Episode Transcript
[00:00:00] Speaker A: Welcome to the Blueprint brought to you by Brilliant Pathways. I'm your host, Manny Tejeda. In this forward thinking podcast, we explore the evolving challenges and opportunities in preparing students for college and career success, offering expert insights, actionable strategies and real life stories to help listeners better support young people as they draw their future.
Today, we're talking about the question that's on the minds of employers, educators, families everywhere. How is AI changing work and what does that mean for our students?
AI is already shaping entry level and early career jobs. We'll chat a little bit more about this. But tasks that used to be reserved for new hires, the researchers drafting the basic analysis, are increasingly being handled or accelerated by AI tools. So that raises some big questions about what does entry level even mean when I can do much of the groundwork? How do schools teach and mentor students for a world where AI is a normal part of the workplace? And maybe most importantly, what do young people still become great at when AI keeps getting better?
Joining me today we have James roth. He spent 12 years at Goldman Sachs, working across capital markets, investment banking, sales and trading.
His experience in financial technology and active investment in tech focused companies led him to FinTech, where he served as head of business development at Galaxy Digital, a digital asset investment company, and most recently as Chief Revenue Officer at Mainbrain Labs, a software platform for digital asset capital markets transactions. He's also a longtime board member of Brilliant Pathways. James, welcome to the Blueprint. It is an honor to have you here today.
[00:01:50] Speaker B: Oh, Manny, thanks for having me. Pleasure to be here.
[00:01:54] Speaker A: So, James, I know it was a kind of a long introduction, but you spent a lot of time in finance, you've worked with AI across different industries. What are your thoughts in terms of AI in the workplace, particularly as it relates to entry level jobs in early career positions?
[00:02:10] Speaker B: Yeah, absolutely.
Yeah. I know there's so much, whether it's doomerism or fear mongering or otherwise. So I'm going to stay out of the realm of science fiction, try to predict the future, which I think the only thing I've learned in my professional career is that nobody knows anything about what's going to happen in the future.
[00:02:29] Speaker A: Right.
[00:02:30] Speaker B: But you know, and I'll try to keep it in, you know, my experience, what I see at least here, right in front of us, and then maybe some, some strategic or tactical things that can be done around where we are right now. I think the issue is, is that things are moving so rapidly and basically right after we have this conversation, it might be obsolete. But, but let's let's hope there's, there's a lot of nuggets that can be taken from here, but in terms of, in terms of what we're seeing, I think things were happening somewhat slowly and it was a bit of a drip and then are happening all at once now that the capabilities of some of these tools have really grown exponentially in terms of their capabilities, their subtleties and the things that it can do and the speed at which they can do them and the breadth of what these tools can do. When I think about my own ARC and career arc and, and what's happening today, if you rewind the clock and go back and let's take an entry level job in finance in terms of a financial model or somebody in investment banking that needs to value a company. And so they're tasked with putting together a financial model which goes through accounting statements and they use a discounted cash flow analysis to come up with a valuation of a company. Very high level.
That process was one born of iterations in Excel. Building a model that requires an intricate understanding of accounting, an intricate understanding of the underlying business model, working through, connecting all the various financial statements together. And that's usually done in concert with somebody who's senior. But you know, you're in there building a model potentially from scratch. Maybe you're using a marginal template that somebody internally at the company has built. But you know, it's a, it's a, it's a cycle of building a model, making a mistake, having something breaking. Maybe there's a cell or a formula out of place. Maybe the thought process or logic behind how to build the model is somewhat suboptimal. And so it's an iterative process where you're learning by doing, things are not working, things are working, you're making mistakes, things aren't adding up and you're able to kind of cycle through. And that process takes many, many, many hours. It could take multiple days over which you're getting tutelage from somebody senior who has more experience. They're helping you develop your thinking.
