Autonomous vehicles and self-driving cars on The Future, This Week

Disclaimer We'd like to advise that the following program may contain real news, occasional philosophy and ideas that may offend some listeners.

Kai So Sandra, what should we talk about this week?

Sandra Oh, there's been quite a few big stories, haven't there?

Kai Ah, there's Google. The US Justice Department and the 11 state attorneys are filing a lawsuit against Google for antitrust. So that's a curious one. I mean, we've been talking about tech lash for quite a while, we've been talking about problems with Facebook, there's been allegations that Apple has built a monopoly around its App Store, there have been talks around Google and advertising, but that the Justice Department should file a lawsuit for Google's core business, Search, alleging that they have illegally built a monopoly, mind you, they organise about 80% of all web searches, and the valuation of the company has eclipsed a trillion dollars, but they alleged that they have struck deals with other companies such as Apple to basically unfairly build this monopoly.

Sandra So this is probably going to be the most significant monopoly related lawsuits in the last decade. But it's probably going to be a very long road ahead, both in terms of the lawsuit and the responses, and the timeline that this will play out over. So maybe we should come back to this when there's a bit more?

Kai Given that a similar lawsuit against Microsoft took about a decade to resolve, I think we have some time to come back to this one.

Sandra I was going to bring up some stories from China. There's been quite a few over the last week. One of them has been that the third quarter GDP numbers point to the fact that China's economy likely grew by 5.2% in July to September from last year, which shows a fairly steep recovery and the fact that consumers are back. But we have covered that quite extensively in a 'China is moving on' special in Sydney Business Insights, hearing from a businessman, an academic, a diplomat and economist telling us the story of China's economic recovery.

Kai And while I was not part of that project, I listened to the episode, and I found it really engaging, really interesting. And it is significant because China's economy really stands out worldwide. If you look at the charts that show the decline in all the big Western economies, China, with a strong recovery and an actual growth, really stands out. And so looking into that is really good. And I think we might not discuss it today, but listeners should just go and look for the episode. And there's also a lot of resources that you guys have put up on the SBI website. I have another interesting story. One that we missed, in fact, in a sense, Quibi, which wanted to be but now is not, the latest in streaming services. So Quibi launched about six months ago too much fanfare, with two very prominent co-founders, ex-Disney studio CEO Jeffrey Katzenberg and Meg Whitman, who was the CEO of Hewlett Packard of course. They wanted to revolutionise streaming with short form content, longer than TikTok, but shorter than other streaming services for snackable content on-the-go, paid for, high profile, very good quality, launched with $2 billion backing. Never really got to sizable subscriber numbers, a bit over a million during the 90 day free trial, which compares to about 50 million on Disney plus when they launched and now they're shutting down the same week that I got a notification that it's now available on Apple TV. So bye bye, Quibi.

Sandra That's a short story. Well, I guess we just did that one. I'm sticking a bit more to my China news because there's been an announcement that Alibaba is taking over China's top Hypermart chain for about $3.6 billion. So Alibaba will take control of China's largest chain of Hypermart. Basically to try to fend off rivals like JD.com. Alibaba has been trying to grapple with what has been really intense competition from the likes of JD.com, from giant food delivery companies, but also from startups backed by Tencent like Missfresh, all going after the groceries and fresh produce market, which, amplified by the COVID-19 crisis, is set to grow two and a half times over the next few years.

Kai So that sounds a lot like Amazon buying Whole Foods.

Sandra In a way it is and it isn't, because Alibaba is a bit late to the game, they're actually doing this move to respond to companies like JD.com, which already have a significant stake in food delivery and so does Tencent. So they're doing this to make sure that they stay in the game. And it's also quite noticeable that in China, the ecosystem has been very difficult to crack for any Western companies. While food delivery in China is very big business, much bigger business than in the US or here in Australia. One of France's largest supermarket chains, Auchan, has just got back from China. Carrefour sold its 80% stake in China, the German wholesaler Metro, also pulled out of China. So it's been a very, very difficult market to crack for the West.

Kai But it also shows the maturing of e-commerce, the fact that this is no longer a standalone thing, supply chains become integrated, the boundaries between online and offline, fast disappearing,

Sandra And in some ways, much faster than somewhere like here in Australia.

Kai So I'm sure we'll return to China and commerce. We did a great episode on Xiaohongshu a little while back with Kishi Pan. And there's much to learn from China. But..

