933: Future-Proofing Your Career in the AI Era, feat. Sheamus McGovern

28:09Jon Krohn:The History of the Universe? 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. I like that.

28:39Jon Krohn:I hope we turn what you just said into one of our animated shorts for this episode. Yeah, yeah. Because that's a great visual. this idea of it that's you know with the universe 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 in the industry three decades in tech, right?

29:19In tech, whatever. Skills never go away. There's just more of them, right? And that's why it's just amazing when people get so concerned, well, this is no longer being needed. Like 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, you know, 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.

30:03Because back in 2015, we were using data science, machine learning, and even AI as a tool, right? it's moving away from AI being a tool to AI as being a collaborative partner. And I do think those collaboration skills you build now will help you over the next decade, because that kind of helps you with it. Can I build something now that can future-proof? So I say, you want to wait or you want to build? 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.

30:54You 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. 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. 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?

31:31In 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 account executive and account manager, different roles, by the way. You have a sales engineer, customer success person, 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. and let's say you were just doing STR 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 and you know that's kind of 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.

32:33And 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 Open Devon, 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. and I really started to research that. He's right. If you think about the average person in office, they're looking at emails, they're doing admin. It's mostly judge work and there's this whole productivity paradox and automation paradox. Even though we got productivity with automation, the problem with the automation paradox is automation still needs oversight.

33:23It still needs judgment. 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.

33:50Jon Krohn:I like that. And that does indeed, you're going to love Sadie St. Lawrence's first book when it comes out. It's got AI orchestration right there in the title. Oh, wow. Yeah, yeah, exactly. She always has great content. She does. I don't know how she does it all. She's unbelievable. So that's great. I love all the perspective that you've provided on what we can be doing as practitioners in terms of rewiring our skill sets and making use of this AI moment, this automation augmentation moment that AI is providing, and using that to make ourselves more valuable, get rid of some of the drudgery of the work that we might be doing.

34:34Jon Krohn:What about the other side of the coin, hiring? So I think there's probably been a shift a bit recently from degree-based hiring, more to capability-based hiring and AI-related roles. Do you want to make some input on that? And then just tell us, yeah, what you're seeing from a hiring manager's perspective. What should they be doing in this moment to make the most of it? Yeah, I think, yeah, I want to choose my words carefully here. because look, going back to the founding of what do you see, right? As I said at the very start, I always believe that anyone could be a data scientist, right? With the tools, the code, the will.

35:17As I said before, even to this day, we look at the profiles of people showing up at the conferences. I go to other conferences and I ask, like, what's your background in? I have people on my show. I ask them, how did you get started in data science? You know how many people tell me they started as a data scientist or an engineer? Very, very few, right? Or a PhD in data science or something like that. So I've always believed in the skills-based hiring right out of the bat. And you've also seen a shift, and this predates large language models and the AI moment, if we can call it that. Like I remember a couple of years ago, I couldn't find my degree from Northeastern when I went for my job at Fidelity.

35:58I was freaking out. And I finally had to go and get it. And I was, I was, um, very nervous there, but I was happy to see that fidelity finally stopped requiring, I think it was a good few years ago, four or five years ago, stopped requiring a degree and, um, IBM and Google and even meta, I think have gone, um, to, to do that.

36:17Jon Krohn:So were you saying, are you saying in that story there that, so you, you lost your diploma and you had to go, so you had to like reorder a diploma and then you had to bring, you had to like bring the diploma to your interview. No, no, no. So even funnier, so just a little background. I am from Ireland originally. Wanted to do electric engineering and wanted to work on better computers. Ireland at the time didn't have very good ones. So I went to Northeastern here in Boston, very good engineering school. And I went there and my last class was in April. And back then in the early 90s, you didn't get a year visa to work around.

36:53So I actually, because I was pretty much broke after school, I went back to London to work in London um and I never went to my graduation in Northeastern so I never actually got my diploma and they were going to mail it to you so um I'm in I'm I'm in London I'm working away um I'm applying for grad school and lo and behold I got a green card in the mail I won't bore you with the details I came back here and then um someone reached out to me from Fidelity and like I didn't want to come work here I passed interview process and everything it's like and then right before I was supposed to start my job they're like um oh yeah one small thing you send you the degree we need to see that degree and that was back in the days before you gotta look it up and i was like what degree yeah i got a degree can i just call northeastern like no they won't let us into their records so i had to go down to northeastern um i will be honest i had a little bill outstanding a couple of grand out the day as well maybe you should have that out No, that's not coming up.

