In The Big Chair - Stan Boland of Five

Five announced $41m in Series B funding a year ago - what progress has the company made since?

To zoom out a bit, everyone in the auto- and tech-adjacent industries has internalized that we are many years away from a time when your average person on the street will see fully autonomous cars (SAE Level 4 or 5) as a common occurrence. 

With that realization I think has come a refreshing sense of purpose for companies like ours - we aren’t wedded to the unrealistic hype and expectations that some see in our industry.

For Five, that meant we took a long and hard look at our long-term strategy and its focus on making deep contributions to bringing realistic, concrete automated driving systems to market by well-funded auto industry companies.  

Since our Series B, we’ve maintained research and development of a prototype automated driving system ourselves, but now with the purpose of helping us productize a highly integrated, comprehensive cloud-based workflow that we deliver to our customers, who can use it to develop, test and design their automated driving systems to achieve production deployment. 

That focus on B2B, not B2C, means we’ve been building solid relationships with several large automotive manufacturers and starting to realize exactly what this industry needs - a merger of skills between tech businesses and also businesses.

Tell us about the highs and lows of 2020 for Five

2020 was a reset for the whole industry, including Five. As a whole in 2020 we started to see a path forward for autonomous vehicle technology, with sharp realism breaking out - meaning clear product targets such as highway driving, trucking and an emerging segment of goods delivery, and less focus on over-optimistic robotaxi goals, which we all accept is pushed out into the future. 

That clarity really helped Five define its minimum viable product and configure its team. We completed multiple proofs of concepts with customers and secured first revenues for the cloud platform we productized.

Of course, we also missed seeing each other during the pandemic, but at Five, we had already been fairly distributed, so the impact wasn’t too great. We pioneered some fun virtual gatherings involving food, games and cocktails, and overall our vice president of people, Michelle Coventry, has made sure that we maintained the culture that has made us a first-choice for a lot of UK software engineers and researchers.

What are the main use cases for autonomous vehicles that you see for the travel and tourism sector?

Long-term, we’re definitely headed to a taxi-like capability that can transport people over well-defined but constrained areas. That’s already being demonstrated in very benign environments in some US and Chinese cities, but the extrapolation of that to very complex urban environments, in variable weather and lighting conditions, is still some way off.

In the meantime, we can definitely expect systems to focus on non-public roads (like entertainment parks), or across fixed routes with no vulnerable road users (like a bus service), or very slowly over more complex routes.   

Many of the traditional auto companies are targeting automated driving features for premium, and eventually mid-market, cars that we as consumers buy.

Most of the service use cases are being pursued by very well-funded tech companies like Waymo, Cruise and Argo. We definitely see even those companies are being much more grounded and realistic in their goals and no longer trying to boil the ocean.

So we can expect to see fixed-route shuttles and slow moving people-carrying in parks, and they will become more functional, cheaper and faster over time.

The autonomous vehicle segment is hot right now with huge investment going into technology development companies, is the investment justified? Why?

That’s certainly true, the reason being that the markets unlocked by automation are measured in the trillions of dollars. In fact we estimate $6.5 trillion by 2030. But it’s a big job to unlock those markets. A major global robotaxi play is a ten year project plus, and probably $10 billion of investment.

But that’s not the only investment activity in automated driving technology today. The breakout of realism across auto industry companies and the concrete product goals they are targeting brings into sharp relief the need to find product fragilities and prove the safety of automated driving systems in a principled way. 

Since the systems are essentially infinitely complex and the world in which they have to be proved safe is also infinitely complex (where we cannot know ahead of time what might be in that world), the auto industry really needs access to machine learning skills, software skills and truly agile development and testing processes, not typical for the well-established auto space.

This is creating really important roles for companies like Five, as well as revenue grants and powerful investment opportunities in the whole sector. 

So yes, there’s a very vibrant investment market, and we expect this to continue and to increase in attractiveness even from here.     

What are your own predictions for the segment?

As I’ve already indicated, even robotaxi firms in the US and China are breaking this over-ambitious goal into more realistic and limited targets and we will actually see these deployed in constrained and relatively benign environments, getting more complex over time. 

We also expect premium cars to begin to offer and expand the first automated driving systems that, for the first time, do not expect and require the driver to be paying attention and in control. That starts with a rather slow highway lane keeping and adaptive cruise control system standardized in a new UN standard called ALKS.

Those speeds and the road complexity will then be increased up to 130kph and will migrate towards more urban-type roads later on.

We’ll see autonomous trucks, especially in big open countries like the US, Russia, Sweden and so on.

And we’ll see some interesting but slow automated driving applications emerge for goods delivery in residential areas, and that will potentially be used for parcel and grocery delivery. 

It’s going to be exciting, so watch this space!

What were your first thoughts when you saw Amazon’s name linked in autonomous vehicle/driving?

It makes sense! For those not familiar, Amazon acquired a very capable Californian firm, Zoox, in mid-2020 for over $1 billion. Zoox had been working on a custom people-moving robotaxi system for urban autonomy.

Over the last decade or so, Amazon has spent billions refining and polishing the last-mile problem in delivery, from logistics and fulfilment to the vehicles themselves.

It was only a matter of time before the company shifted resources towards autonomous vehicle technology. It has been telegraphing their intention to do so for several years, but it has really ramped up its efforts in the last few years through direct investments and acquisitions. 

Last-mile delivery, as explained above, is one of the areas where autonomous vehicles make the most sense, especially in light of the pandemic and where we will likely start to see autonomous vehicle technologies take off outside of consumer vehicles. 

When and why did you first become interested in autonomous vehicles?

Like many of our team, I’m attracted to solving complex problems that apply novel science and engineering, build new services and create economic value. And I have to say that there is probably no problem being worked on today that is as complex as automated driving.

But it does also happen to be a segment where we have the ability to bring together computer vision, machine learning and robotics talent here in the UK, so that makes it a very tantalising space.

We identified this back in 2015 and have been working in the sector for almost six years now.  

What the biggest challenges facing the rollout autonomous vehicles currently? Regulatory/consumer perception/tech/financial…

Boy, that’s a big question and it’s all of the above of course.

The biggest problem is the state space explosion that comes with the very large number of known and unknown environmental factors for automated driving, and the state space explosion that comes with the complexity of an automated driving system that has to take in sensor inputs, make sense of them and determine driving action, safely.

That’s made incredibly challenging because every sensor, every module and every subsystem is faced with constraints and makes errors, sometimes minor but sometimes not. All of the financial and regulatory challenges are consequences of this complexity.

So it’s a science, maths, robotics and engineering problem at its heart. Companies like ours need to figure out exactly where to focus best on this problem space, and we choose building very sophisticated cloud systems for supporting virtual test and validation. In itself, it's already a big space.

Others may decide to try to do everything and that really amplifies the funding challenge. But we can be sure that automated driving is coming and these challenges are being met!  

Given your entrepreneurial background, what other technologies do you have your eye on that might benefit travel and why?

Well I really like electric aero engines - perhaps that’s my early career at Rolls Royce coming through! Plus some of the latest concepts for lightweight mass rapid transit systems that can be deployed sympathetically into towns and cities are also really cool.

Integrating those with local automated vehicles for a seamless service seems like a really good idea. But frankly, we already have so much to extract from the many applications for driving automation to keep us busy! 

What has been the biggest learning for you personally out of the pandemic?

We’ve learnt that we can do most things in a virtual world. 

I’ve also learnt that it’s possible to avoid a lot of international travel and still get at least as much face time with customers. 

It’s also taught us how dependent we are on goods delivery, and that perhaps the priority to automate that needs to be a lot higher up in everyone’s plans.