Sandra But still probably not the topic of today.

Kai Probably not. I found another really interesting curious one that we don't have enough information about to do an episode on but I thought I'd mention it. Did you see the news about a drug selling robot service on Telegram?

Sandra Are you suggesting more research is needed into the Televend service which automates the entire illegal drug selling business on Internet, they're open 24/7, they probably deliver too?

Kai Well, I want to put it on the record here, since we're a university-hosted podcast, no I am not intending to do research into this. But it is really interesting that those services should spring up. This was reported in VICE News. Apparently, if you want to sell and buy illegal drugs on the internet, it's really cumbersome business, you know, to be engaged on the dark web, you have to have elaborate VPNs, you have to engage with very slow websites. According to the article it's really complicated business. And so here's this new service on the encrypted Russian messaging app Telegram that promises to bring together sellers of vendors and buyers of illicit drugs on a platform that is completely automated. So basically a robot eBay for illicit drugs. So a really interesting service.

Sandra And we did have a whole Corona Business Insights episode on how COVID-19 and the pandemic has transformed organised crime. Here's another instance of it, very hard to sell drugs on the streets during lockdown, but now it's also automated open 247, and I think this is where we should stop with this topic.

Kai I agree. Probably better.

Sandra If you get the munchies after the last story, one more out of Asia, Impossible Foods, which we've covered extensively here and its rival Beyond Meat. Both companies making plant-based meat alternatives have now launched in Asia and are planning to move into China. Impossible Foods open stores in Hong Kong and Singapore, before entering the Chinese market. And Beyond Meat has already signed a deal to start production in Shanghai. So both companies are going after the rapidly growing Chinese market. And we'll put in the shownotes our conversations around alternatives to meat.

Kai But the topic that we really have to look into this week is autonomous vehicles. There have been a tonne of different stories, variously announcing that self-driving cars have finally arrived, or pronouncing that they are nowhere to be seen and really hard to do, they won't be accepted by users, or that Tesla is releasing its self-driving functionality this week. So a smorgasbord of announcements that pointed us to look into this a bit more deeply.

Sandra So back on The Future, This Week are autonomous vehicles.

Kai Self-driving cars, let's do it.

Sandra This is The Future, This Week from Sydney Business Insights. I'm Sandra Peter.

Kai And I'm Kai Riemer. Every week we sit down to rethink and unlearn trends in technology and business.

Sandra We discuss the news of the week, question the obvious, explore the weird and the wonderful and things that change the world. Last week, there was a flurry of announcements from Alphabet's Waymo, to General Motor's Cruise, to Tesla to Baidu's Apollo, but there was really only one story we could do. And that was the one from Forbes...

Sandra "You May Not Have Heard, But The Autonomous Vehicle Has Now Arrived".

Sandra But it really hasn't.

Kai It kinda has...

Sandra But it really hasn't.

Kai No, it hasn't.

Sandra So we both have the look at the story and started having flashbacks because for the past few years, these autonomous vehicles kept arriving. And somehow they're still not here.

Kai So for a few years now, self-driving cars were being announced, they were arriving, Tesla's self-driving functionality kept arriving. Uber was about to have a breakthrough in self-driving car technology and a lot of the timelines end around 2020/2021. So it is an interesting new story to see that supposedly self-driving vehicles have now arrived.

Sandra And whilst in our experience over the past few years, the conversation has gone from a lot of hype around autonomous vehicles about three years ago, to more conversation around electric vehicles recently. The narrative around the autonomous vehicles being 'just around the corner' is still very much present. But very, very little has actually changed in how we drive around, the types of vehicles that most of us see on the road, and types of vehicles that are being produced by most companies.

Kai And if you don't live in Palo Alto or a few other places, in the US, you won't actually see self-driving cars as part of normal traffic.

Sandra So we should first have a look at the Waymo and Cruise and Baidu, Tesla announcements this week and see if the autonomous vehicles are really here. But then more interestingly, let's have a look forward at what really can be done in the future of autonomous vehicles, and what is it likely to take to really have them on the roads.

Kai So let's have a look at this particular story first, so Waymo, which is of course, Google's self-driving car programme, has just announced the launch of a self-driving taxi service in three smallish cities in the Phoenix, Arizona area, Chandler, Tempe, and Mesa. All towns of a population between 200 and 500,000 people. So there is now a self-driving taxi service where taxis with no driver will carry people from A to B. So self-driving cars have kind of arrived.