37:54Jon Krohn:That's so funny. Am I going red, John? Am I blushing? I was wondering if you were sunburned before. No, definitely. Yeah. So anyway, so back on track to the credentials,

38:09like you, I learned so much from my podcast as well and from speakers at ODSC, instructors like yourself. Like we had Daniel Rock and Sam Manning on and they did some great papers about, Like when people talk about AI exposed jobs, I take that with a little bit of a grain of salt because I kind of look at the jobs and they broke down the tasks and see what tasks are very repeatable. But, you know, and a lot of experts talk about this credential collapse, saying that AI exposed or AI augmented jobs. If you're good at AI, you no longer need credentials. And, uh, look, the way I look at it is because actually, um, in my work for a VC firm, I'm not.

38:50Yeah.

38:50Jon Krohn:Let's, let's talk about that really quickly, just so that people have a context because so in addition to ODSC, all the work you do with the open data science conference, you're also, you're the founder of cortical ventures. Is that right? No, absolutely not. Okay. They found me. Um, I was on the founding team, I guess I would say, but I would say, uh, Jeremy Arshon, from Data Robot fame and Igor Taber. He was at Intel and their startup team there. Two pretty big deals. When they were thinking of starting a VC firm, they reached out to me. So I wouldn't say I'm a founder per se because I'm a venture partner, which is a fancy way of saying you're part-time.

39:28So I do help find deals. So if you've got a startup ID there, shameless help plug, send me your deck. We'd love to see it. Connect me on LinkedIn. in. But what really got me involved is they promised me this role of head of AI. So I have a small AI team there. So this allows me to build out AI. So I've been drinking the Kool-Aid, Kool-Aid, eating the dog food. So we started about almost four years ago. And we said when we started Cortico Ventures, we will use AI to find the next generation of and support the next generation of founders. So we've been using tools everywhere from kind of like using some of the signals I used to use in the hedge fund world.

40:09Signals then we were using agents. Now we're using agentic workflows. But even in cortical, like when we look at startups, look at founders, degrees still matter. But what's shifted now, it's like it's one signal only, right? You know what I mean? And you asked about employers, what are they looking for? But even on the people looking for job site, what they should be looking at, it's like they want to see your skills, but they also want to see proof, right? And John, you and I have been preaching this for years. what's on your GitHub? What's on your LinkedIn? Are you posting about stuff? If you learn something, write a blog about it.

40:44Are you speaking at a meetup? Are you volunteering at events? So you want to show, you want to build a project portfolio and show proof of skills. And to hiring managers who are new to this, yeah, degrees still matter, I think. But look for people, what they're doing in the community, how they're contributing, where they're contributing. look for portfolios um and look for excitement and look for people who are excuse me eager to learn who are coachable flexible because that's what it is because when the whole industry is in flux when you got to rewire your skills um you have to maintain flexibility and you got to be very coachable right and look at this um crazy talent board that's going on between open AI and meta, right?

41:32$100 million bonuses and offers. And yeah, and one thing as well, I say to people like going back to the speed of change, just get used to the fact that AI roles and the required skills will always emerge much faster than traditional degree programs. And I think this recognition of skills is long overdue. and to be quite blunt, I think universities are doing graduates a disservice in how they train computer science and data science degrees at the moment. So yeah.

42:11Jon Krohn:Yeah. I was recently reading that, so Noam Brown, who's now a pretty well-known researcher at OpenAI, and he was on this podcast some years ago.

