Full Episode TranscriptThe Healthcare Crisis We’re Facing

Lisa: Hi everyone, welcome into Pushing The Limits. Today I have my friend Cern Basher with me again. So Cern, welcome back to the show. So cool to see you again.

Cern: Hi Lisa. It’s great to be here. Thanks for having me.

Lisa: We’re making this a regular thing. It’s really cool. I’m just so stoked to have you on and we’ve been doing some interesting stuff together.

Recently I just did a video on robotics and healthcare. I’m very passionate about making something like this happen down in New Zealand especially. And we’ve been having some conversations around that. A lot of people look at me absolutely like I’m crazy when I’m talking about Optimus Robot being in healthcare or robots in general being in healthcare.

But as someone who’s lived the caregiver experience for over a decade now, this is something I’ve thought really deeply about and something that I wish I had. It’s not going to be coming in the next few minutes, but things are really rapidly changing now. You’re an expert in Tesla and XAI, SpaceX, in robotics and Optimus and what’s going on there.

So let me paint a picture for you before I let you dive in. The WHO says that we’re going to be short somewhere between 10 and 15 million healthcare workers in the next four to five years. This is an absolute disaster coming at us. We have an aging population worldwide.

New Zealand is the largest importer in the OECD countries for health workers. Every second nurse and doctor at our local hospital is from overseas, which is fantastic that they come here, but we’re all after these nurses and doctors and caregivers and there’s a shrinking pool of them. Every country is facing this demographic inversion where we have a lot more older people, and this is going to collapse at some point.

We already have 8,000 vacancies in our healthcare system that we cannot fill just in the hospitals. That’s not counting the aged care facilities which are also running at about a 15% vacancy rate. I personally have caregivers that I have to get in. It’s an impossible task to get a really good caregiver and to be able to afford that as a family.

Already one in four people is looking after loved ones — usually elderly loved ones, sometimes with children as well, disabled or some such thing. This is a huge portion of our society and we are heading for an absolute unmitigated disaster if we don’t do something. So that sets the stage, Cern. What’s your take on how can robots in general, and perhaps Optimus specifically with your Tesla expertise, help in this case?

The Washing Machine Analogy

Cern: Yeah, the situation is pretty bleak if we don’t do something. And the good news is we have a wonderful opportunity to push forward some new technologies that will allow us to not only deal with these problems but actually make healthcare a really pretty amazing experience. And in addition to that, turn the tide on rising costs in healthcare.

If you think about New Zealand importing all these doctors and physicians, the cost of that must be astronomical. There’s probably a decent amount of turnover where those people decide that they want to go back home after a period of time. So the cost must be increasing. The quality of care is probably going down.

Wouldn’t it be nice to be able to deploy a technology that turns the tide on that? Improves healthcare, drives the cost down, makes everybody’s life that much easier. And this is what we’re looking at with robotics.

The way I like to think about it — this is kind of a silly example, but I think very instructive — how many people wash their own clothes today? Don’t use a washing machine? How many people don’t use a dishwashing machine? All these mechanical tools that we’ve had before.

The idea of having to wash your own clothes seems probably pretty crazy to most people. And I think in the future, the idea of not having robots do a lot of very boring, repetitive, in some cases dangerous tasks that humans do today — I think the idea of that is going to seem kind of strange in the not too distant future.

Self-Driving Cars & The Technology Pipeline

Lisa: And so we’re on the cusp right now of the early days of that transition towards having intelligent robots autonomously do a lot of things that humans really shouldn’t be doing. That’s starting with driving cars.

Cern: One of the most challenging problems. I think many people thought that we would never see vehicles drive themselves. That the world out there is too complicated for vehicles to navigate the million edge cases that you could encounter. And yet here we are today. Tesla and others are proving that this technology works and it’s getting better all the time. It’s not perfect — no new technology is ever perfect. But we are on the cusp of being able to have autonomous vehicles and that technology is going to flow right through to robotics and we will be able to bring this technology into the healthcare field as well.

Lisa: Full self-driving is starting to roll out in America. There’s a couple of hundred robo taxis out there at the moment and that’s going to scale. In China there are trucks running around on the highways without drivers, without even steering wheels, which for us here in little old New Zealand might sound absolutely crazy, but that sort of thing is happening already.