And then you're obviously doing the execution of actually building the tools in the model and maybe putting that into a PowerPoint presentation. Fast forward to. And there's a lot of learnings to take in that process. Fast forward to today. You know that that has been condensed into about 30 seconds. And the models are in some instances better than what a human can do, at least in terms of not do, let's say, but better than a human can do at that speed.
And the intricacies and the subtleties and it's debugged everything and everything kind of comes out beautiful and it's beautifully formatted and then a PowerPoint presentation can be put together after that. Now, maybe that PowerPoint presentation and that model requires tweaks. It almost certainly does. Maybe it requires. Oh, there was an underlying assumption that was off and so we need to go ahead and fix that. But. But the actual doing of the execution of the build took about less than a minute. And I think that is exciting in some ways because maybe it means that the work then moves to a more higher order thinking work, but also scary at the same time because it means a lot of the work that was done by entry level workers is now going to be done by a machine and potentially with far fewer humans needing to manage those machines. And so the implications are huge because that's one small window into finance. But that's happening across the knowledge workforce and it just has big implications and frankly, it can be exciting and scary at the same time. And it's very uncertain. I can't remember another time in my life when things have been this uncertain as it pertains to the future of most things, let's put it that way.
So again, don't want to speculate, but that's the real tension there, as I see it, which is the apprenticeship model, which in so many disciplines was driving how the learning was done for junior people that ultimately move up and make it to a more senior role, whether it be in medicine, whether it be in law, whether it be in finance. But a lot of that knowledge work was, was done with tutelage from more experienced folks. And when that process gets broken and the model, or, you know, writing the brief in law was how you learned through doing well, the process starts to break down a little bit. And the question is which employers are going to lean into the training model of it's not just shove everything into the machine and let the machine do it, it's how you think about what you put into the machine. It's the taste and judgment that comes from a senior person with mastery over something and then how you manipulate the AI to give you the results that are vastly better than what a human could do on his or her own. And so that's the tension. And I don't think firms have really kind of figured that out, but it's real, it's here, it's happening, and we're. It just depends at the pace of which it's getting adopted at each particular institution. I'd Say smaller companies, easier to adopt, easier to get in the door, bigger companies, it's happening a little bit more slowly because there's the logistics of moving it through.
But, you know, it's, it's coming, it's there.
[00:08:14] Speaker A: Yeah, that's, that's all very interesting and a few things that come to mind for me, you know, the, the making mistakes, that's huge. That's a learning opportunity that could go away. Right.
The ability to work with more experienced folks going back and forth. I mean, that's what we talk about mentoring, right. Like showing you the way. So there's a lot that could be lost.
So. And you talked a little bit about how that would happen or could have implications in the workplace, especially as, as you folks go into positions where there is a model for teaching you the way through apprenticeships. But what do you think, thinking about in the school setting, what does that mean for schools? How do we inform what we should be teaching and how should we be mentoring students? When it's not so much about the information, right. But. But more about the critical thinking skills that you might need to figure out how you asked AI to give you the information or how do you assess the information to make good judgment into. Is this good data on what's been fed or like, how do you make that determination but really want to focus in how do you think that plays out, will play out in schools? And I know we're not making predictions, but thinking about base of that stretch that we just talked about, how would that. What could that look like?
[00:09:41] Speaker B: Oh, super loaded question. Lots of ways to go there.
I can.
Let me just start with, I guess, a couple of observations. So you're absolutely right in terms of learning by doing and learning by making mistakes.