42:24Noam wrote recently on LinkedIn about how a lot of university AI programs,

42:32Jon Krohn:Introduction to AI, it has nothing to do with the kinds of skills. It has nothing to do with large language models, transformers. it's about, um, you know, kind of decades old ideas around, you know, hill climbing strategies, uh, you know, things that are completely out of date, but that, you know, the curriculum moves so much slower than universities can adapt. And, uh, yeah, I mean, I, I don't have my own kids yet. Uh, and so it'll be a while before the kids that I do have are, you know, thinking about going to college, but I could easily imagine, even if I had kids of college going age today, I'd be saying, you know, think about whether all that investment, especially in the U S it's so expensive in a lot of cases, a lot of the rest of the world that isn't, there's all kinds of social things that I think are great about university, but in terms of employable skillset, it isn't necessarily the best, uh, the best use of capital.

43:30Yeah. A hundred percent agree. Um, yeah, I really agree there because I don't want to be hypocritical. My son is actually going to college and I think it's a great experience. And I understand, I know why universities are in a bind, right? You got to teach the fundamentals. You can't be teaching the latest hottest thing and there's another latest hottest thing. So you have to teach the foundations. Like when I was doing computer science, I learned about databases and operating systems and some of that stuff is still true. But look, the way the universities are teaching it right now, when Chattapit came out, they banned it.

44:03Instead of trying to control it and have people help them learn with it, they banned it. As you know, I also teach a boot camp. We have a boot camp with ODSC to try and get people ready for the firehose of learning they're going to experience at ODSC West. So we run a boot camp every year or before every conference. And I've moved towards incorporating AI into my learning and just telling them. We did a podcast with Natalia from MIT Media Lab here. Natalia, I won't say her last name because of the budget, but she wrote the viral paper. She was one of the core writers of your brain in ChatGPT. So if you're using ChatGPT all the time, using LLMs to do your work, you're going to lose your skill set, right?

44:46Part of rewinding your skill set is not to be overusing ChatGPT, but universities need to get on board with letting students use AI tools, be comfortable with them because look, it's a complete disservice. I'm sure you've read the report where computer science graduates' employment rate is way up, right? More so than the norm. I think it's something like 7 % or 8 % versus the other one is 3 % or 4%.

45:11Jon Krohn:Computer science grads' unemployment rates are way up? Unemployment rates, yeah, are way up. Yeah. Now, I read that report a month ago, but like everything, people are sticking an AI label on and blaming the AI for it. I think that may have a lot to do with, because look, most of the computer science hiring in the us is big tech hiring so it all depends on what apple meta google um and a few others i'm forgetting are uh are hiring right so i think it's a bit to do with that because there's other trends that support that so everyone's saying that um they're they're blaming the fact that they're not trained in ai further unemployment rate and also the fact that um you know uh these folks don't have ai skills.

45:57They have skills that are easily replaceable with AI, like the junior skills are easily replaced with AI. And I think universities need to pay attention to that because I'm mid-level career now, mid-level management, I suppose. So I'm not trying to diss anyone and scare anybody. But if I was at a company, I would much rather hire a young junior computer science graduate who was excited about AI, ready to learn, learn, learn, constantly rewire their skills. Then someone's like, eh, yeah, vibe coding is just a trend. It'll go away. I don't need to learn this stuff or I'll use it a little bit. Yeah.

46:39So I think there's a lot of work universities need to do there. And I think they'll catch up. I think they will cop on. I know Northeastern here is doing a good job on that. BU, I talked to some folks there, BC. So I know universities know this, they're just trying to figure it out. But yeah, so I think the whole, I think the shift to skills-based hiring is real. And yeah, I think it's real.

47:13Jon Krohn:All right. So we've talked now a lot about individual skills. What about teams? So what do you see, you know, in your role as what's the title there again, part venture partner, venture partner at cortical ventures at the VC firm. What do you see in successful AI startups? How do they get their team skillset? Right? Yeah. Yeah. I didn't mention before, but I'm really grateful for, um, Jeremy and Igor and Mike and, um, give me a window into this world. And, uh, it's a fascinating world. Um, what a fascinating world. and it's a very privileged position because I'm definitely not on all of the calls because of my other work and my other roles.

47:54But I do have a lot of conversations with some extremely bright, fascinating people building some amazing, amazing things. And I wish Cortical could have funded them all. But the power law in VC is very strong. So don't hold me to that. not all my decision so that crap yet yeah so there's this whole um you know come on everything starts in the valley and kind of spreads out right there's this whole um let me let me roll back here um because startups are an interesting beast for people who have never been involved in but they're like they're like time machines right so an ai startup is like a time machines an AI startup is like a time machine.