The robo taxis, the full self-driving — this is like an Optimus but on wheels, isn’t it? It’s the same technology behind it, the same brain that goes into the cars that will go into the robots.

Cern: It’s the same sort of technology that’s involved here.

Timeline: When Will Robots Be In Healthcare?

Lisa: I think one of the first things we have to address is where is the technology now? How far out is this? Are we talking four years, five years? Ten years? Twenty years? What is the time range that you think is realistic to be actually deploying these into healthcare settings on more than a pilot scale?

And then let’s discuss some of the concerns that people have — the fall risks. These are big heavy machines. They’re not just stationary like our washing machine stuck to the floor. It’s not going to get up and walk down the hallway and cause havoc. We’re used to machines being stationary, locked to the ground. We’re even used to seeing in car factories those arms that move around, but they’re still stationary. They’re not fully autonomous humanoid robots, which is what we’re talking about here.

Cern: I think it’s important to first of all understand that we are still in development largely with this technology. But that said, I think it’s important not to make the mistake of saying, “We’re just going to wait until this technology is perfect before we deploy.”

The reality is that the humanoid robots, even as they are today, are useful. Even if your humanoid robot could only move something from point A to point B, that’s a very useful endeavour. Particularly if we focus on the healthcare environment and think about a hospital setting — all the supplies that must be moved within a hospital constantly. Think about freeing up humans from that task and how valuable that might be.

Sometimes it takes a long time to get the human to deliver the item because that group is backed up, whether it’s the pharmacy, whether it’s the food service. Having robots that just do that constantly — that’s very doable today.

Now, those robots today might be on wheels and you can mitigate the fall risk with that. That technology is already out there. The idea with the humanoid robot is now you have this all-purpose robot that functions in the environment that’s been truly built for humans. Things with wheels can’t go upstairs. Humanoid robots are built more for the human environment.

Lisa: In a hospital, think about all the different cables and things — the humanoid robot should be able to navigate that environment much better than other types of robots can.

Cern: I think it’s important to realise that long term, this is an opportunity, and in the short term, the sooner that we begin to integrate this into organisations, the better. Not only because you get some initial benefits, but also then you play a role in determining what the future looks like.

Lisa: And this is the opportunity for New Zealand, isn’t it?

Safety, Risks & Managing The Transition

Cern: Definitely there are risks. Any new technology, you have to be very careful. With the self-driving cars, for example, the technology is largely supervised, although I think that’s going to be changing relatively soon. We’re going to see Tesla roll out their robo taxi network with more and more unsupervised vehicles with no one inside the car.

On the humanoid side, we’re going to see more and more of that as well. But as these robots develop, you make sure that they’re operating in a safe environment initially. And then as they prove themselves, you release some of the shackles on the bots and you can have them integrate more with people.

One of the things that a lot of the robot makers have done is make the robots relatively small and relatively lightweight. So even if they do fall, it’s not like a 200 pound machine falling on you. It might be something like 120 or 140 pounds. Which is still a lot — you don’t want that falling on a child or even an adult. But there’s things they can do to mitigate a lot of the risks.

Tesla Optimus: The Hands Are The Game-Changer

Lisa: Where is Tesla at with Optimus specifically? We see all over the world robotics companies at different levels of development. From what I’ve seen with Optimus, it’s the hand that really stands out as being the key difference. Our hands are the most dextrous part of our body. That’s what we do everything with. I think Tesla has spent a lot of time on developing the hands specifically. Is that correct?

A lot of the other robots we see doing kickboxing — which is probably not a great thing to be showing because it gives the wrong message — we see them dancing, doing flips, doing crazy types of things. But a lot of them don’t have the hands that are necessary to do the fine motor control, which is really where we unlock a lot of the power of Optimus. Is that right?

Cern: That’s right. Hopefully kickboxing in the hospital environment is not something that’s necessary! Although it might be entertaining.

For Tesla and a lot of the other robotics makers, they’re developing this technology and there’s really sort of three general components.