And I also think in terms of being able to learn, it's all about, in some instances, it is about repetition and it is about making mistakes. So, like, one of the things that I think a lot about when I think about schools is, and this is going to seem further afield, but I'll bring it back and it'll kind of make sense, which is if you think about the advent of the phone and the iPhone, that is, or the smartphone, and you think about GPS living on the phone. Well, GPS on the phone was able to create mapping technologies where all of us now probably don't get into a car without turning on our GPS and getting to a particular location. Right. There's no longer are the days of pulling out the atlas and trying to find your way. Now what science has shown or what studies have shown. And I'm obviously not a scientist, but I've read the results. And effectively what it's shown is, is that the portion of the brain that's responsible for wayfinding, I believe it's called the hippocampus, but has shrunk dramatically since the advent of the smartphone because nobody is using that portion of the brain to find things or locate berries or, you know, the human needs have changed so dramatically. And what that shows, at least to me, is that if you don't use it, you lose it. And so that is the very scary proposition for AI. And already studies are showing us that when you leverage AI to do everything inclusive of critical thinking, those parts of the brain and your ability and those skills atrophy. And so that is ultimately, I think, the temptation that we need to resist as it pertains to education. And so for young folks, you know, it isn't, I don't necessarily think it should be about putting our heads in the sand and pretending like AI is not here and it's not going to impact you. But I do think that everything now moves from the not how you do it, which is, call it the execution or the mechanics or the formulas or the steps. It's more about what you do and why you do it. So, meaning human taste, human judgment, which problems matter to solve. What's a good question? What's a bad question? Why does this approach work?
What would break this approach? So it's more about.
And again, this is more what I'll call, I don't want to say softer skills, but it moves more towards, in my mind, critical thinking and about how to think. And why would a potential assumption hold as opposed to not another one? And so it will, even with the advent of AI, it'll still require a human who actually knows how to think. And it requires mastery of a subject. So in terms of, you know, what that means is it means teaching judgment. I also think human skills, so being able to socialize, being able to public speak, being able to debate, being able to communicate effectively, those human soft skills, I think will be dramatically elevated.
And so I think in terms of. And then the other thing to think about is that AI is developing so rapidly, whatever we teach a student today about AI is going to change in the next three to five months.
Because when I first started playing with these tools, call it, you know, a year ago or more, or whatever it was to today, the step function higher has just been dramatic and so amazing to watch that it's almost like what I was using early on in terms of AI, the rules have completely changed and now there's a whole new paradigm of what to use and that'll change even forward. It might just be, you think something and then the AI does the job as opposed to actually needing to type something, God forbid.
So what I think what it should mean for schools is not outsourcing any of the thinking to AI.
And you, you know, it's, it's this.
I was explaining this to my son the other day around who's 10, and he was asking me about computer programming. And so I started showing, oh, you know, here's Python and here's what we can do, and let me show you how these small math problems work. And then we went over to an AI tool and I said, you know, build me a replica of Space Invaders. And it did in about 30 seconds and you have a full fledged game that kind of rolls down. Well, all of that was built on Python underpinning everything in terms of the back end technology.
But now he's gonna be in a position where the temptation is probably too great and he'll want to skip all of the hard work of understanding what's going on under the hood and jump. So I think, sorry, this is a very long, kind of long winded way where I kind of moved around. But I do think if we skip to the AI and that's all we kind of teach students and we don't teach them how to write, how to read, how to do core concepts in terms of math, we don't teach them, you know, how to think, then all is lost, you know, and the machines have definitely won. Because what the science probably will say is that the more we use it, the dumber it makes us. So, you know, pretty soon we'll have nobody thinking if all we do is AI. But, but the biggest thing that I would say is, is that the folks that are best at AI are the masters of their craft that can leverage AI to make them 100 times more better at what they're doing because it is able to provide a broader aperture through which they can look at things. Whether that's a researcher who is telling an AI exactly where to go and what to look for, or it's able to find research papers in other languages that you wouldn't otherwise have been able to do if you didn't have access to AI who could translate it for you. So there's lots of examples of where this can elevate, but that's only once you've learned mastery in all these subjects so that you can leverage it appropriately. I do think there's a lot of applications for teachers today because they're masters of their craft and they understand how to use it in such a way.
But, but that's just a long winded way of saying I think we have to be really careful about when and how we introduce these things because I just think the temptation will be too great. In the same way social media, it's, you know, I'm not reading, I'm going to be scrolling, right? And I think that is the, the easy button is just way too easy to hit.
