for AV. Off-road haulage is a multi-billion dollar industry that is mired with concerns around safety and efficiency. So far this year, in the US alone, nine miners have been killed and 185 have been injured in accidents involving powered haulage. Year to date, this is the highest number of powered haulage fatalities since 2006, according to the Mine Safety and Health Administration.

But autonomy is not just an investment in safety, it’s an investment in efficiency. Autonomous haulage can reduce tire wear, fuel consumption, maintenance costs, and operating costs. The benefits for site owners are clear: improved safety, reduced cost, and higher productivity.

What we’re doing with A2B allows the mining industry to realize these benefits now. With our cutting-edge technology, mines can get up and running years faster than with existing systems, and have a return on investment in their first year of operation. We’re able to do this because we’ve upgraded our on-road autonomy stack for off-road driving by combining it with data collected from mines in the Arctic Circle to the Australian Outback, and everywhere in between. The result is what we believe to be the most advanced AHS in the world — and we are very excited to see our customers begin to roll it out.

It’s Time to Deliver

As I see it, the best way to advance autonomy for the long run is to work on products that deliver value — and save lives — right now. So, instead of chasing the dream of on-road AVs, we’re focused on shipping AVs to off-road customers today. We’re working with some of the world’s largest mining companies as well as some smaller operations to roll out A2B, and I’m grateful for the chance to be back.

It’s time to deliver.

Anthony

Simplifying Autonomy with Pronto was originally published in Pronto on Medium, where people are continuing the conversation by highlighting and responding to this story.

One thoughtful public response was from Brad Templeton in Forbes. Even though Brad challenged some of my data, assertions, and conclusions, I am grateful that he embraced this type of dialogue to elevate the current safety conversation. I’ve followed Brad’s writings closely for several years; he is a high-integrity individual and a respected thought leader in our field. One specific assertion in my post that Brad and others in the media and across the Twitter-verse keyed in on was that, based on publicly available data, level 4 autonomous vehicle testing fleets in California crash more than average drivers. This seems difficult to believe and there are several reasons why we should be skeptical. The conclusion that level 4 autonomous vehicle testing fleets in California crash more than average drivers means that the AV industry is putting the public at risk far more than it should be. I want to unpack my thinking and data more here since this is a serious claim.

To give autonomous test fleets the chance to look their best, I decided to use the SHRP 2 dataset, which was age-adjusted in a Virginia Tech study to account for some age groups being over-represented in the original dataset. This data also tries to account for human drivers severely underreporting crashes. Put simply, the SHRP data makes human drivers look worse than the other credible sources of data on traditional driving. (As others have noted, NHTSA data suggests that humans are actually several times better than this. I’ll stick with SHRP to try and help the prototype robots out so that they look the best as possible when compared to human drivers.) An interesting side note: it was reported last week that GM Cruise was also using the Virginia Tech modified SHRP 2 dataset to define their launch milestones.

The SHRP 2 data was first used to compare human driving to AV fleets several years ago after a University of Michigan study pointed out that Google was involved in a higher than average number of crashes. Google responded by commissioning the Virginia Tech study mentioned in the prior paragraph. That study used SHRP 2 to argue that AV testing fleets were safer than human drivers if we took into account unreported crashes. Below is a graph pulled directly from that study. SHRP 2 Age-Adjusted refers to manual driving. Note that references to “Levels” in the graph below do not refer to the SAE levels of autonomy, but rather types of crashes.

An interesting thing to note is that in both the Michigan and Virginia Tech studies, the safety of AV test fleets is supported by referring to a lack of at-fault crashes involving prototype autonomous vehicles. Such claims have been repeated in many reports and media accounts since then. In my previous blog post, I noted how safety theater uses legal liability rather than preventability to argue safe driving. It turns out, self-driving vehicles are significantly worse at avoiding preventable crashes. We all encounter situations on an almost daily basis where a driver could easily cause a crash, even on purpose, and yet not be legally at fault. Anyone who has had a vehicle in front of them brake unnecessarily hard can confirm this.

