940: In Case You Missed It in October 2025

0:00Jon Krohn:This is episode number 940, our In Case You Missed It in October episode. Welcome back to the Super Data Science Podcast. I'm your host, Jon Krohn. This is an In Case You Missed It episode that highlights the best parts of conversations we had on the show over the past month. To start things off in episode number 933, I ask Seamus McGovern, founder of ODSC, the world's largest data science conference, if we need to radically rewire our professional skills to keep step with the rapid developments in AI. You mentioned earlier in this episode about how you've been doing this meetup tour across the United States.

0:37Jon Krohn:You mentioned Chicago, New York. And before we started recording, you mentioned to me that when you speak to people, a lot of these people you speak to are concerned about the pace of change due to AI advancements. And so you talked to me about this idea of rewiring professional skills. So what does this mean? How can people be coping with the pace of change that is coming about because of these rapid AI advancements and proliferations? Yeah, I got so many of those questions. And I've had to do this myself. You have to think about it in two parts. One is you have to kind of get comfortable with the speed of change, right?

1:25And I think most of my career, I've been comfortable with that. That's one of the reasons I left very large companies and started my own startup. I want to do things quicker, faster, right? So it's not for everybody. But you do need to get comfortable with that speed of change. And there is actually an advantage there as well, because it's been proven throughout time, you've seen this over decades and centuries, that the pace of change always outruns most people's ability to absorb it, right? And that, of course, leads to a lot of angst, but also leads to opportunity. So if you are one of the people who learn how to deal with that, right, and what that means is you have to kind of be comfortable with this continuous need to rewire.

2:17And I think that's what people have to understand, that the speed itself is the new challenge, right? And this is something I've been kind of studying a lot. Like there was a study I might send it to you for the show notes, but I did read this study in AI exposed industries. And I'm not just talking about data science now, but AI exposed industries or jobs rather. The skills turnover, something like 30 something percent. So more or less, I don't know if it's compounded or not, but more or less every three years. I don't quite believe this, but even in the low case, the skills requirements turn over.

2:56And you can kind of see that like people now have to learn prompt engineering, vibe coding. and um you know that's and that's and that pace of change it leaves a lot of gaps right your company may be moving fast um you're moving slower you're moving you're moving fast the company's moving slower the industry and stuff like that and then i've listened to a lot of your podcast john of course and you can see as well that compute is doubling every six months right that's having um a big problem and so getting comfortable with the pace of change is is important because I started listening to this show last year about the history of the universe.

3:37It's my AI detox program. Great show, History of the Universe. What's it called?

3:42Jon Krohn:The History of the Universe is the name of the show. History of the Universe. It's on YouTube, History of the Universe. It's my AI detox. It's about astrophysics and astronomy and all that kind of stuff. Believe it or not, until about four or five years ago, I didn't know the universe was expanding. So I always think of AI skills like the universe. They never stop expanding. And just like the universe, we don't know what it's expanding into, right? You know, the universe is a bubble or there's multi-bubbles and all that kind of stuff. But I'm going off on a tangent here. But anyway. No, that was a good analogy.

4:10Jon Krohn:I like that. I hope we turn what you just said into one of our animated shorts for this episode. Because that's a great visual. This idea of, just like the universe expands, skills are expanding. uh you know the quantity of them you know like we talked about earlier you know data science branching off into ai engineer and lots of other more specific paths and now ai engineer will will branch itself into lots of other sub careers um so yeah i think you've nailed it with that analogy anyway i've taken you off track now yeah yeah and um and and and look i've been i've been in industry three decades in tech right in tech whatever um skills never go away there's just more of them, right?

4:57And that's why it's just amazing when people get so concerned, well, this is no longer being needed. They still need COBOL programmers. But anyway, back to the rewiring. Yeah. So when I talk to people about rewiring, they're like, okay, well, what does that mean in practice? And yeah, it's such an important thing because I kind of take an obvious, very optimistic, almost, I would say, hardcore view on that. Because I really think, and I'm doing this myself for myself, I think, you know, AI is moving from, we're moving away, because back in 2015, we were using data science, machine learning, and even AI as a tool, right?