Sandra Well, they've only kind of arrived for a number of reasons. First is that the service is only available in a very limited area. And that's because of the way Waymo has approached the technology problem this time around. And that is by building a very detailed three dimensional map of the environment, so that the car can better deal with unexpected obstacles that might come up on these predetermined routes.

Kai So what we're saying is that self-driving cars haven't really arrived in the world, they have arrived in three places that way more has basically turned into elaborate 3D computer game environments, where they have now a complete 3D software representation of those cities where those taxi services can go, but they cannot really venture outside of that. So they've taken off the table a whole lot of complexity that self-driving cars would normally have to deal with, by having this complete digital representation which comes at a huge cost of creating in the first place. And mind you, we still don't know whether the self-driving cars perform really well in bad weather circumstances. Arizona as a place was picked because weather conditions are usually fine. These are also cities that don't have the same kind of complexity and traffic or pedestrians in the streets like you would find on the streets of Sydney, for example. So we're talking about a launch of a service under very favourable conditions, with the extra effort of creating the 3D representation.

Sandra And the announcement from General Motors with Waymo's rival Cruise has really been a similar type of story. And even though this one was launched on the streets of San Francisco where pedestrians abound, traffic conditions are a lot more difficult, and so is the weather. It came with a small disclaimer that said that the vehicles are only allowed to travel to speeds under 30 miles an hour and are prohibited from driving in heavy fog or rain, which is probably really not good news for people in San Francisco.

Kai No, and there's only five cars so far that they are testing, and as you said they're only in certain neighbourhoods with strict speed restrictions. So self-driving cars are arriving again, under very favourable conditions in very restricted areas.

Sandra And before we look at what this all means, we must also bring up the Tesla story because, as with all the other announcements, Elon Musk has been announcing for the past three or four years that Tesla cars have all the hardware needed for full self-driving and people have been asking where this is.

Kai So Elon Musk has just announced that it is coming this week, but only to some drivers who have a history of very safe driving. And he said that Tesla is approaching the software update very cautiously. The car will also be in a mode where it drives very slowly and safely because according to Elon, "the world is a complex and messy place". Well, you don't say.

Sandra So again, it remains to be seen whether self-driving cars have really arrived.

Kai Also, it needs to be clarified that drivers still have to be in full alert because the self-driving feature, much like Tesla's autopilot, can disengage at any moment when something happens. So drivers still have to be ready to take over.

Sandra But before we try to look forward to what next for autonomous vehicles, it's worth reminding people that it's not just American car companies having a go at this. Just recently, China's Baidu announced its own AI-powered car platform called Apollo.

Kai And the announcement came in September at Baidu World 2020. This is the sixth version of Apollo. And Apollo is significant because it's not a particular car, although Apollo did team up with the Chinese state-owned carmaker FAW Group to create a pilot car. It's more like a platform that Baidu offers to about 210 partners globally.

Sandra And much like the other projects that we've discussed, a robotaxi service powered by this platform has been tested in China in multiple cities, including Beijing. So in this flurry of announcements, we thought it would be important to try to look forward at what will it take to have autonomous vehicles on our roads in our everyday lives? As we've said, over the last five years, nothing really much has changed in the way most of us experience driving.

Kai And so what we can see across these stories is that there's a number of different ways in which self-driving cars can be approached from a technology angle. But that's not the only thing that needs to be solved. So apart from the technology, there's questions about the usability, how do these things fit into our daily lives, the economics of self-driving cars, issues of finance, insurance, and the legal issues of what happens when there's an accident, and also matters of infrastructure and the built environment.

Sandra So the first thing it will take for self-driving cars to get here will be to solve the technology problem. And whilst with many of these examples, we are very, very close, we're 99% of the way there, we're still not quite there. But the story is do tell us that companies have taken different approaches to solving the technology problem.

Kai And so from the announcements this week, we can glean at least four different types of how to approach the technology problem. So the first one would be what most companies are trying to do, which is largely based on equipping cars with a whole range of different sensors and cameras, LiDAR, radar, and to then feed all the data into the kind of deep learning algorithms that we have covered on the podcast previously. And while that seems to be on paper a promising way forward, the technology is still struggling to, you know, as you said, go to that hundred percent, to deal with the so-called 'edge cases'.