[00:16:07] Speaker A: Right. You mentioned a couple of things here that I want to go back to, like the skills that we got to make sure students are developing. Right. But for, for a teacher, for instance, who don't shy away from, from the tool, but also don't shy away from having the students develop like the public speaking, the critical thinking.
So those are all great examples on how our teachers might use AI. And quite frankly, they might even have the students think of other ways in which they would want to learn. Like how do you use AI to teach history? Right?
[00:16:43] Speaker B: Oh, totally. I think there's so many interesting things that can be done. For example, I'll use my son as an example because I'm not a formal teacher. But, well, let's take it in two steps. There's one which is anything that a teacher does that isn't, you know, impacting the students directly and leveling them up, that is rote and can be automated away. So, for example, grading, right?
The ability to, let's say, you know, you're a writing teacher. Well, you can train an AI to, hey, this is what good structure looks like. This is what good sentence construction looks like. This is what good, you know, a good outline looks like. This is what.
How I'd like this particular persuasive essay to be formatted.
And here are the arguments that we'd like to see as more of a higher level concept. And the AI can use that rubric to do a first pass on the various essays that would normally take a teacher, let's say, you know, 20 or 30 minutes to read. And then they can distill that down to kind of the more interesting or more nuanced parts of the grading process so they can provide more elevated feedback.
That's potentially one example.
Another would be on assessments. Hey, generate 50 practice problems for algebra.
Maybe the teacher then picks 10 out of those and modifies the three based on common misperceptions. She's seen and then maybe she adds it, maybe she will add something around an actual interest that a student may have. So the customization can really get there. So you could say something like, give me five lesson plans contingent on how somebody learns. You know, is it more tactile? Is it more discussion based?
It's more kind of working smarter as opposed to working, you know, harder when maybe, you know, she, or maybe she can produce a video and then she has it animated. Right. Right away, as opposed. That might be more engaging. I did something with my son the other day where was asking me a little bit about Napoleon. Well, you, one interesting kind of concept is you feed AI, you know, his, his writings, his thought process, his logic. And then you literally can debate Napoleon through the AI, right through a voice, or ask Napoleon questions or why did you do this? Or what did you do there? And it kind of can spark a debate. So there's lots of hypothetical. What I've actually found, which is interesting, is that when, as you start to get into this, if you're intellectually curious, you actually end up doing more work because you're like, oh my God, this can go further and deeper and wow, this is kind of creative. So sometimes there's a paradox of choice there when the whole world is your oyster. But I do think that, you know, those are just a couple of examples of how maybe we can take the teacher's game to a higher level of engagement and do away with a lot of maybe the rote or boring or like, call it like, you know, lesser interesting things there and they can spend more kind of time with the students in elevating kind of those, those skills.
[00:19:50] Speaker A: So no, those are great examples. I appreciate you saying that.
So let's flip it. And now let's, let's think about, for instance, your son as a student. What would you say are the skills and things that they, that they need to be really working on and developing to set them apart. And in the future, going back to obviously joining a field where AI is part of the workplace, what could be some of the things that they can do now?
[00:20:20] Speaker B: Yeah. So I mean, for me, I think as AI keeps improving, what should the students spend more time being great at? I think it's foundational learning. So that would be honestly, maybe call me old fashioned, but I do think it goes back to the basics, which is like reading, writing, math, logic, verbal communication. Like we gotta dial those in and try to spend as much time on those with as little AI helping as possible so they can get the reps, particularly when they're also younger too. I would say, you know, it matters whether you're 18 or 12, right? Obviously, for where you are. My son's 10. So in this example, you know, I'm trying to hold off on. I'm just giving him snippets of what AI is possible of, but not giving access to it so that, you know, you have to go and earn those hard yards. So that would be. Those are a few things, I think taste is another. So, meaning, like, what does good look like?