5:41It's moving away from AI being a tool to AI as being a collaborative partner, right? And I do think those collaboration skills you build now will help you over the next decade, right? Because that kind of helps you with that. Can I build something now that can future-proof? So I say, you want to wait or you want to build, right? That's kind of important because when we look at AI, it's quite clear now it's either going to be, AI is either going to augment or automate, right? And both of those present opportunity. So I really think that when I talk to people about rewinding their skills, first and foremost, stop worrying about your skills being replaced and start thinking big picture now what's possible with AI.

6:28You know, I remember you said this in one of your podcasts. You said something like, with AI, you can do work now that was previously impossible. That's absolutely true. I'm doing stuff that I would never have done without AI before. or like, and forget about the role of a data scientist or a machine learning engineer for a second. I talked to a lot of startups in my other role. And also I saw it at ODSC. All of a sudden we had these people from sales and marketing showing up and they're all talking about agents. They want to learn about agents. If you think back, maybe a startup yourself, when sales was a decade ago, right?

7:04In sales, there was, you basically had an account manager and account executive. Now in sales, you have a lead gen specialist You have an SDR, you have an account executive, an account manager, different roles, by the way. You have a sales engineer, customer success person, a revenue officer. If you take AI, a lot of those roles can be rolled up and AI can either augment or automate those. And then think about that salesperson. Let's say you were just doing SDR or you were just doing account management or you're just doing sales engineering. you can now do a whole lot more, but you're going to have to rewire your skills because of AI, right?

7:46And, you know, that's kind of, and again, I've been studying this a lot, and the more questions I get about it, the more it's kind of a continuous loop, the more I kind of study it. And, you know, as you know, we have our own podcast, which we need to have you on or back on. and we had, was it Robert Brennan? Sorry if I'm mispronouncing his name, but he's the CEO of Open Hands, which was an open source version of OpenDevon, which allows you to automate your work. And I asked him the question, shouldn't people be worried about just replacing their jobs? He's like, look, Seamus, most work today is drudge work.

8:31And I really started to research that. He's right. Like if you think about the average person in office, they're looking at emails, they're doing admin. It's mostly drudge work and there's this whole productivity paradox and automation paradox. Like even though we got productivity with automation, the problem with the automation paradox is automation still needs oversight. It still needs judgment. And so, yeah, I think the new roles are going to be, as I said before, you rewire your skills, going back to data science and AI, less about building models from scratch, more about designing workflows, managing supervision, evaluation, and, you know, less to be builders and more orchestrators, or less to be, you know, building from scratch and more orchestration.

9:23Jon Krohn:orchestrators in an expanding AI universe sounds like a great way to rethink our approach to white collar work. And indeed, many companies are actively encouraging their employees to work alongside AI applications. Garobi's mathematical optimization guru, Jerry Yurchison, tells me how Toyota optimized their planning process to manufacture their vehicles. Something else that you have for us, I think that's completely new since your previous appearances on the show, are some interesting new real-life use cases of mathematical optimization. So you have, of course, alluded to some of them. We've talked about the burrito optimization game, or as a toy example, or the new Garobine coffee example.

10:07Jon Krohn:You've mentioned that application areas like supply chain, logistics, those tend to be areas that use mathematical optimization a fair bit. But I'd love to dig into a few more cool real-life use cases that have cropped up in recent months? Yeah, so we recently had our, what we call the Groby Decision Intelligence Summit. It's our fancy sort of event that we put on ourselves. We invite customers, we invite prospects, we invite anyone who's interested in learning more about optimization. You know, we invite you to come. Last year, it just finished up a couple of weeks ago. We were in Vegas. um and so we're bringing in like you know um uh super cool customers that are doing really cool things and we had a couple talks that that i thought were i like how you had to pause there because you're like it like had company names go into your head and then you can't say them so we get a some pause yeah really cool companies um but i will mention a couple there's a couple that I can mention.