Sandra And that can include challenges to the physical technology such as low light conditions, or fog or rain.

Kai Yeah, well, who wants to drive in the rain, right? You know, Germans maybe.

Sandra Or indeed cases where humans are involved, say someone's at the pedestrian crossing, they have no intention to cross the road, but they are standing there on the phone, cars are still unable to detect whether or not there is an intent to cross in the person standing at the pedestrian crossing.

Kai Because they're not capable of you know, making those judgments like a human driver would make. So there's a lot of these situations where the self-driving technology still gets caught out. And that kind of technology is what Cruise is using, which is now testing in San Francisco. But this is also what Uber is trying to do. And there was an article recently in Gizmodo, which was titled "Uber Is Here To Remind You That Autonomous Vehicles Aren’t Even Close". And it reports about the struggles that Uber has with its cars, frequent disengagement of the technology, to the point where Uber's cars can only drive a couple of kilometres without a problem occurring. So a lot of issues with actually getting these algorithms ready for primetime in a range of different situations.

Sandra And that's pretty much the problem with the second technological solution, which is micro-mapping the environment, rather than reading the environment through all the sensors, it's an approach that micro-maps the environment, and then just sets the car loose in what is practically a simulation of the built environment.

Kai And that's what Waymo is doing in those three cities in the Phoenix area. And Google, of course, is very experienced in mapping environments, we have Street View, we have now 3D representations of some cities. But to bring it down to the level that is useful for these cars means that a lot of effort, a lot of work has to be done to create these elaborate 3D maps, that then reduces the complexity for the algorithms, because they then only have to deal with the moving objects rather than the environment as a whole. The downside of course, is that...

Sandra Plant a new tree?

Kai Yeah, exactly. You plant a new tree, you have to basically remap all the time and keep those maps up to date. So it's not something that scales easily to create cars that can drive everywhere. And it's also not clear whether a car that can drive in the one city, with its algorithm trained on the city of Chandler, for example, would also then drive in the neighbouring city of Tempe. So while we have those services, they are not the kind of self-driving car that you could put in the hands of everyday consumers to just drive everywhere.

Sandra But there's two other approaches to trying to solve the technology side of the problem. And, Baidu offered one of the solutions, which was to develop the platform, but then to give the platform to as many carmakers as possible to ensure that enough data is gathered to optimise the algorithm. So Baidu, as we've mentioned, partnered with the Chinese state-owned carmaker FAW Group to do this and to ensure that the platform is in as many vehicles as possible.

Kai And it refers to its platform as an "experienced AI driver", so they basically want to build the brains around which manufacturers can then create self-driving cars by connecting sensors. And the flip side, of course, is that Baidu gains a lot more experience gathering the data from all these partners, without having to build cars themselves.

Sandra But to round this up, there's also one company that has a very different approach to solving the technology problem, and that's been Tesla.

Kai So most self-driving cars can easily be spotted in the streets, because they have these bulky domes on the roof that hosts the LiDAR sensor technology. But we all know that Tesla cars look just like normal cars. So Tesla wants to solve the problem without the sensors, substituting for this with orders of magnitude more learning data that feeds into its algorithms.

Sandra Because Tesla has built the entire car basically around a computer means that its entire fleet is collecting data on everything that the car does and sees all of the time.

Kai So Tesla has almost a million cars on the road. And it collects all the data which it can feed into its self-learning algorithm, improving its autopilot functionality, gradually taking over more and more of the functionality of driving, to the point where it hopes to actually have fully self-driving functionality, well right about now or very soon. And estimates are that the amount of data that Tesla collects is at least 1000 times more than what a company like Waymo or Cruise could collect with its smaller fleet of test vehicles. And so, if it is indeed a matter of collecting more data to train the self-learning algorithms, Tesla should have a strong advantage.

Sandra But before we turn to what will it take for self-driving cars to get here, in terms of the economics, the business model, the legal structures and the infrastructure that has to support these cars, it's worth bringing up one more model for autonomous vehicles that actually comes from heavy equipment makers. And that is Caterpillar who makes trucks and bulldozers has taken a very different approach to autonomous vehicles and has been quite successful in responding to the pandemic.