There’s the robot physically — making a robot body designed in such a way that it can function and has all these different degrees of freedom so that it can operate in the same space that humans do. We’ve made some great strides across the robotics community and particularly with hands. The hands are looking more and more human-like every day.

The second is the AI part that operates this physical body. If you think about a child learning about gravity — the ball rolls off the table, the child learns that it falls to the ground. Humans learn that over a period of time. We have an intuitive understanding of physics. We know that if a ball rolls behind the wall, the ball hasn’t disappeared. The robot also has to learn physics and that’s a big part of the AI development.

And then the third thing is the robot’s brain — you can think of it like putting a large language model inside a robot. The cool thing about these robots is they’ll be able to speak to you in any language.

Lisa: Amazing.

Cern: In a hospital environment where you’ve got people coming to a country like New Zealand, it would be cool to have a robot that can speak any language and you don’t have to find an interpreter and wait 30 or 45 minutes. There’s one application right there.

There are other robotics companies focusing on just the hand. They’re making a bet that the hand is the key component — that the robot companies will do everything else and may purchase the hand from them. It remains to be seen how this industry shapes out.

And then the other great challenge — and this is where Tesla will excel — is in the ability to make millions, tens of millions and eventually hundreds of millions of them. To be able to scale the technology.

Tesla’s Manufacturing Scale: Fremont & Texas

Lisa: We talked on our last podcast about the Fremont factory where they’ve already shut down a couple of the lines to make Optimus robots, and they’ve already broken ground in Texas.

Cern: That’s right. They shut down one of their auto lines in the Fremont factory in California to make a million robot capacity line, and they’ve begun to build a factory in Texas that will have the capacity of about 10 million robots per year.

Lisa: 10 million! That’s going to take a few years to build obviously. But the one in Fremont is a million robots a year type of setup. And the price of these things — I’ve heard you talk around $75,000 US. I’ve heard other things where it might be getting down to $20,000 to $30,000 US eventually. That’s really a key factor.

If you say in an ideal world about $20,000 to $30,000 — a nurse in New Zealand might cost $90,000 to $100,000 a year to employ. If you’ve got a robot that can do this 24 hours a day — it needs charging and some maintenance but generally speaking it can work around the clock — that’s going to be one hell of a savings for the healthcare system.

Without displacing workers. I don’t see displacement in healthcare. I do see displacement in other industries, but in healthcare we just don’t have anywhere near enough. These will be augmenting the current staff and doing all the drudgery tasks.

The Moxy robot is currently in 25 hospitals in the US. It’s already done a million deliveries and saved over 600,000 hours of staff time doing the drudgery — picking up the soiled linens, delivering medications, delivering something from there to there. Yes, it’s a one-armed robot on wheels, but this proves the concept. That’s already saving those hospitals a lot of time and money and it bodes well for the future.

My dream is to see a pilot happen here in New Zealand. In a place where the staff are trained, a small number, we’ve done the data — in a small setting where it’s completely controlled, to run this whole thing to get the data.

Fleet Learning: Why 40 Robots Can Learn In 6 Months What Takes A Human 25 Years

Lisa: Optimus or any of the robotics companies needs the actual real world data, not just the simulations. Can you explain fleet learning and how this data collection that we could do as a first phase would play out in reality?

Cern: If you think about a human from a baby up to age 25, where you might have peaked out in terms of your learning journey. You spend the first 25 years essentially learning about the world and how it works. We learn sequentially. We don’t teach physics to three-year-olds. But they’re learning physics on their own. It takes a long time.

And by the way, your learning journey is different than mine. I don’t really benefit directly from your learning journey unless you come and help me understand. With robots, you can have one robot learning how to do something and another robot learning how to do something else, and at the end of the day, they both know how to do those things. The learning is not contained to that specific robot. It’s spread across the entire fleet.

Let me put this in context. If a robot learned at the same rate that a human does — if you had a baby and a baby robot, they might both take 25 years to get to the same level. We don’t have time for that. But if a robot can learn in parallel and spread it across the fleet, you would need about 40 robots. Those 40 robots could learn the same amount that one human could learn in 25 years — in six months.

Lisa: Wow. That fast. Just 40 robots.

