Fan Favorite: AI in Higher Education and the Supply Chain, Trump’s Tariffs Hit Lesotho

0:00Goldman Sachs 2026 outlooks examine the trend shaping the global economy. From global growth and regional perspectives to deep dives into asset classes and portfolio allocation, listen to Exchanges Outlook 2026 from Goldman Sachs. Bloomberg Audio Studios. Podcasts. Radio. News.

0:33This is Wall Street Week. I'm David Weston bringing you stories of capitalism. Unicorns used to be rare, but they're popping up more often and getting bigger by the day. We look into why larger companies are staying private for longer and whether we should be concerned. Plus, we're paying a lot of attention to the tariffs President Trump is imposing on China and Europe and Brazil. But we go to Lesotho to see what the Trump tariffs mean for a country whose exports may seem small by U.S. standards, but are big enough to threaten its local economy. And the pandemic made the world far more aware of vulnerabilities in supply chains.

1:11What is technology doing to overcome the problems we found? But we start with a major public university transforming itself and its students for a very different world. Michael Crow is the longtime president of Arizona State University. He's won just about every award there is for innovation in higher education. And his institution is working to make sure it keeps up with the times. We decided to change everything. Change culture, change design, change intellectual structure, use technology, move in new directions. and then we began measuring learning outcomes. For all the talk about changes in higher ed, including just this week, when the University of Virginia became the first public college to agree to Trump administration oversight, the biggest cause for change, AI, doesn't come up all that often in the political debate.

2:00For nearly a quarter century, Michael Crow has led the university that is by some measures the largest in the United States. Few people have thought as much about American higher education while also having the ability to influence it on such a large scale. But when we first talked with him a little over a year ago, we didn't cover artificial intelligence. A lot has changed since then. AI will leave a lot of white-collar people behind. Another uncomfortable truth linked to AI. AI. AI. AI. AI. It's not a bubble. Obviously the most disruptive technology in the history of mankind. Egalitarian access to knowledge is at the highest level in the history of our species.

2:40What we have is a walking, talking, reference library on any subject, and we never had anything like that in our society before. We wanted to know how artificial intelligence is changing the American college experience, what it means for the teachers and students who have now made it a regular part of their lives. But we also asked Crow about the outcomes for recent graduates, and how his school is preparing students for an economy that is moving very fast. What does that do to teaching? I mean, back in the olden days, we wrote essays. We wrote blue books back when I was there. How do you do things like essays and evaluations?

3:16Well, I think what has to happen, and we've experienced this at ASU with our 6 ,000 factor members, several thousand of which are already AI trained, is you have to up the game. Perhaps we were learning too slowly, too incrementally, too much in a regimented or industrial way. With the AI tools that are available now, you can up the game, enhance the question complexity, enhance the answer complexity, expect more of the students. We had somebody give their test out of the business school to an AI system and get everything right instantly. Well, then the test is too easy. The test is too simple.

3:44So it's basically a way, in our view, to accelerate learning, to broaden learning, and to speed learning. So you have to look at it as a new way to basically make the game more intensive. The model is always changing. So Plato, you know, was against the written word. He thought everything should be thought through verbally and communicated verbally. There were unbelievable forces against the development of the printed book. And so the internet and its development, the web and its development all had people that were against it. And so AI changes the model in the sense that it speeds it up and intensifies it.

4:15It personalized the learner's experience, but it doesn't teach those core things. There's no values being taught. There's no values being experienced. There's no lived experiences being built. So what we really have here now is we just have this massive hyperspeed calculator capable of going to all of the digitized information, you asking a question about that information and getting the most probable answer. It's all about the questions that you're asking. It's not about the answers. It's about the questions. And that's what people need to really figure out. Does it change the notion of cheating?

4:45I'll bet humans have cheated for quite a while. It does change the nature of what is your work. Now, if you're answering a complex question and you're using a reference library and an AI system to answer that question, that seems legitimate. If you're using it to produce your analytical response that's supposed to be demonstrative of your ability, well, then you're cheating. Now, you have to then build a system which recognizes the ability to gain access to these tools. Now, sometimes it's just going to be you in there by yourself taking the test because they've got to know that you know how to ask the question.

5:17You know how to derive the answer, that your brain works in a certain way. Beyond that, the AI systems are going to enhance learning in every possible way, and the idea of cheating will change. At this point, are there some things that you can learn that AI cannot teach you? Absolutely. I mean, an AI system can't teach you to be innovative. It can't teach you to be creative. It cannot teach you grit. It cannot teach you the discovery process. It cannot teach you—I mean, it's a machine. It's an advanced hyperspeed calculator. It can do things that you can't do. It can think around corners that you can't see.

