After studying for an engineering master’s degree at the University of Oxford, he moved to the University of Sydney to obtain his PhD. His storied career has taken him from subsea navigation to the Australian Centre for Field Robotics and MIT, before returning him to Oxford.
Today, he serves as the BP professor of information engineering at the University of Oxford and the chief technology officer of Oxa (formerly Oxbotica), a company pioneering universal autonomy.
As the UK makes moves to bring automated vehicles to our streets, Newman spoke about the “metadiscipline” of engineering, the shift toward industrial automation and why we shouldn’t fear the rise of “embodied AI”.
How did your journey into engineering begin?
When I was six, my parents asked: “What do you want to be?” And I said: “I either want to be the pilot of Thunderbird 4 or an inventor.” Since I was a kid, I was always a maker, whether it was making a smelter or a programmable robot out of Lego, or going round town with 50p and buying bits of electronics, trying to figure out how the hell it all works and being totally thwarted, but loving the idea. I’m an inveterate maker. I’m an engineer!
How did that interest evolve?
I was fortunate enough to get a place to read engineering as an undergraduate and, oh my god, I loved it. Every year, it was like: “Oh, wait, the previous year was a bit fake. And this is actually how it works.” There’s always the next level of security clearance. There’s always the next thing.
Then, for me, the big moment was when I decided to do a PhD. I was the first PhD student to go to the Australian Centre for Field Robotics, which then became a behemoth of robotics. My supervisor, Hugh Durrant-Whyte, said: “What do you want to do?” And I said: “I’d like to build a submarine.” He said: “Go on, then.” Right… so I did.
During your time in Australia, you began working on how machines navigate. What was the core challenge?
I remember meeting with a friend. We were crossing Parramatta Road in Sydney and he said: “I’m working on this problem: can a machine never get lost?” If a machine has sensors on board, without being given a representation of the world, can it just figure out where it is? That was the SLAM problem – simultaneous localisation and mapping – and I was fortunate enough to be working on it at the time. And that’s now sitting on your phone.
There’s something about robotics that I find utterly fascinating. It’s about compromise. You have finite compute, finite size, finite energy, finite time, but you’re immersed in a physical world where Newton matters. If you make a decision and it’s a two-ton vehicle, that maths, that decision, is still with you 10 milliseconds later.
How do you tackle problems as an engineer?
I’ve got this thing. I never go for a walk and say: “I have an interesting mathematical problem” – though I have many, many interesting mathematical problems to figure out. What I tend to do is go: “I’m really pissed off that a robot can’t do that. Why can’t they do that?” Then I find that all the problems that are in the way will come up immediately in front of you. Examples would be spatial, temporal calibration, multi-modality localisation or perception. It would be data-driven planning. These are all things that are just in the way of the goal of building machines to do something useful.
Given the current hype, many might call Oxa an AI company, but you seem to prefer engineering company. Why?
It’s a kind of engineering, isn’t it? We’re a robotics company that uses the right tools to do that, which, of course, includes AI. We also need to mix it with other things. There are techniques that are just mind-blowingly powerful, but they need to be couched and supported by other techniques that are not AI.
That’s an engineering job, fusing these different solutions together, and engineers build products, right? Those products go and do something useful. I think engineering is the metadiscipline. Saying that somehow AI is separate from engineers – no one should be thinking that because we’re the ones that engineer the AI to do something useful for us.
We’re speaking as Claude Code is taking over the world. Some worry that autonomous coding agents and generative AI will take the engineering out of engineering. What’s your take?
That’s like saying: “You know what, I really think we should stick with punch cards.” Or, you know, “we should use assembly code”. That’s nuts. The whole engineering enterprise is a tide that lifts all ships.
What I see is great engineers going: “Awesome. I’ve got a way better spanner now.” It’s the speeding up to build things we need to do that is fantastic. And there is no excuse for not using it. None. You must. I’ve had the most amazing experiences using it. You want your engineers to be as much in that “I wonder if” creative moment as possible and as little in filling in boilerplate stuff to get to the bit you want to do.
Where are we in the UK with autonomous vehicles?
We’re going wholeheartedly into IMA (industrial mobility automation), and so we’re looking at ports, we’re looking at airports, we’re looking at warehouses, we’re looking at yards where we think the economics makes sense. If you’re trying to interfere with an automotive supply chain because you want to do passenger cars, we say good luck to you, because those guys do not move fast. They are very hard to interfere with and it’s a race to the bottom on price.
Another thing I’d say is that the approaches some are taking on doing the on-highway stuff seem staggeringly expensive. It seems mind-boggling that you need billions to train something to do that. There are cheaper ways to do this.
How does the complexity of a port or an airport compare with driving on a public highway?
I wouldn’t say it is a simpler environment than on road in terms of what’s going on, but there is lower variance in the wildness. In the ports of Sydney and Botany [a suburb of Sydney], no humans came in so they closed it off.
From a business perspective, from an engineering perspective, that makes a ton of sense. You can pick those off in terms of economic order, and you could pick them off in terms of an engineering order. That’s much harder to do if you’re building a passenger car and it will drive itself because you’ve got to do the whole thing.
When happens in a world where things get automated at scale?
I think the question you’re getting at is: what about the labour market? I went back to Brisbane after 20 years of doing the first automation there and talked to the stevedores. They said: “Look, this has been interesting because there’s not been a single accident. We’ve not had one person being maimed.” And we don’t have people having a horrible life out in the baking sun working these crates. They’re indoors. The job has changed.
I’m very positive about technology. It does make new jobs. We’re already hiring individuals for a completely new type. Our autonomy solutions engineer is kind of like when IBM made their first machines. Those machines that got sent to businesses came with people who helped operate them, helped maintain them and helped train the business on how to use them. That led to companies having their own IT departments. That’s what we’re seeing now – new types of jobs that just never existed because of the autonomy.
I can imagine a world where someone says: “What’s your job?” And they’ll go: “Oh, I’m a driver. I drive 17 vehicles in, you know, Perth, Brisbane, Kuala Lumpur.” Some call me pathologically positive around this.
Do you feel the perception of engineering as a profession is finally shifting in the UK?
It’s always been the case that in German it’s Herr Engineer, whereas over here an engineer might be someone that fixes your boiler. It’s just that we smudge the two together. I think the Royal Academy [of Engineering] is doing a good job. I think as a country we celebrate engineering wins a lot.
I think the word “engineer” is slowly changing its meaning. I’m really proud that we’re proud about engineering as a profession. If I could, in this article, get 10 more people to be an engineer, happy days.
Looking back at your six-year-old self, is this the future you imagined?
You’re the first person that's ever asked me that question. Yes. I think I’ve always believed that the machines are coming and it’s going to be great. These machines are going to drive themselves because that’s next, and we can’t be condemned to a future sitting behind steering wheels. That’s not our future.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.