Cern: It doesn’t take much. Small scale deployments over a six to twelve month period in specific areas, particularly healthcare, could be really powerful in helping these robots learn all kinds of different things.

The training methodologies — both the simulations — the way to think about simulations is it’s kind of like reading a textbook. You can learn a lot from reading a textbook. It’s helpful, but it’s no substitute for the real world experience.

Lisa: That’s a really good analogy, Cern. We all know that we read something and sort of get it, but we don’t actually understand until we’re actually doing the doing.

Cern: Same thing with the robot. The simulations are great, but ultimately you want to get them out in the real world and actually see how they operate and then you refine the training from there.

New Zealand’s Unique Opportunity

Cern: This is where it’s going to become absolutely critical in the healthcare field because there are very specific things that the robots need to learn. I think New Zealand is potentially in a very unique position to offer a lot to robot companies who are looking to develop robots to be capable in the healthcare field.

If you think about it — New Zealand has a national healthcare system, it’s an English-speaking country that’s kind of manageable in scale. Smallish. And you’ve also got a stated digital modernisation goal with some proactive people wanting to push forward.

Here is an opportunity to help New Zealand in a major way with a problem that every country in the world has — to create this learning environment for these robots that then can be copied and pasted around the world.

And by the way, that’s job opportunities for New Zealanders who are involved in this training. It’s not just solving the healthcare problem. It’s creating new opportunities. If you become the leaders in this field as it relates to robotics and healthcare — not to mention if you embrace these new technologies and develop good relationships with companies like Tesla — the opportunities that could come off the back of that are just absolutely enormous.

Overcoming The Terminator Fear

Lisa: You get a lot of people that are frightened of this technology. All we’ve ever seen in our movies is the Terminator. That’s what comes to mind for people when they think of these machines. And then we’ve got this whole AI revolution happening and we’re all going through this turmoil and massive change.

These are complex conversations that legitimately have to be addressed. When you have Einsteins basically with these brains inside machines, there’s the cyber security risk, the privacy risks, they’ve got cameras. We have to work out ways to do this in a safe manner.

But that’s where we need a seat at the table if we want to be influential as a country in helping develop these regulations. Where do we want to be getting these robots from? Friendly nations or not so friendly nations? There’s a whole lot of conversations and laws that need to be enacted.

I liken it to how I approached AI. At the beginning I was like, I don’t want to have anything to do with this. This is going to take jobs. But the more I leaned into it, the more I realised this is coming regardless. It makes huge opportunities. It’s going to provide opportunities to those who lean into it.

The exponential power that I now have at my fingertips compared to even a few weeks ago is insane. And I’m only able to do it because the technology is coming into play. It’s still clunky and it still breaks, but it’s the first movers who will have the advantage.

Lessons From History: The Hammer, The Loom & Artificial Ice

Cern: Throughout human history, all humans were afraid of the new technology. Think about the weaving looms and the rise of the Luddites. I could imagine the first human when the hammer was invented saying, “I’m not going to use that tool. That looks dangerous. I can hurt my fingers.”

But throughout human history, the arrival of new tools has only been a positive for humanity. Yes, there’s risks. With a hammer, you can hurt yourself. But look what you can do with a hammer that you couldn’t do with anything else. This is no different in my opinion. The implications are far more far-reaching, but if we have these problems, we have to find a way to solve them with some new tools and approaches.

Here’s one quick story. When artificial ice was first invented, the people that sold natural ice labelled it as ungodlike ice. Artificial. Synthetic. They were very successful for a long time getting people worried about fake ice. The problem was there were bacteria in the lakes on the ice they were harvesting and people were getting sick. Eventually society did a 180 and realised natural ice wasn’t safe and artificial ice was much safer.

Part of the hesitation comes from entrenched interests that don’t want to change because they benefit from the current system. Part of it is that as humans, it’s hard to constantly disrupt yourself. That’s uncomfortable.

But if change is coming — and it’s pretty clear that with robotics and AI and automation, it is — would you rather have a role in shaping that change, or would you rather that change be forced upon you and you don’t have any say at all?

Lisa: Being first is hard. Being first, you make mistakes. There’s lessons. You trip up. But being last is not the place to be.