In my analysis of the most recent data, I tried to use a nearly identical methodology as the Virginia Tech study. I reviewed each AV crash report filed with the DMV and only counted crashes that occurred in autonomous mode or immediately after disengagement. If an injury was reported, I counted it as an injury accident (Level 1 in the graph above). In the Virginia Tech study, they used data from inertial measurement units to differentiate between crash types defined as Level 2 and Level 3. I combined Level 2 and Level 3 crashes into one field as I did not have access to that data. I did not count incidents where a test vehicle was attacked by angry members of the public (which happens surprisingly often), accidentally struck by a golf ball, or other things that might make you think that an autonomous “crash” was not really a crash. In order to get the crash rate, I added up the Total Miles Driven for each fleet (which is provided in the DMV Disengagement Reports from 2015–2018) and divided by injury crashes and property damage-only crashes.

Despite all this, my analysis still shows that in the years after the Virginia Tech study was published, the reported rate of “autonomous mode” crashes in California more than tripled: it went from 8.8 crashes to 27.2 crashes per million miles by the end of 2018. The national average, according to that study was 20.2. And it’s even worse than these numbers suggest because the 27.2 figure for level 4 fleets includes prior years, when fleets used to crash less frequently; while the 20.2 figure for human drivers has improved over time. Also, my figures don’t take into account that most of the testing in California has occurred in Santa Clara County, which has a crash rate that is less than half of the national average in the comparison I’ve outlined. We haven’t even begun to do this analysis in Arizona and other states where the majority of level 4 prototype fleet testing occurs, because there are no reporting requirements on crashes or miles driven.

The implication of the data is that today in California, a professionally-trained driver — whose job it is to monitor the safe road performance of a system that is supposed to be capable of driving more safely than a real person — is actually significantly more dangerous than the average driver. It’s not looking good for 2019, either. We don’t yet have the data for miles driven so we can’t calculate the crash rates, but there has been a noticeable uptick in crashes in the last four months when we look at the number of crashes that have occurred this year compared to last.

Notably, just two companies out of the 62 companies with testing permits in California for level 4 test fleets were involved in the 132 out of 156 reported crashes. These companies also by far drive the most miles in California. The takeaway from this should be that if companies are claiming to be so far along in autonomy, they should hold themselves accountable by demonstrating, in a clear, transparent manner, that they have achieved a certain amount of progress to justify this amount of driving on public roads. I was encouraged to see Brad’s recent article in Forbes calling for test programs driving many miles to go beyond government-mandated metrics and be more accountable in getting safety right. If your test fleets with safety drivers are crashing far more than human drivers, it should be clear that you are nowhere near being able to deploy real level 4 vehicles. As a result, the amount of testing should be limited in quantity and constrained only to areas where testing can be performed at least comparable in safety to a human driver.

We clearly need to shift the thinking around testing from one where companies try to show progress by driving lots of miles, to one where they are incentivized to make as much progress on as few miles as possible. This latter philosophy is what we subscribe to at Pronto. I am really encouraged to see that we are not alone. Aurora Innovations put it well in their safety assessment report: “We keep the fleet only as large as our engineering team requires to develop the driver. Others have maximized fleet size in order to maximize the number of miles traveled, assuming that such a strategy will maximize their learning. We see such a strategy as an unnecessary expansion of the fleet’s driving risk. Such a situation creates an excess of data at little to no value.”

At Pronto, we made a decision to not have a level 4 prototype fleet today at all. We believe that a full level 4 system is many years away. Our engineering team can get all the learning it needs to improve our software by developing a commercial-grade level 2 product, Copilot. We have made a conscious choice to build a product that is designed to augment, not replace, a skilled driver. We believe this sets the right incentives and leads to safer roads — a human driver plus Copilot is safer than a human driver alone.