5:53But, of course, so can a dog. And so it's a powerful analytical tool to enhance our mental capabilities, not to replace them. In order to ensure that AI is a springboard rather than a crutch, Crow says students and teachers will have to raise the bar. And one place he's already seen signs of AI's ability to supercharge progress is in the school's research programs. It's almost unbelievable. We have probably 50 research groups that are using advanced AI to solve unsolvable problems, to figure out how to process materials or manage the Mississippi River in a different way in terms of the flow of the water and the flow of the dirt and other things that go down the river.

6:33We've got people doing advanced chemistry now. We're using AI systems to think beyond the way that we normally think to create more revolutionary opportunity. There was a study recently done by some of our faculty at the speed with which you could complete the work equivalent to a dissertation in genetics, 14 days. What that means then is that the PhD that normally takes four or five years to set up the experiments, do the experiments, do the work, be evaluated, maybe the PhD student of the future will do the equivalent of 20 PhDs. That will speed up the cures for cancer. That will speed up the analytical tools that will help restore sight in human beings.

7:09That will speed up the techniques that use electromagnetic current to affect people with Alzheimer's disease and other neurodegenerative diseases, all of which are computationally limited. Certainly the world has changed enormously since I came out of college a long time ago now. But my sense is the rate of change has really increased, maybe even geometrically, how fast it's changing. How do you prepare graduates today for a world 20, 30, 40 years down the road that I can't even imagine? So there's no way to prepare someone for something you don't know what it will be, except one thing. What we call, we're attempting to take all the people that are coming to our university, 120 ,000 degree-seeking students, 700 ,000 other learners who are just taking courses with us digitally and otherwise.

7:53Can we help create you to be a master learner? Can we help you to be a person capable of learning anything, adjusting to anything, adapting to anything? It's really that, because we don't know what all of the adaptations that will be required are. We do know that you should be grounded in, you know, American history and economics and the role of democracy and certain subjects in math and science and so forth and so on. And then after that, we find a learning path for you to take where you learn to learn. We don't care what your major is. You can, I met a kid the other day who was majoring in opera and physics.

8:24Great, fine, fantastic. That's how that kid learns. And so that's what we're after. How do you create universal learners capable of learning anything? That's the pathway. way. We hear at 40 ,000 feet about a shifting employment situation for recent college graduates because of AI. Are you seeing any of that in the real world? We're not seeing that in our graduates. Now, the problem with people talking about all college graduates, there's more than 20 million people in college. A couple million go to what you think of as sleepaway colleges. They go to places where you're living on campus. The other 18 million go to college in some other way, community college, online, some kind of other course and so forth.

9:02So we're not seeing any change. You know, we're seeing the same level of anxiety. We're seeing the same level of the process where, you know, more than 95 percent of our students that graduate as undergraduates are employed or in graduate school within the first year, almost all within the second year. So we're still, you know, seeing good ROIs. But what we are seeing is students, you know, who are quite savvy, you know, adjusting their trends. So we're seeing a slightly downward trend in computer science and a slightly upward trend in double majoring and triple majoring. More people moving into analytics and supply chain and all kinds of other things.

9:39And so the market for learning is also adjusting. We measure our success based on who we include. Crow hopes the size and scope of ASU will help with that adjustment, allowing students to react quickly and build new skills for the changing world. And in a school made famous for opening its doors, rather than being exclusive, he thinks the most important skills of all can come from unlikely places. We've even got ways now that we're using advanced AI-enhanced robots to help people that aren't qualified to get into a particular college to do what they want to do to get them qualified. Guess what?

10:19when we get them qualified they have more grit and determination than anyone else who sort of walked into it from high school and they outperform everyone there's a theme going on right now that every college has been overrated yeah oversold what do you say to parents what we have is a way for your your child your student to learn on the path that's going to enhance their ability to be most adaptive throughout their life so don't worry about their major so we get these parents that say well my kid needs to major in accounting so they can get a job or they need to major in anything other than political science or history or English where they'll never get a job.

10:51That's actually not true. Some of the hottest things that we have that we're producing right now are English majors that can code. And so they have a broader perspective and they can code. And so we provide free coding classes to everyone in the institution. We provide other ways in which you can double major, triple major, take other kinds of things. And so what we say to parents is, let's find the way where your kid is going to smile while learning, while preparing themselves to be a master learner, and you'll have to worry about them less. If they take a fixed thing in a fixed way, in a fixed pathway, they could find themselves in an alley and no way out.

11:26We're trying to make sure that that doesn't happen. Coming up, transforming the U.S. economy by taking the bugs in supply chains and turning them into features. That's next on Wall Street Week.