The Caregiver’s Perspective

Lisa: I don’t want to see old people left alone to die. I don’t want to see a health system that cannot cope and provide the care that we owe to our older citizens. And I see that happening right now. I’ve experienced it and lived it.

I’ve been lucky enough that I’ve been able to advocate for my mum. Not everyone has an advocate. There are a lot of people going under the bus right now. And this is a way for us to change that for the millions of us heading towards being older. This could be us in 10 years. All it takes is one event and you need support. This could be any one of us any time, and we need to move now.

Privacy, Companionship & Early Detection

Cern: We worry about older people at home alone who don’t have someone to help them. We worry about people in care facilities not getting the right attention or being abused. Cameras go a long way in making sure they’re safe. You’ve got to thread the needle between that and privacy concerns.

Having a humanoid robot in the home that can provide assistance — not the least of which something they can talk to. You’re never alone. Ever. And these robots won’t only talk to you in any language, but eventually will probably talk to you in any voice. The robot could sound like you. So the robot with your mum might sound like you, and that might give her a lot of comfort.

That’s a level of care that we cannot provide today no matter how many people we throw at the healthcare problem.

Lisa: It’s very complex with a lot of ethics that need to be thought through. But there are so many lonely people in the world.

Cern: And if the robot is watching your behaviours and listening to you, there are certain things it may be able to alert the physicians to. “The speech is slurred” or “I noticed when they’re walking, there’s a problem I didn’t see last week.” Early detection could go a long way for everybody.

AI In Healthcare Right Now: Real-World Examples

Lisa: I use that already. I went to get an X-ray done for my mum. She’d had a fall and I wanted to check her hip. They did the X-ray. While they were doing it, I was standing next to the nurse who took a photo off the computer screen. I had to wait 10 days for the results from the radiographer.

Two seconds later, I walked out of that room, uploaded it into an AI, and had an answer. She was fine. I still want the official result, but I had a reasonable level of comfort that in the next 10 days I didn’t need to worry that the hip was broken.

Another example from my personal life — my husband had an eye issue. We took a close-up photo, uploaded it because we couldn’t get into an ophthalmologist. It told me exactly what was going on with that eye and what I needed to be concerned about.

This is not replacing doctors, but sometimes you cannot get access on a Sunday in rural New Zealand. If my mum has a problem, I’m going to AI as my first port of call while I’m waiting to get to the doctors. That’s what people are doing all around the world now.

Digital Optimus & Macrohard

Lisa: What is Digital Optimus? What is Macrohard? And what is Terafab? Tell us about the whole Digital Optimus story.

Cern: If you think about Optimus as being the robot that does physical work, Digital Optimus is the robot that can work with software. One’s physical labour, the other is the work you do coordinating things on a computer — taking data from somewhere and feeding it somewhere else, any sort of office work, anything a human does by looking at a computer.

Software like Digital Optimus, or another name Macrohard, can operate the computer just like you do.

Tesla is using their full self-driving software technology that understands visual inputs. The AI takes what it’s seeing on the road and makes a decision about where to take the car. It’s using that same fundamental technology of looking at a computer screen and everything on it and responding to it just like a human would.

Lisa: Parts of that exist already now with these agents. They’re doing parts of it, but it’s in a very disjointed manner.

Cern: What’s missing is the visual part. You can’t just put these agents down and have them visually look at all the different things going on in your computer screen and make a decision like a human can. The Digital Optimus could do that by taking the visual inputs. It’s a very powerful system.

Tesla also realised that they’ve got these 9 million vehicles around the world that are like computers on wheels. You can actually use the computers to run Digital Optimus. Think about them as little mini data centres on wheels that are cooled, that are powered with big batteries.

Lisa: And a possible income stream for Tesla owners who could rent out their extra compute while sitting in the garage.

Cern: That’s right.

Terafab: 50X The World’s AI Chip Output

Lisa: Tell us about Terafab.

Cern: Terafab is a very ambitious project that Elon has talked about a couple of weeks ago. He’s come to the realisation that if they’re going to produce as many autonomous vehicles and humanoid robots as he’s thinking — into the billions — and also putting AI data centres in orbit, he’s calculated the number of computer chips they need.