48:43The skills that they're betting on today, that often becomes the industry standard tomorrow, right? And we've seen this time and time and time again. You talked about, we just talked about skills-based. I've never met a startup. they don't look at static resumes. They always want to hire people who are flexible, who can rewire their skills almost in real time. Flexible, can have multiple roles, and are coachable. And that's been around for a while. And now because of AI, the new thing they're doing is this concept of tiny teams. And tiny teams is a fascinating concept. And the idea is you build a small, flexible, highly capable team, and you want to emphasize this hybrid skill set and this deep collaboration.

49:42And the reason you can do that is AI is either going to augment or automate a lot of the routine, not only the routine tasks, but a lot of the expert tasks, right? So, you know, you do a startup, you used to have to hire a legal counsel, you do all your whatever. now you just use chat tpt or a legal platform um you can use ai assisted coding um you can build your mvp the front end demo used to go and bring to vcs you can really literally build that and build that in the day thanks to uh replet and lovable and all that stuff so what you end up seeing is um founders prioritize quality hires over um quality hires with you know domain in expertise, hybrid roles.

50:29And another thing they're doing, which I don't think is healthy, is they're really focusing on revenue per employee. Not sure if it's relevant to this discussion, but you see startups like, I think I told you earlier, I think the number's right, Lovable, the AI says a coding platform, 10 million in revenue in two months with about 15 employees. and I think a better one is not as impressive, but Cursor. No, wait. Cursor went to 100 million with 45 people. I don't know what the timeline was. 100 million in revenue? And all it was was, okay, I'm not, I shouldn't be dismissive, but it's basically a wrapper around an LLM, right?

51:12But to their credit, they've done some amazing work training that. So the key thing which you see in startups here that's relevant to our conversation is they really work a lot to build out the AI tooling. It's kind of like when I said we started Cortical. That's why I was brought on to Cortical for it was, can we use AI to build an AI VC firm? And this is why I pay so much attention to this, because this is what we're trying to do in the team at Cortical. Using AI tooling that augments or automates away and many of those drudge tasks we talked about And, um, you, you know, before really required a lot of efforts.

51:50Um, you know, so, um, and I do think startups, as I mentioned before, are kind of a window into the future of work. And, um, yeah, so, so very, very interesting. Um, yeah, I'm not, did that kind of answer your question? It did, Seamus.

52:07Jon Krohn:Yeah. Sorry. Sometimes I speak for so long, I kind of lose, lose the start. you have you have a unique way of speaking you did bring it back you have brought it back on every question but you uh your you know you you have such extensive knowledge and we kind of we have this journey hopping with you uh you know away from my question and then back my apologies no it's all great politician and uh yeah so going beyond so we've talked a lot in this episode about the skills and how they're changing, how they've changed over the past 10 years. So, you know, going back to the beginning of the episode, we talked about the Open Data Science Conference, how you founded it.

52:49Jon Krohn:And we kind of followed this journey of how the data science and then AI skillset changed over those 10 years. We've now spent a lot of the episode, you know, the past half hour or so talking about rewiring professional skills, the kinds of things that people should be looking for when hiring, the kinds of teams that people should be putting together by combining all these skills together. And then, yeah, most recently, how to make a successful AI startup, the kinds of people that are employed in those organizations, the kinds of skill sets that are involved. And so now I want to take a look to the future.

53:25And so there's a lot

53:28Jon Krohn:of concern amongst developers themselves, anybody who's writing code, anybody in data science, anybody working on AI, that AI itself could put us all out of a job, that artificial general intelligence or artificial super intelligence could make a lot of human work obsolete. And I think there's maybe, when people think about that as a concern, things that involve physical manipulation of the world, I think people might feel like that's a little bit safer for a little bit longer because robotics is hard and it takes time to be able to, you know, physically manipulate the real world. So all kinds of things like, you know, physical therapist, massage therapist, uh, you know, construction work, there's all these kinds of jobs that, uh, you know, seem relatively safe because they involve physical manipulation of the world.