Kai So Caterpillar last week announced that it is growing significantly its services business, that revolves around its remote control and autonomous operations technology that they can not only integrate into new machinery that someone would buy, but they can actually retrofit into existing fleets of bulldozers and diggers and it's large trucks that are used in mining for example.

Sandra So this was really accelerated tremendously by Caterpillars response to operations being shut down in many parts of the world because of lockdowns, due to the pandemic. Caterpillar attempted to help its clients by providing autonomous driving technology that can just be bolted on to already existing machines. So rather than trying to build the machines that will do this, because of the restricted environments that heavy machinery is deployed in, Caterpillar can afford to just build the technology separately, and then deploy it afterwards.

Kai And while the field of autonomous vehicles is of course, much bigger, it includes drones and other delivery vehicles and planes and so forth, we'll set that aside and might come back to this in a future episode, we want to look at the economics of bringing onto the roads self-driving cars.

Sandra The economics of self-driving cars have always been an obstacle. And initially, the obstacles were related to solving the technology problem, in that vehicles that are loaded with sensors are fairly expensive, the sensors are produced in very few places around the world. LiDAR has become a lot cheaper over the last few years, but it's still a fairly expensive game.

Kai Except for your new iPhone, which will have LiDAR technology in your camera.

Sandra You mean your new iPhone?

Kai Exactly.

Sandra And cost really isn't the problem just due to sensors. But just consider that micro-mapping an area such as Waymo did for three cities took them almost three years, if you try to deploy something like that at the national level, the costs involved and the time involved would not make a lot of economic sense.

Kai At least not at this point in time. But which also means that those economic issues will slow down the arrival of self-driving cars considerably.

Sandra As will the fact that a lot of the self-driving cars at the moment are still monitored, many of them still have a person in the car to be able to drive, very few are allowed to be totally driverless, many of them are supervised from control centres and so on.

Kai Which to me, again, raises the question, what are we actually optimising for? What is actually the end goal to achieve? Because if it is, indeed, safety, as is often touted, you know, the argument being that human drivers get distracted, they get tired, but a robot, a computer would not, we would say and we've said this on the podcast before that we have already solved this problem. If we look at a human driver paired up with the safety technologies that go into a modern car, such as a Tesla or a Volvo, we can already achieve very safe assisted driving that is ready to be rolled out across new cars at a much lower cost than what it would take to get to fully self-driving cars.

Sandra But on the other hand, if you're trying to solve for replacing the driver all together, so that the driver can now relax and read in the backseat. Or if you're trying to solve, like Uber, for having an entire fleet of share cars that have no human drivers, then we're still miles away from solving this.

Kai And that is bad news for Uber, of course, which many say will only become a profitable, large scale service, if it indeed manages to, you know, get rid of all its drivers and become fully autonomous. Not only is Uber struggling, but the technology is still years or decades away from a mass rollout in any types of conditions in which you know, Ubers will have to operate.

Sandra And then there's of course, questions around the fleet of cars. Because while a company like Tesla sells directly to consumers, and people still own Tesla cars, companies like Google and the Waymo service is not really trying to build something that's offered to consumers, but rather to build fleets of taxis, much like we've seen with the implementations of the Baidu platforms as well. Which means that you still need some structures, in the case of Waymo it's Avis, that manages that fleet of taxis. You still need infrastructure that will house these cars, that will service these cars, that will insure the cars and so on.

Kai And that is not a trivial question. Because both Google and Uber are technology companies, they're largely software companies. So it would mean a reversal in terms of their economics, in terms of their capital spend. If Uber, instead of putting the cost of owning vehicles onto its fleet of drivers, would have to take on the capital of actually owning all these cars. The same for Google, which you said is partnering with Avis. Building those autonomous taxi fleets is, from a financing point of view, from a capital point of view, a very different business that isn't necessarily at the core of what Silicon Valley tries to do.

Sandra Of course, to get autonomous vehicles on the roads it will take not just solving the technical problems and figuring out the economics of it or the financials. It will also take figuring out issues of accountability, legislation and so on and so forth. A Mercedes Benz executive was asked a few years ago about their future autonomous cars and whether or not they will save the cars, drivers and its passengers, and sacrifice the life of pedestrians, or the other way around? And famously, he answered that, "if you know you can save at least one person, at least save that one", foreshadowing that the car company would always privilege the lives of the drivers or the owners of Mercedes cars, to those of pedestrians.