11:44Goldman Sachs 2026 outlooks. From global growth and regional perspectives to deep dives into asset classes and portfolio allocation. Goldman Sachs research examines the trends shaping the global economy. For insights to help you stay a step ahead, listen to exchanges. Outlook 2026 from Goldman Sachs.

12:10As markets move and headlines break, what matters most is context. A Bloomberg subscription gives you unmatched reporting, sharp analysis, and powerful tools that help you connect the dots. Subscribers get unlimited access to Bloomberg.com and the Bloomberg app, including exclusive stories, premium markets tools, Bloomberg television, and podcasts, all in one place. Make smarter moves with the global benchmark for business news. Learn more at Bloomberg.com slash podcast offer.

12:45This is a story about featuring the things we thought we needed to fix. The pandemic focused all of us on the need to fix our supply chains. But what if we could use technology not just to fix our supply chains, but to rethink them altogether? Could we use them to transform our markets, our workforce and our economy itself? Stopping the spread of the deadly coronavirus has meant stopping the spread of more than just germs. Goods parts and people have been halted in place. I think the pandemic sort of brought logistics to the limelight from this like tucked in the corner industry to really, really be in the forefront of all of us, of consumers and society.

13:27And when supply chains stop, economy stop. Going back to the COVID days in 2020, we started buying online like we had never seen before. And so what happened then with all of that cargo rushing in, it was like taking 10 lanes of L.A. freeway traffic and squeezing them into five. But what if we could use technology not just to fix supply chains, but to rethink them altogether? Could we use them to transform our markets, our workforce, and our economy itself? Logistics is the largest cost for the manufacturing and retail economy. And if you drive that cost down, like anything else we have seen in the history of economy, you unleash the next epoch of economical activity.

14:17Lior Rahn recently became COO of autonomous trucking company Wabi after co-founding Uber Freight back in 2017. Their cost of goods moved is sometimes 30-40 % of their cost of operation. It's a prohibitor for them to scale and to be bigger. If you take that cost down, you essentially democratize access to logistics. You encourage more economical activity. if you reduce the friction, in that way it's friction in the system. If you reduce the friction in the system on manufacturing, on distribution, on moving those goods around, you just unlock a bunch of new economical opportunity. You unlock manufacturing strength.

15:00You unlock new predictability back to COVID. You unlock resiliency. And that set the stage for a rapid economical activity. The stakes are enormous. but how to get from here to there? Ron says the answer lies, in part, in that artificial intelligence everyone's talking about. I think AI allows you, the way I think about it, really like three levels of optimization and supporting function. One is logistics is a highly fragmented, highly manual ecosystem. You have hundreds of thousands of operators acting as glue, overcoming all those information challenges, doing a lot of repeated manual tasks, which is really sort of like taxing the system from a cost perspective and from an efficiency perspective, and more importantly, just from a time perspective.

16:00The second level is optimization. Can we now actually start looking at the supply chain holistically and start driving smarter decisions? Can we optimize those empty miles? Why does it need to be 40 % empty miles if we know everything and we could connect everything together and we can actually start smartly designing the network so you can actually minimize those empty miles? And the last level is using AI to truly drive better decision making. Let's have a chat GPT for my supply chain. And the most important one is unleashing AI to the physical world. With physical AI and self-driving, which I think is really sort of the deepest disruption and the most profound change that we'll see with AI in supply chain in the next decade.

16:52Part of what Ron envisions for the future is happening right now in the Port of Los Angeles. There's so much here, and I think we're just scratching the surface. Gene Sirocco is the executive director of the Port of Los Angeles, the busiest container port in the United States. One of the areas that we stepped into right away was digitalization of all the port's information. How many ships are coming in, containers, what trucks and trains need to be planned? how do we get our great skilled labor at the right place anticipating the cargo that was coming in. So we worked with the Wabtech company to develop the first information sharing system for a port here in the United States.

17:33We call it the Port Optimizer. We can now see cargo 40 days before the ship arrives into Los Angeles, giving us ample time to plan all those things, the skilled labor, the land, the machinery, and just that intuitiveness about how we're going to handle things if something doesn't go to schedule. And normally that's the case. There's always an adjustment that has to be made in the supply chain. So now we can see things every morning, a dashboard of information about the velocity, the vital statistics of the port. And I could tell you after about 90 seconds of a review how we're doing and what we need to do next.

18:10That digitalization started to introduce prescriptive and predictive analytics. Then we start to get into really, really interesting engineering work. We're going to have a project here on the Vincent Thomas Bridge to resurface it. So we're using geospatial mapping and great companies like Esri and the Jet Propulsion Lab to help us now simulate traffic patterns ahead of time and in real time, allow drivers to move around with much more knowledge than they had before. It's going to help this trucking community out in a tremendous way. What Soroka is pursuing within the Port of Los Angeles, Roger Penske is doing on roads coast to coast and even around the world.