54:25Jon Krohn:But if you're in a, in a purely cognitive kind of role, which I think up until the past few years, it kind of felt like that's where you were safer. You know, really go to university, get that graduate degree, really develop your specialization because the machines, they're good at the physical stuff. You know, they're getting better at the physical stuff. Got robots and factories, but cognitive work, that's safe. And now all of a sudden, you know, since the chat GPT moment that we talked about earlier on this episode, you're like, wow, Now machines are carving out more and more and more of a broader cut out of the possible cognitive tasks that we can do, doing it at a superhuman capability.

55:11Jon Krohn:And how much does that concern you? How does that, you know, how fearful are you of, you know, the cognitive ability of machines? And yeah, how can people, is there any way that we can prepare as individuals or organizations for that kind of future? Yeah. You know, again, through my role at Cortical and most definitely my role at ODSC and then for going to these meetups and talking to people. I do think about this to the best of my ability. So there's a lot of, a lot of egg being spilled on this. A lot of smart people talking about this, but also I got to say a lot of nonsense as well. And maybe I'm, I'm speaking nonsense as well.

55:59So I don't mean to be, speak ill of anybody. But when I do go to meetups and conferences and people are generally concerned. It makes me a little bit pissed off, to say the least, because there's so much angst out there and a lot of clickbait headlines. So part of my role, like your role, is to educate people in data science. And this is why people should listen to this podcast, right? So yeah, preparing for the future, I try and figure that out myself because how do you deal with anxiety? How do you deal with worry? You try, like most things, it's a big problem. You break it down into smaller problems, right?

56:44And I do believe, without having to repeat myself and the stuff we already talked about, I do believe there is a new kind of skills hierarchy out there. Whether you're in an AI data science role or a non-technical role, I do think that's important. And I don't know, I'm trying to flesh this out in my head because people ask me this all the time and I want to have a better answer. But I do think number one from the teaching I've done myself is think about the first step is or in the hierarchy or whatever that is. Look, at this stage, ChatGPT, basic prompting, using Gen.AI, it's all table stakes, right?

57:32that's the price of entry everyone should be doing that um and with each of these tiers you can kind of look at a base tier and then how do you how do you do uh you know more of a an expert tier and so for those was in the role we should think of the expert here as well so so beyond doing that and i can this is why um i loved listening to you uh your podcast with curl uh curul um the former host of uh the super data science podcast when he explained llm to me i finally got it and then i was finally able to teach other people i forget what podcast episode that was but that was one of my i think yeah so you're talking about the the transformers episode

58:11Jon Krohn:that he did yes um yeah i'll you keep talking i'll find the episode number that is one of our most popular episodes of all time but yeah so in that tier one table stakes are basic prompting stuff like that. Becoming more expert will be understanding LLM's limitations, understanding the architecture, hallucination bias, understanding garbage in, garbage out, data-centric AI, data hygiene, cloud basics, and of course, AI awareness. And then the next step up from that, if you want to be an engineer and stuff like that in Future Proof, it's all about keeping current with the tools that are out there, like keeping current with what's going on with OpenAI, Antropic, Gemini, Google, building those systems, prompt engineering.

58:57And take, for example, prompt engineering. That's the second level. I actually teach a class on that as well. But now we've got prompt programming and prompt templates, which is a part of that as well. And that next tier is about learning to customize model workflows and definitely starting to work with domains. and of course, efficient engineering. So that's very important. But like Sadie, I think the next tier after that, because I'm kind of going on levels here, the next tier after that is definitely orchestration. So I can't wait to read her book because it's always been over decades, like new skills, new tools.

59:38It's always about, as you said at the very start, connecting those tools, wrapping them, integrating them into product workflow, putting them into production, and then managing latency and cost. You know that the paradox we talked about earlier, not the productivity one, but anyway, I'll come back to it. But yeah, so when new things come out, I think it's called the automation paradox. Whenever something is automated, it's never in the ether. There still has to be oversight and judgment applied to that. So orchestration integration is very important. And then the expert level of that is, you know, these multi-agent systems, tool chains, pipelining, and then continuous learning.