Kai But the problem is a little bit more complex than that, of course, and you mentioned accountability. So if I was deciding whether I wanted to buy a self-driving car, I would want to know what happens if this car actually is involved in an accident and kills a pedestrian, who is liable? Would I be liable as the owner, who now has no agency about what this car that I'm sitting in is doing because the robot is making the decisions? Or is indeed the car company liable, killing this pedestrian? And before those questions aren't addressed, why would I actually purchase a car like that?

Sandra And maybe this is a good place to just remind people that safety is still very much a concern for autonomous vehicle makers, even though companies like Waymo claim that trials that involve driving 20 million miles without killing anyone are a good track record.

Kai So what we need to know is the death rates in the US are actually one per 100 million miles. So the pronouncements we can make from 20 million miles are rather limited. And also those miles could be driven under quite favourable conditions, in good weather, and on streets that are not necessarily all that challenging. So there's still a lot of questions around this that remain unanswered.

Sandra And these questions are still very much discussed. While the German Ethics Commission came out with a report three years ago, proposing a certain set of rules that would provide guidance to self-driving cars, we know from a long running experiment from MIT called the Moral Machine, that it's really difficult to come up with a set of rules by which cars operate, and then translate that into legal frameworks into answering accountability questions, and so on and so forth. The research project that MIT has been running for a few years now offers a set of dilemmas that people can answer on the internet, whether or not they would choose to save the old lady or the three children, or the dog versus the three, and so on. And up until now, after counting about 40 million decisions from people in 233 countries, they found out that other than people opting to save more lives when possible, and people rather than animals...

Kai The results otherwise aren't really as clear cut because different cultures value different things.

Sandra Depending on whether people were coming from Confucian countries or were coming from countries where Christianity has historically been the dominant religion, or certain groups from Central and South America, they showed very different preferences, for instance, for sacrificing older lives to save younger ones, or the other way around. So there are no easy decisions that car companies can make at this point, if they want to build morality into their machines.

Kai And MIT is looking into other aspects around trust and also usability of these cars. There is research now around how these cars interact with their passengers. For example, would they have to have natural language interfaces? How do you build the kind of trusting relationships with these alien machines that people would have to get into? How do you make these technologies palatable to end consumers? So there's a lot of unresolved questions because polls have shown that while people are generally interested in these technologies, when asked about would they get into a vehicle right now? About 86% said that that was out of the question. So there's lots to be done to build trust into autonomous vehicles. And we'll put a recent episode of an MIT podcast in the shownotes that looks into this aspect a bit deeper.

Sandra And the final point we want to bring up today is that of infrastructure. Once all these questions are answered, there is still the question of the infrastructure that these cars will need to use. And we're not just talking here about high speed 5G networks that the cars can connect to, but also the quality of the roads, the signage on the roads, potholes and so on, and indeed, the suitability of cities to accommodate these. Everything from parking spaces for fleets of autonomous taxis, to potentially roads that exclude pedestrians and favour autonomous vehicles.

Kai So much like bike lanes or bus lanes, there could be changes to the built environment to accommodate yet another special type of vehicle on our roads.

Sandra And this is still a decision that we will need to make. Historically we've always tried to adapt the roads to the cars more than the pedestrians. Jaywalking laws were essentially invented to transform cities into safer places for cars. But we've seen a recent move, especially with COVID-19, towards making cities more pedestrian-friendly and rather excluding cars from city centres, which would make it even more difficult for these autonomous vehicles to find their place in our lives.

Kai And this is where we want to leave it saying that COVID-19 has thrown up a whole range of new challenges and opportunities for autonomous vehicles, which we might come back to in a future episode of Corona Business Insights. See you next week.

Sandra On The Future, This Week. If you enjoyed this podcast leave us a rating.

Kai Thanks for listening.

Sandra Thanks for listening.

Megan Sandra Peter is the Director of Sydney Business Insights. Kai Riemer is Professor of Information Technology and Organisation here at The University of Sydney Business School.

Sandra With us every week is our sound editor Megan Wedge,

Kai And our theme music was played live on the set of garden hoses by Linsey Pollak.

Sandra You can subscribe to The Future, This Week wherever you get your podcasts.

Kai If you have any weird and wonderful topics for us, send them to sbi@sydney.edu.au.

Kai You may not have heard but the autonomous vehicle has now arrived. But the best is yet....

Sandra No!

Kai No?

Sandra No.