An attentive human driver (Copilot includes a driver monitoring system to ensure attentiveness) is the central safety feature and requirement for a robust commercial product. This demands a much higher safety bar and keeps the system simple, while still providing incredibly valuable learning (since our customers would be driving their trucks on the exact same routes even if they had not purchased Copilot). At Pronto, we are excited to be delivering Copilot to customers soon and making the roads safer for everyone.

Shining a Spotlight on the Data Behind Safety Theater was originally published in Pronto on Medium, where people are continuing the conversation by highlighting and responding to this story.

We drive too many miles, which puts lots of people at risk

In the early days of self-driving vehicles, we thought that more time on the road meant greater progress and likelihood of success. ‘‘Total miles driven” became an important metric for judging progress, and unfortunately it remains a central theme to this day. Companies deploy hundreds of prototypes to collect low value miles, but don’t seem able to make use of the collected data. In reality, total miles driven is a leading indicator of how likely a self-driving vehicle program is to cause injury or property damage. The unjustified focus on “total miles driven” puts pedestrians and the motoring public at risk and undermines a real safety conversation. High quality miles are what we need, not countless vehicles driving for the sake of driving. Self-driving prototypes should only be deployed on a limited basis; safety claims should not be anchored to the absolute number of miles driven or simulated; and companies that drive many miles on public roads should be open to routine and unannounced, independent, third party safety audits based on total miles driven.

And so we injure too many people

Taken as a whole, companies that have registered with the California DMV for AV test permits are involved in more injury crashes and product damage-only crashes than the average driver. In 2017 and 2018, prototype self-driving cars were involved in over 100 incidents in California alone, most of which were in autonomous mode. This is insane considering most autonomous vehicle testing occurs outside of California. I want to underscore that an industry built on road safety potential, with professional safety drivers behind the wheel (whose job is to ensure the safe operation of a vehicle), is somehow ending up in crashes more frequently than average drivers in regular vehicles. The reason this is not widely known is because our industry masks the data instead of implementing practical safety solutions.

But — when we get into crashes, we point fingers at others

Self-driving programs should focus on preventability rather than at-fault collisions. The further implication is that AVs drive well, but human drivers crash into them. Generally, self-driving vehicles with safety drivers do a good job at avoiding collisions where they would be found at-fault. However, they are significantly worse at avoiding crashes that should be preventable. This is because self-driving technology has not advanced enough and safety drivers don’t intervene early enough.

Safety theater causes real problems

The miles per disengagement metric is a bad metric for measuring progress and is not meaningful in terms of safety. Companies inflate their miles per disengagement to appear further along and use their own absurd definitions of what a disengagement is — effectively erasing thousands of safety-related disengages. Companies that report their disengages honestly appear much further behind. Furthermore, this metric incentivize safety drivers to not disengage in unsafe situations.

Relying on the wrong metrics or not looking at real data has unfortunately propelled us into the realm of safety theater — meaning creating the illusion of safety instead of actually delivering on safety. This type of safety rhetoric is more harmful than helpful. For example, NHTSA provides an appropriate forum with its voluntary safety assessments, but companies misuse this well-intentioned process as a public relations tactic to submit glossy safety brochures with little substance. Companies should take these reports seriously and report on real safety efforts by showing that if they are driving many miles, their collision rates are better than human drivers. They should also be able to justify the number of miles driven, not just proudly declare that they have been driven. Autonomous vehicle companies have a responsibility to speak up about efforts they are making to address safety.

In addition, proclamations by AV executives that fall into the categories of general safety platitudes, mixed with bogus disengagement numbers, and brashly optimistic product launch dates create the perception that prototype vehicles are safer and further along than they actually are. Such statements are dangerous because they create an environment at some companies that they must win a race to autonomy or else close their doors. This thinking wrongly motivates companies to deploy hundreds of test vehicles prematurely and at the expense of safety. It is important that we have real conversations about the current capabilities of the technology and how far it still has to go.

So why Pronto?