1:00:25And then after that, it comes into management. And I'm a big believer in this, you know, you really need to focus on those human-centered skills, right? Like when you're not stuck coding all day or AI is augmenting or automation a lot of your work, That communication, that collaboration, being adaptable, domain knowledge, whereas fintech, healthcare, finance, startups, health law becomes so important. And then, you know, building leadership around that. What's your AI strategy? What's your AI strategy? Understanding ethical judgment, what to do there. Safety, auditing, compliance. I know a lot of people now going for AI risk compliance certificates, which I think is quite smart.

1:01:17And then I think the last thing to really future-proof is you've got to learn to learn. You've got to stay very curious about this stuff. Stay in the slipstream. You've got to get comfortable navigating ambiguity. And I think a big part of that, if you want to become expert at it, is developing foresight is really important in this field now. Anticipating what AI will enable, what are the next requirements going to be, system thinking, creativity, and understanding those responsibilities. Because to be honest, it's not that hard. That's what I've been doing most of my career. That's why I kept on shifting.

1:01:58That's why I got into data science earlier. And that's why I got into AI early. that's kind of what I got into VC because I wanted to kind of have a window into what's next and I got to say as well even when I started ODSC and they met us back in 2013 then ODSC in 2015 it's all about building your brand long gone are the days when you depend on one employer. John I got to say you were the master of building a brand right people can take a lesson out of your book build your own brand, do your work write blog post learn be curious you know um skills in the right now they've got a they've got a shelf life or a half life of much less than what they were like you know i've i spent decades honing my coding skills and you know it's it's it's like um loss aversion right you don't let go those skills but you gotta you gotta sunset some skills and um now i focus on being the best prompt engineer I can be, unless about coding.

1:03:03And when I teach my machine learning class and my coding class, I teach them the fundamentals of it, but I'm not like, okay, now it's going to be about how to prompt this. So yeah, you got to sunset because there's only so much time in the day. So you got to sunset some skills and you have to be comfortable with learning new skills continuously.

1:03:24Jon Krohn:Nice. Well, that's a great message to finish off on here, Seamus. Before I let you go though i need a book recommendation from you as you know yes i do because i i listen to your podcast um so i'm gonna do a couple of book recommendations is that allowed i'll allow it i'll allow it or you can always edit out the second one but um no no you can have to the one i've been preaching to the choir when i talk to people is um like most of my answers is going to to be a long one. I'm going to try and keep it short. But like most of my answers, so think about the models when they first came out. There was one type of model.

1:04:01And then reasoning models when they came out last year were a very big deal. And then instant models like, you know, system one thinking, system two thinking. Well, a book I read about many years ago, it was very influential to me how I started to think about data science and then machine learning and then AI. It's never let me down. And that's Daniel Kahneman, Thinking Fast and Slow. Have you ever read it?

1:04:26Jon Krohn:I knew it. Yeah, I love Thinking Fast and Slow. I've talked about it on the show a lot. And usually in the exact same context that you just brought it up, which is this idea of system one thinking being your fully automatic, fast thinking. And that is like what GPT 3.5 was like, or GPT 4, where you get the, you know, it's like stream of consciousness, just being spit out in real time, like the words coming out of my mouth right now, the tokens that I'm spewing. Uh, whereas, uh, thinking slow is, you know, careful, logical thought, breaking things down step-by-step, maybe getting out a pencil, a whiteboard marker.

1:05:05Jon Krohn:And, uh, that's what we now see with reasoning models. Yeah. Another one, um, because I interviewed him from my, um, he was a guest on our show. Um, and that was, uh, you think I'd be able to pronounce the name correctly, Arvind Narayanan, and he wrote the book AI Snake Oil. So a good book to read. And, you know, and not necessarily a skeptic around AGI and stuff like that. And then you asked me about AGI and I think I never answered it, but I don't know. I think we will get to AI eventually, but not in the timelines people are talking about and calling what we have today AGI is like calling a skateboard, a self-driving car because it has wheels.