11:16There's some that I can't, sadly. Again, we're the best kept secret in decision making. I guess that's what, if you're going to come away with anything, optimization and Garobi is the best kept secret because people don't like to talk about us because, yeah. Why would you spill the beans? But there are two presentations that I really liked. One was from Toyota and they're talking about how they used optimization for for planning of vehicle manufacturing. So they're getting, you know, sort of, you know, demand forecasts of like, okay, this is the number of this type of vehicle that I expect to, you know, to, that customers would want in this region at this time.

12:05So you can sort of see if you're thinking about by region over a certain amount of time, the whole sort of fleet of Toyota vehicles that they offer, you know, that's a pretty big problem. And now you're thinking about like, okay, manufacturing that, how can I best manufacture these things, these cars, at minimal costs and everything, you sort of see all of the small things that trickle into making a car is a very complex process. But so they ran through how they're sort of building tools. And there is an aspect of LLMs and natural language in this as well, but they allowed their planners to sort of interact with an optimization model that, you know, an optimization team built this optimization model, but they allowed sort of their planners to interact with that and do scenario tests and what if analysis on all of these sort of things, I'm like, well, what if the tariffs on this particular thing, you know, what if tariffs go up by, you know, from 0 % to 10 % and then next week they're 80 % and then the week after that they're back down to 10 % and then sometimes they're 30%.

13:24You know, this is a, it's an insane time to try and plan long-term manufacturing right now. It's like insane with all this sort of fluctuation of particularly tariffs. but it they had a tool that was at optimization in the back had sort of an LLM sort of interface where where the planners can really interact with this and say okay well what if tariffs are this or what if you know my supply of this this thing was cut in half or something it's like interacting with with the optimization model in a very natural way and getting all these sort of cool scenarios and really being able to understand, okay, what if this happens?

14:06What should I be doing? How should I be manufacturing things at sort of like at some macro level and really making decisions that will impact the company? And it's just providing like a whole new way to access optimization to people who don't, they're not gonna be writing the models, they're not gonna be doing any of the Python coding, but these are the people who are making the decisions who have all that have all this sort of smee expertise all this business expertise all this uh foundational knowledge of like i actually know how to plan um you know manufacturing for cars and stuff like that i know all of this i don't know optimization but now we're now like this this this uh you know this group at toyota they did an exceptional job of sort of blending the two and letting people interact with that so So that was one super cool case.

15:01And the other one is with Total Wine. The other one that I could mention is Total Wine. And it's, again, a similar problem of like, how can I, it's a similar problem because it's kind of supply chain-y. But it's, you know, essentially if you think about what a Total Wine store is, it's a massive store that has all the beer and wine that you could ever want. Like anything you're interested in finding and depending on state laws, there's maybe like liquors and stuff like that, too. And here I thought it was a platform for getting complete complaints. I love it. But they what I really liked about their story is sort of the think about like the complexity of decision making that can happen within within something that it's like, OK, well, I buy a bunch of beer, I buy a bunch of wine.

15:52OK, but you're sort of thinking about about again, about complexities and in the presentation, the presenter is talking about like, OK, I want to buy just like one brand of beer or something. The choices that you have in just that single sort of brand is pretty massive. Like, am I buying massive cases? Am I buying individual six packs? How many am I buying, you know, you know, 20 cases of 24 cases of 18, you know, all that sort of stuff. When when am I getting them? How often are they coming? How often are they arriving? And everything like that. And then I think about that for pretty much every beer that exists in it, particularly in like North America.

16:35Or are you importing them? Every wine is sort of you sort of it's a massive, massive problem and not easy to solve. But what I really liked about this problem is the Toyota folks that I just mentioned and a lot of our customers, they have what we call operations research expertise in-house. Even the Toyota example, the person who presented it did not have the traditional background of our common customer. He is an AI person, but had some mathematical chops to him. And so it was like not super. He took to it a little bit faster than I think some would. But Total Wine folks, they were a team of data scientists.