Given my concerns with safety in my industry, some may ask why I’m working at Pronto and, specifically, with Anthony Levandowski, who has dealt with a lot of criticism with regards to safety. I’ve worked with Anthony for a long time and I know that some of Anthony’s comments and actions on safety have been taken out of context or were inaccurately portrayed. All that being said, I do not think that Anthony has gotten everything right with respect to AV safety. He has made mistakes and said brash things. But, Anthony is someone who is always looking to improve himself. What I value most about him on safety is that he doesn’t resort to safety theater. He is committed to getting AV safety right. All of us at Pronto are. At the end of the day the only way to improve road way safety is to deliver a safe product. Anthony and Pronto are on schedule to launch a Level 2 system which will make the trucking industry safer. And that is why I am excited to be at Pronto.

Closing the Curtains on Safety Theater was originally published in Pronto on Medium, where people are continuing the conversation by highlighting and responding to this story.

I know what some of you might be thinking: “He’s back?” Yes, I’m back.

The “why” is fairly simple: I’m back because it’s my life’s passion to make the life-saving potential of autonomous vehicles a reality. “Reality” means real autonomous vehicles solving real problems for real people, while actually improving safety on real roads.

Over the past 15 years, we’ve witnessed numerous advances in self-driving technology. I’m proud to have played a big role in it. At the same time, I’ve admittedly grown frustrated — and at times impatient — with the industry’s inability to deliver on its promises.

Instead of delivering products that could help people right now, we talk of “visions” or the “the future of mobility” ad nauseam. Rather than democratizing the technological advances we’ve made for everyone’s benefit, we reserve them for extremely narrow and non-representative “demonstrations” or “proofs of concept” that have scant hope of ever scaling in a commercially-viable way. And we avoid intellectually honest conversations about road safety through aspirational platitudes and glossy marketing materials that are passed off as “safety reports.” Collectively, we have the resources and capabilities to do much better.

In the past, my refusal to toe the corporate line by pointing out the above has landed me in hot water. Just as my openness to bluntly address technology’s capabilities and its limitations have led to me being labelled “reckless”. Road safety is not something that is well-suited to sound bites and short attention spans. It involves tradeoffs and important, nuanced discussions. I am disappointed in how shallow the discourse on this front has been and continues to be. But I also recognize that some of my prior comments have not helped.

Over the past 18 months, as I’ve been painted into a villainous caricature, I’ve had lots of time for introspection. In some sense, the deeply personal hardships my family and I’ve had to endure have provided an important opportunity to step back and honestly reassess my past as well as plan for the future. One thing I continue to stand by is my safety record, which is second to none in this industry. My teams have a demonstrably impeccable track record. At every company where I have worked, safety was better during my tenure than before or after. On the safety front, I always look forward and am proud to work with and be challenged on a daily basis by the best in the business.

Today, I want to share a few more things I’ve recently learned that I think will prove central to making autonomous vehicles a reality. To begin with, several trends have become increasingly apparent:

First, despite the vast sums of money and time dedicated to developing and rolling out autonomous vehicles, there are no real autonomous vehicles today. There are only increasingly complex and expensive demonstrations of those visions I mentioned above.

Second, true level 4 or 5 vehicles will not arrive for many more years.

Third, the reason why nobody has achieved this level of functionality is because today’s software is not good enough to predict the future. It’s still nowhere close to matching the instincts of human drivers, which is the single most important factor in road safety.

And fourth, traditional self-driving stacks attempt to compensate for their software’s predictive shortcomings through increasingly complex hardware. Lidar and HD maps provide amazing sensing and localization of the present moment but this precision comes at great cost (with respect to safety, scalability and robustness) while yielding limited gains in predictive ability.

Put simply, the self-driving industry has gotten two key things wrong: it’s been focused on achieving the dream of fully-autonomous driving straight from manual vehicle operation, and it has chased this false dream with crutch technologies. So we now find ourselves in a position where regulators, investors, and the public are waking up to the reality that nobody is ahead in the race to deliver autonomous vehicles because the real race is a marathon that has not yet begun. The last 15 years we’ve mostly idled at the starting line.