He’s looked at the semiconductor industry worldwide and determined that they produce about 2% of what just Tesla will need.

Lisa: Only 2%? The entire world?

Cern: 2%. The entire world’s AI semiconductor output is roughly about 2% of what Elon wants to do every single year. He’s been pleading with the industry to grow faster. They’ve been doing that, but it’s not enough.

So Elon has said, “If I’m going to bring this vision into reality, I’m going to have to build this massive semiconductor factory, the likes of which the world has never seen.”

The first step is building an advanced technology fab on their campus in Austin, Texas, where they’ll be doing research and understanding how to make the semiconductor building process even more efficient. If Elon is successful in building Terafab, they’re looking at 50x-ing the output of AI semiconductors in the world today.

Why Tesla Is Building Its Own Chips

Lisa: How does this differ from the chipset that NVIDIA is producing or Taiwan Semiconductor? Are they all different chips for different purposes?

Cern: Taiwan Semiconductor and Samsung make semiconductors for all kinds of customers all around the world. They have fabrication plants set up to change and make different chips. One customer is NVIDIA with their specific chips, then another customer like Tesla with their AI4 chip. That introduces a lot of inefficiency.

If you could have a semiconductor fab that just made one chip — that would be a much more efficient system. When it comes to Tesla’s AI4 and future AI5 chip, that’s a very purpose-built chip. Other companies’ chips might do the job, but Tesla knows what they need from their chip to be optimal for the full self-driving software.

If you can make your own chip and get rid of the stuff you don’t need, that reduces cost. If you can build a fab that just makes your chip, that massively reduces cost. Elon’s seeing an opportunity to massively reduce the cost and massively increase the output.

Lisa: There will never be enough chips in the world for the demand we’re facing. And 50x-ing — that sounds insane. But then a lot of things Elon takes on are. And he’s saying 80% of those chips will be used for space.

AI Data Centres In Space

Cern: If you think about NVIDIA’s chips today in data centres on Earth — a lot of those chips are used to train AI systems. Once built, people run them, and every time you run an AI system, you need those chips because the AI is constantly thinking. It’s not a software program written once that just runs.

If you think about a future where everything runs off AI thinking in real time, the amount of inference chips you need is a colossal number. Elon’s determined that the lowest cost way is to put all that in space. Run the inference in space because the energy costs using solar will be far cheaper than anything on Earth. And we get the added benefit of freeing up all that energy we’re using now on AI for other purposes.

He’s saying roughly an 80/20 split. We’re going to have a lot of cars and humanoid robots on Earth, but our big brains in the sky.

Lisa: Is this a collaboration between SpaceX, XAI, and Tesla all in one?

Cern: It is. Some have estimated that by 2028, it’ll be cheaper to put a data centre in space than to build one on Earth. A lot of it hinges on SpaceX’s ability to get the cost of launching Starship down and proving reusability. Another launch coming up in April.

If you can unlock reusable Starship and launch many per day, eventually one every 5 or 10 minutes — it may actually be cheaper to put things in space than to pay for cargo on a cargo flight across an ocean, on a cost per kilogram basis. Then you’re going to see a lot of activity in space. Not just AI data centres — manufacturing too. There are manufacturing processes that could benefit from a low-gravity environment.

Chips Designed For Space

Lisa: Is there a radiation resistance issue? Does he need special chips in outer space?

Cern: There’s a radiation resistance issue but I think that’s easily solved. There are some NVIDIA chips today in space that are operating. The biggest issue is cooling. AI data centres in space typically have a big solar array to generate power, but then you need a big radiator to cool the chips.

What Elon showed a couple weeks ago was a design with a fairly small radiator. He says they think they can design a chip that can run hotter, so they won’t need as much cooling. No one today is designing chips for space. Elon’s saying, “I have to be the one to do this because I see the need at scale.”

It makes sense for SpaceX and Tesla to partner on this. Tesla is already accomplished at designing its own AI chips. SpaceX with their Starlink satellites has a lot of relevant componentry. More and more people are saying, “We see Tesla and SpaceX converging. It makes more sense to put these two companies together because their interests are truly aligned.”