17:24They were people who did not have a traditional sort of operations research background, industrial engineering. Those are some of the common degree types that people have who have been exposed to linear programming, mixed integer programming, the mathematical optimization things. That's where you typically learn that. These were people who were, you know, I'm a data scientist, been doing that for a decade now. Oh, we have this new problem type that we're trying to solve. machine learning is not cutting it. What else can we do? Oh, okay. I've learned of mathematical optimization. Now we need to actually do it.

18:00And so it was a total success story of, of taking a team of people who did not really know how to do this right away. Understanding their learning, their pain points and stuff like that, understanding what worked for them and what didn't. It was just a great story to hear that, that this stuff, you know, that if you're, if you're listening to this now and you're like, oh, well, you know, I'm, you know, I don't have time to listen. I don't have time to learn all of this or, or I don't know the benefits of should I just hire or something like that? You know, that's also complicated. It could be time consuming and blah, blah, blah.

18:34You can be done in house. You can build a team that can take care of this, that can do this at the scale. And, and I think this is where a company like Garobi, this is why I love working for the company I work for, is we don't just like hand you the software and say, good luck, have fun. You know, as long as your check clears, you know, blah, blah, blah, you know, we're not going to talk with you. We have an exceptional sort of support team that helps you with this. So if you get stuck, you know, not stuck with like, hey, I don't know how to build my model stuff but like hey this is taking a lot longer than i thought to run or or we're getting like these these um sort of error messages or or we have issues with this or that you have people you said you know when you submit like a a ticket with us you have someone with a phd in optimization or decades of experience that looks at that and thinks here's how i can help you Um, so, um, so they leveraged that and they used sort of our, you know, they used our, our, our support system to really sort of help them.

19:52And, and now they're saving, uh, I mean, I, I don't want to mischaracterize the number, but it's a lot of money. Um, and they're being able to reinvest it then. And that's, that's, what's really great about these projects. These optimization projects is, is, you know, yeah, you're saving money typically, but it gives you an opportunity to reinvest and make things better elsewhere. So those are a couple of really cool customer stories that I was able to hear. And there's tons more though, tons, tons more.

20:21Jon Krohn:Saving money is of course, one of AI's foremost benefits to corporations that want to improve their margins. But anyone who uses AI must also stay aware of how they are using systems, these systems and tools. As technologists, it's so important to keep asking ourselves, how can we use AI to build a better, fairer, more equitable world? On the podcast, we frequently cover how AI comes with ethical risks that have to be weighed against its promising productivity and efficiency gains. In episode number 935, researcher, broadcaster, and author of the best-selling book, Technology Is Not Neutral, Dr.

20:55Jon Krohn:Stephanie Hare, joined us to discuss her thoughts on how we can install ethical boundaries for our AI use. On the note of developing your book and coming up with these ideas of how technology ethics are treated, not just in the West, but all around the world, something that you've brought up a number of times is the idea of whether we should have something like the Hippocratic oath that they have in medicine for technology. And so it doesn't seem like that's, I don't know, it doesn't seem like it's probably a practical thing that we're going to have like an international technology Hippocratic oath come about.

21:35Jon Krohn:It's a nice idea. But so maybe instead of a symbolic oath, are there practical, non-negotiable checkpoints that maybe should be embedded into tech product development life cycles? Or, you know, yeah, there's some kind of tool set like a Swiss army knife that technologists could work with that maybe is enforced in some way and isn't considered to be a luxury. I think you've hit on the rub of it, which is the enforcement question. The reason I liked the Hippocratic Oath, by the way, is not because it's like a mandatory thing. Like not even all medical schools around the world require that now, and it hasn't always been required for doctors.

22:18and it was actually recreated or rebooted, if you will, after the Second World War. Because of course, as we all know, the Nuremberg trials, after the Second World War, there was a special doctor's trial because doctors were actually very instrumental in the Nazi regime's murder of many citizens of several European countries. And they had a special trial for that. And so that led to a sort of reckoning and a crisis within the medical community after the war, which was like, how is it that a bunch of people who are supposedly trained to help keep people alive and indeed healthy and thriving, how on earth were they among the first instruments of murder in a tyrannical regime?