Please don’t take this somewhat sobering assessment of our industry as pessimism or cynicism. After all, I do have a well-deserved reputation for brash optimism. Over the past three years, amazing gains in machine learning and a new breed of tensor processing hardware have made it possible to pursue a different, ultimately much more promising, path toward solving the self-driving challenge. That’s what my new company — Pronto — is all about. I’ve reconnected with a few old and very qualified friends and gotten to know the best young minds in our industry that have fresh, truly innovative perspectives. Together, we are building the best self-driving stack in the world.

To be the best means overcoming the biggest challenges that self-driving technology faces head on. Our approach? Much better software. After all, the best and safest drivers don’t necessarily have the best eyes. They have the best brains and the most experience. We are building neural networks from the ground up that combine experience-based AI, end-to-end deep learning, and crowdsourced data with advanced computer vision to deliver a highly scalable and flexible driving stack. Nobody else is doing this.

We are not building technology that tells vehicles how to drive. Instead, our team of engineers is building tech that can learn how to drive the way people do.

Our tech does not shy away from the rich complexities of real-world driving. Through better prediction and decision-making software, we are able to navigate previously vexing “edge cases,” such as very low light, direct sunlight glare, heavy rain, snow, construction zones, etc. in a safe, scalable, and repeatable manner on a wide variety of highways without mapping them. Because — let’s be honest — “edge cases” is actually a misleading (albeit convenient) term of art. These adverse driving conditions and much worse occur regularly in the real world. In order to roll out autonomous vehicles at scale, we are tackling them first rather than saving them for last.

Our new approach has already enabled us to make great progress. We drove a vehicle coast-to-coast without any human intervention — a technical milestone accomplished in October (link). But we will not claim it to be something it is not. The age of autonomous vehicles crisscrossing the country by themselves is still quite a ways off, as I mentioned above. Ours is a highly-capable “level 2” system and not more.

Or perhaps I should say “highly capable level 2 system and not less”? Because for all the talk of going straight to level 4, we think it’s much more exciting and promising to first develop and scale a truly great level 2 system. Better to do several things — braking, steering, and throttle — super well on a wide variety of real roads and conditions rather than attempting to do everything else that driving entails in a very artificial manner. The reason this is so exciting is that we can capture some of the benefits of self-driving technology today prior to the arrival of the perfect autonomous vehicles of tomorrow.

The first step is to deliver a commercially-viable ADAS product that makes driving safer for everyone. It augments the driving experience by reducing the cognitive workload for drivers, allowing them to focus their full attention on monitoring the road ahead. The market that we believe makes the most sense to engage first is the commercial trucking industry, which stands to benefit tremendously from an aftermarket product that will help truck drivers deliver their cargo anywhere in the world with greater safety and comfort than ever before. Another bonus, it’ll cost orders of magnitude less to deploy than anything else that’s been promoted so far.

We’re excited to announce our first product: Copilot by Pronto. Copilot delivers the most advanced ADAS features that have previously only been seen in the top luxury cars — and much more — but it is built specifically for Class 8 vehicles, with driver comfort and safety top of mind. I am sure you have many questions about product specifics, our customers, and how we will ensure that it is used properly. We will share lots of details on all of that and more as we roll out Copilot in early 2019. It will be the first stand-alone real product for a real market that the self-driving industry has delivered, and we are excited to be blazing the trail to the reality of how autonomous vehicles will ultimately come to fruition.

But for now I’ll end by just saying that I’m super excited to be back. I’m excited and humbled to have an opportunity to lead an amazing team. And I’m more excited than ever to accelerate the real-world introduction of self-driving technology in a safe and business-minded manner.

Drive safely.

Pronto Means Ready was originally published in Pronto on Medium, where people are continuing the conversation by highlighting and responding to this story.