Lisa: They’re really AI-powered businesses. He must have planned all that out long before any of us saw it.

SpaceX IPO & A Possible Tesla-SpaceX Merger

Lisa: We’ve got the SpaceX IPO coming. Do you think that’ll be a 2027–28 situation with Tesla and SpaceX combining? Is that positive for Tesla investors?

Cern: It’s a little bit of a mixed bag. I do think the IPO will likely go ahead. In terms of a merger with Tesla, I think the earliest would be 2027. It makes a lot of sense operationally because the companies’ needs and directions are headed in a fairly similar path. Tesla will benefit from AI data centres in space and SpaceX benefits from Tesla’s semiconductor work and humanoid robots. They really need each other.

Tesla may also build a solar panel fab to make the panels that SpaceX needs for these satellites. There are some issues — when I wear my investment hat, I like being able to allocate to more than one company. If Tesla’s down, I can buy that. If SpaceX is up, I can sell some. If it’s one company, I don’t have that opportunity.

But I do think that together, for Tesla investors, adding SpaceX to that mix increases the long-term opportunity for both companies. To the extent SpaceX is successful in putting AI data centres in space, that helps expand the number of humanoid robots and robo taxis we can have in the world.

The Energy & Supply Chain Risk

Lisa: In the current world situation with the wars, the Straits of Hormuz, the bombing that affected the helium supply which affects the semiconductor supply — do you see any risks coming from what we’re going through at the moment?

We in New Zealand are one of the worse off as far as supply chains getting fuel stocks down here. We’re facing possible shortages of fuel in the next few weeks. We may have to start prioritising. All the tariff wars, COVID before that — we’re seeing the world in a big regime shift with global supply chains breaking.

Cern: In the short term, it’s a problem. It could be very disruptive. We’ll find out which companies have secured the supply of what they need. We’ve been through disruptions during the pandemic and it’s not fun. It takes a while to unwind.

The good news is that necessity is the mother of invention. It’ll force people, companies, and governments to figure out other ways of doing things that long-term will be better.

I feel like every country should have a minister of “what if.” It should be their job to walk around and say, “What if the ships that bring oil to our shores don’t show up?” COVID should have taught us that we are dependent and we don’t want to be dependent for essential things.

Lisa: Maybe instead of a bigger government, every person before they graduate from school has to do a “what if” analysis about something.

Cern: Actually, AI could do that much better. Put it into AI — where are the vulnerabilities of this country, the supply chains, the market structure? Have it as a real-time dashboard that’s constantly running and analysed.

Lisa: I might develop that app. The “what if” app. Might be a bit complex.

Where Robots Go To Learn

Cern: New Zealand is in a remarkably interesting and promising position. It’s a small enough country — it should be relatively easier to collectively determine what’s best for the nation. You don’t want to consume somebody else’s standards. You want to be involved in setting the future standards.

Jensen Huang of NVIDIA said that everything that moves one day will be autonomous. It just makes too much sense. Not just vehicles, but moving products, stuff around a warehouse, around a hospital. Anything that moves should be autonomous.

We need to stop thinking of robots as a distant science fiction thing. We already have robots and self-driving cars in the world today. People need to get comfortable with these new tools. Most of the time, fears we have prove to be largely unfounded.

I’m not minimising privacy issues — we’ve got to do serious thinking about how you construct a privacy regime in a robotic world. New Zealand could be — instead of just this amazing tourist destination and an amazing place to live — it could be where robots go to learn.

Lisa: Robot students! They have to go somewhere.

Cern: That’s good for Air New Zealand. All those unused seats on aircraft coming back — fill them with robots on their way to New Zealand to learn.

Lisa: You spent your first 10 years of your life in New Zealand. You love New Zealand as well as I do and you see the opportunities. Thank you so much for your insights today on what’s actually happening in the world and this far-reaching conversation.

The thing that gets me is the interconnectedness of this whole web. When something goes right or wrong, it’s going to have a very big knock-on effect, and we need to stay in tune, stay up on it, and be very proactive. Thanks Cern, for your insights as always. You’re a very wise man and I really enjoy our conversations.

Cern: My pleasure. Thank you, Lisa.