What does weak reasoning look like, What a strong reasoning look like? How do you ask really strong questions? Because that's what, at least right now, as the AIs exist, the chatbot, that's what makes somebody very good at using AI, which is it can spot weak reasoning, it can spot holes in logic, and it can ask really strong questions to get to an even, you know, a third or fourth level effect there. And then I also think we have to teach kids humanity, right? So the soft skills of like, courage, integrity, leadership, empathy, I think machines, maybe we can encode them for that, but probably not, in my opinion. So I think also teaching of humanity is a huge one that I think we should put an emphasis on. So I think being deeply human, deeply discerning, I think all of those things are super important for the kids as they go forward. And a lot of what I just mentioned includes a lot of those softer skills which are around developing relationships. Because unless I'm missing something, humans will still do business with other humans.
So, you know, being able to be persuasive. And so I think what school may end up looking like is, you know, maybe there'll be more engagement with a tool like an AI that can engage with the student at their level, be more individualized, gives decent access. But you have teachers that are really spending the time guiding in these other areas where they're just, you know, maybe you aren't going to have to learn as much rote skills and it'll be more like instead of learning facts, for example, it'll be more around, okay, let's talk about how to debate using said facts, or let's talk about how to craft an argument using said facts as opposed to, you know, hey, memorize these facts on what happened in, you know, the Korean War, for example.
[00:23:23] Speaker A: Right.
And I think that that's really, that's really awesome. I want to reflect again on the mentoring piece because I think when you started, you know, how does it look? How is this changing in the workplace? That's so important and it's one of our, you know, core practices.
But also on the soft skills that we call the essential skills, I think that, that leadership, teamwork, that, that, that could be all embedded into how learning happens.
And to your point, those are the skills that perhaps are not, we know, are not with AI Maybe it could be in the future, but still the transactions will happen between humans and we continue to use it as a tool.
[00:24:07] Speaker B: Yeah, I also think, I also think, I mean, maybe this isn't directly related to what you were just saying there, but I think goes back to, you know, what can we do now and what do we teach kids? I think, I think it should always be like this, but now it becomes dramatically more important that the process matters more than the output. So how do we have you create an outline? How do we have you identify the right sources?
How do we isolate your reasoning and then maybe reflect on how AI was able to help or maybe it failed you in a certain place, and then you actually learn what they actually understand through that. And so, you know, I think that's, that's a big one. And then to your point, it's how do we get more discussion, more presentation, more debate, more teamwork, more mentorship.
Those are way more valuable. And so I also think, just taking a bigger step back, I think it also then will boil down to helping kids find out what they are encoded to do and what motivates and what drives them. So, meaning more emphasis on the what problem are you going to solve with your life?
Because that is what hopefully can outwork a machine. Right. The drive to ask questions that are beyond the realm of curiosity. It also will create the drive to continue on when things get hard. I think spending a little bit more time on that, as well as to, you know, where you should go into a career, I think is really important. And that will just, you know, that'll be highlighted as, as, as AI becomes really good at, you know, kind of most things.
[00:25:56] Speaker A: Yeah.
Well, James, thank you so much. Just to recap, AI is changing not only what entry level work like, but also the skills students will need, the experiences that students have in the classroom.
It doesn't make our work with students less important. If anything, it makes their human strengths more important than ever.
For educators and school leaders, this is a moment to rethink how we prepare young people and for students, use it as a tool to handle the busy work, but not to replace your curiosity, your character, your ability to connect with others, and your capacity to lead. The advantage will go to those who use it wisely and bring something deeper to every classroom, every project and workplace. Thank you for letting the Blueprint be part of your growth journey. And remember, at Berlin Pathways, we believe learning isn't about following a script, is about building the curiosity, empathy in agility to navigate in a constant change. Thank you all so much. My name is Manny Tejeda and I'll see you next time.
Support for this podcast comes from Brilliant Pathways, an organization that had spent over 30 years helping students make opportunities happen.
Connect with us wherever you get your podcasts, and thank you for joining us on the Blueprint. I'm Manny Tejeda, your host, and I'll see you next time.