18:53We do 500 ,000 vehicles. We're in four continents and nine countries, and we have our truck leasing, rental, and logistics, and we have 44 ,000 people. So it's a real enterprise. From artificial intelligence, what we're using in the truck leasing business, We're taking, downloading 200 ,000 vehicles a night with operating data. When I look at the amount that we capture on an annual basis, it's a billion units of data. We run 5 million, 600 million miles with our trucks. And when you take that, we need somewhere to aggregate it. So we're starting to use what we call Catalyst AI. It's an AI product where we look at the data, we diagnose the data, we go to our customers, and we use this as a connection.

19:39The Penske company's use of AI is not limited to monitoring its massive fleet. It also uses it in the maintenance required to keep that fleet on the road. We then become to build a process where we have predictive maintenance. So AI is telling us this truck should come into our shop. You might have the same truck but a different duty cycle. So when that truck comes in our shop, what happens is the mechanic goes to the computer, He scans the barcode or the VIN number on the truck. He puts on a headset with a mic and what we call guided repair. And that guided repair takes him through the whole maintenance process on the trucks.

20:20And then that data, because it's live, goes into our SOS center, which is our call center for breakdown. This whole avenue, downloading the data, taking it, making it real. Then we take it, and for the customer, we'll go back to you as a fleet, and we'll take and give you your data live and how we compare location to location you have. We can look at that and then determine what are the things that we have to do to take action. We have 15 ,000 customers, and many of them don't have the opportunity to have this kind of data. So it's sticky, it's important, and it's taking all the technology that we can get.

20:59Whenever we begin to talk about automation and AI, people raise questions about jobs. What's the effect on jobs? Are you going to eliminate a lot of jobs in supply chains? It's a hard job. And many folks, many young folks do not want to take on this job. Being 300 days on the road, not being able to raise a family, not being able to have predictable access to sleep and food. I think for the foreseeable future, as you start rolling self-driving gradually, you augment those jobs. You fulfill the empty spots in the demand for those jobs. You don't replace those jobs. There's going to be plenty of time for the existing population to retire over the next few decades.

21:45As the truck replacement cycle, again, it's like four million trucks in the United States. It's going to take a while until self-driving will actually sort of make an economical dent to the driving population. So I see that as a gradual job transition on self-driving. I think if you look at knowledge workers, I think we'll see a faster transition because the digital AI is even more prevalent and will scale potentially even faster. I think there it's about an opportunity to up-level. So at Uber Freight, we've automated a bunch of those manual repetitive tasks, but then those folks have been up-leveled and they're now more sort of an orchestrator of an orchestra.

22:27They are the brain behind actually orchestrating all those agents, because those AI agents can now do this job and that job and that job, and I'm sort of more the integrator. There is no replacement for the human connectivity, so I can spend my time focusing on customer needs. I think in the idea that people think that jobs will disappear, that it'll be just talking to AI, I think it's overhyped. Chris Kaplis is the executive director of MIT's Center for Logistics and Transportation. I think there's a lot of overhype in that it can truly take over a human's full job, when in fact jobs are tasks and it'll take over some tasks.

23:06And so then there's certain tasks that each one of us in our jobs do that cannot be done by AI, to include the person filming us right now. It's more the mundane things that get automated, and then the things that can help enhance what a human does. I think there's a continuum on how you automate things. We've been automating stuff forever, right? But there's a line of, this is mundane, and over here I still need a human. That line keeps moving as AI gets smarter. I don't think it'll go all the way. Artificial intelligence may well transform our supply chains, but to get there, it's not enough to have AI alone.

23:43It has to be linked to autonomy. What is AI without autonomous vehicles? What are autonomous vehicles without AI? I fundamentally see the next decade really about building the autonomous infrastructure of the industry. And I view self-driving as the most profound help in driving efficiency and safety into supply chain. The whole digital infrastructure in the end of the day can drive 10, 15 % optimization of the network. When you start actually moving assets in a self-driving way, when you can move an asset from six hours a day of being utilized with a human driver to 20 hours, 24 hours, When you can drive safety into our roads with thousands of unnecessary deaths and half a million related incidents in the U.S.

24:34every year on trucks, you can make a profound change and a profound help to the industry. This is the time. I think it's a confluence of many things really coming together to make the next year, not even five years, the next year transformative from self-driving in the industry. Like, this is game time. This is real now. It's happening.

25:01Up next, Lesotho may not come to everyone's mind as a textile exporter. But it was on President Trump's list of tariff recipients for not playing fair. We'd go to the country in Southern Africa to see what a small kingdom can do to stand up to the world's largest economy.