23:02And I was really fascinated by that because my second area of study was history and specifically World War II history. So I was like, Jesus. And they revisited the training of doctors because of what happened in World War II, right? That reboot came as a response to an acknowledged, universally discussed around the world problem of horror. And I was fascinated by that because of the way that we think about trust. Doctors tend to be quite trusted, you know, put a stethoscope and a white coat on them. And you're like, Oh, you know, you'll do what they say. It's very difficult for a lot of people to push back against a doctor.

23:42They have more training than us, et cetera. And often when you approach a doctor, it's because you're unwell, you're injured, you're sick, or your family member is. So you need to know you can trust them. So I was thinking about those sorts of concepts, you know, the historical reality of trusted, intelligent people betraying that trust in the worst possible way that they possibly could. How do you then come back from that? How do you restore trust to a profession? Why do some medical schools do something like a Hippocratic Oath and some don't? The fact, by the way, that the original Hippocratic Oath versus what's said today is largely rewritten.

24:22Jon Krohn:So what was... They don't do it in Greek? No. A lot of them have rewritten it. And I kind of like that. It's basically just, you know, the first one is, first, do no harm, which I think is totally appropriate for a technologists to embrace as well. And then second, which is the mission statement in my book, is like, how do I maximize the benefits and minimize the harms, which I personally think is a bit more realistic or utilitarian way of thinking about it, which is there's going to be some harm. You cannot make the omelet without breaking some eggs. So fine, choose it. Choose it mindfully, build it in, have a discussion.

24:58It could be democratic. We should all be thinking about this. That implies that people have to be around the table. There's knowledge, there's consent, blah, blah, blah, all this stuff. So that was the only reason I was thinking about it. And the reason I liked it for the medical establishment and thought it might be useful for technologists is precisely because it isn't enforceable. It's not about getting a driver's license. You are not allowed to drive your car unless you have a driver's license and insurance. And if you don't have those things, you could get arrested, sued, et cetera. This is part of joining this community.

25:32It's an ethos. And it's a sign, I would hope, in the best engineering schools, the best business schools, et cetera, that we teach ethics. And indeed, that is actually true in lots of professions. So lawyers have this, accountants have this, civil servants have it. Here in the UK, the civil service ethics code is really serious. I have several friends who are several servants here and I really admire them. Their sense of commitment to something larger than themselves is part of their professional training. So I think it would be lovely, this is just my own take on it, for technologists to have that in their formation and for them to think about it a lot.

26:14If we treated our careers as a vocation, why do you get out of bed in the morning? What are you building? That would be something that I think could help not just with how we design and live and create, but also for our relationship with everybody else, you know, the users of our products, our customers, but we're also our family, our friends, et cetera. So it's just an articulation of the value statement, but I don't think we need to add more regulation to it in the sense of, you know, you can't code unless you've done this thing or you can't create something if you, unless you've got a, you know, the world does not need that.

26:53But, you know, nothing, you don't have to be regulated to do the right thing. You could just decide to, you know, not be an asshole. Yeah.

27:00Jon Krohn:It's kind of this idea, even when you said, you know, the first line, I guess, of a typical Hippocratic oath of the first do no harm. It's interesting how with technology, often the primary incentive is first make a profit. it's like our first generate arr well is it though like i would say that's for like companies that's for a lot of people sure but like a lot of people are out just tinkering right or like necessity is the mother of all invention you know the person who invented the washing machine you know or what i'm just looking around it's you know now i'm like everything in my house toilet gina tool yeah you're usually doing it to solve a problem right where you're like god damn I cannot take this anymore.

27:47I want scissors for left-handed people. I know the world is mainly right-handed, but there's a whole crew of people who are not being served and they can't scissor things without hurting their hands. I shall invent it. I think it's often, hopefully, coming from that. Yes, there are people who always start with the profit motive first, good for them. But I think a lot of innovators are more, they're problem solvers. And then they're like, oh man, if I did this, I can make bank. Why not? There's nothing wrong with that. But I think the best stuff comes from solving problems.