1:05:54I think I heard that somewhere. Kind of funny. So AI Snake Oil is a good book. And then for my AI detox, as I said, History of the Universe, great series. And then one of my favorite all-time books is now a Netflix series, but The Three-Body Problem. And then I think, what was it? the um dark forest and i forget the name of the third metrology but there's my recommendations

1:06:19Jon Krohn:yeah third body problem that's gotta be one i've gotta read that you know i've been looking for a good novel to pick up and uh lots of people have recommended that you gotta it's time yeah uh and then ai snake oh you mentioned that at ods he used to me in person he's someone i've got to get on the show i just haven't had yeah i haven't had space but i'd love to get him on it's great and then i I did look up while you were speaking that episode, that introduction to LLMs and Transformers episode that we did with Kirill Arimenko. How could I forget? It's episode 747. That's right. 747. Exactly. Exactly.

1:06:52Jon Krohn:It's flying. All right. Seamus, for people who want to get more of your brilliant thoughts, obviously they can make their way to ODSC West at the end of October, basically every year, Halloween time. uh this year it is the exact dates you must have 28 to 30th 28th to 30th there you go yeah yeah and uh yeah so catch you at odsc west or if you're on the east coast uh odsc east in the spring i'll probably see you there at either of these i definitely want to see you there john and um yeah yeah yeah for sure beyond the conferences where else should people catch you how can people follow you after the episode, Seamus.

1:07:38Yeah. I love this podcast, but if they want to listen to another, if they've got room, our podcast is where I kind of speak my mind and they're definitely not brilliant thoughts as you well know, listener from listening to me for the last however many minutes it was, but I do my best. I do post on LinkedIn and yeah, but please come meet me at a meetup near you or come to odsc west and come to odsc east i i don't have all the answers but i love hearing the questions nice love it seamus such a treat to have you on the show

1:08:15Jon Krohn:my apologies that it took so long to get you on and hopefully it won't be too long before you're back uh but you know the twist is going to be that in two years when you're back agi is going to be hosting my show. You're so welcome to be a guest. I don't believe it. I'll make a bet on that. I just, that's satire. I'll put money on that. Yeah. I did make a few bets on that already with people, but John, look, this has been an amazing experience. First of all, Super Data Science, I always tell people, this is the podcast people have to listen to. I've learned so much from it over the years. And I've learned so much from your talks, the talk you gave at Ed Donner at odsc um west it's on it's on youtube um love that and so many people have said to me that was uh have listened to on youtube and loved loved loved it and um yeah again to people um like yourself who come to odsc our instructors you know we owe you a debt of gratitude and can't thank you enough and um and i should not be on this podcast because i love all the guests you have so i'd much rather from all the other brilliant people you keep on having on there.

1:09:26Jon Krohn:And thanks for mentioning that Agentec AI Engineering YouTube video that Ed Donner and I did that's recorded at the Open Data Science Conference. And for people out there who have listened to that or have watched that YouTube video, it's this very Seamus on this podcast today that introduces me and kicks off that video. So thanks for the enthusiasm there, Seamus. And yeah, catch you at ODSC soon. Catch you at ODSC, John. We'll see you there.

1:10:21Jon Krohn:the pace of technological change outstrips people's ability to absorb it, the tiny teams phenomenon in AI startups, and practical strategies for future-proofing careers through continuous learning, building personal brands, and developing orchestration skills over pure technical coding. As always, you can get all those show notes, including the transcript for this episode, the video recording, any materials mentioned on the show, the URLs for Seamus' social media profiles, as well as my own at superdatascience.com slash 933. Thanks, of course, to everyone on the Super Data Science podcast team, our podcast manager, Sonja Breivich, media editor, Mario Pombo, partnerships manager, Natalie Zajski, our researcher, Serge Macisse, writer, Dr.

1:11:02Jon Krohn:Zara Karche, and our founder, Kirill Arimenko. Thanks to that whole team for producing another awesome episode for us today for enabling that super team to create this free podcast for you. We are so grateful to our sponsors. They literally keep the show going. And then you can help us out by checking out our sponsors links, which are in the show notes. And if you're ever interested in sponsoring an episode yourself, you can find out how at johnkrone.com slash podcast. Otherwise, share, review, subscribe, edit videos into shorts to your heart's content. But most importantly, just keep on tuning in.

1:11:39Jon Krohn:I'm so grateful to have you listening. And I hope I can continue to make episodes you love for years and years to come. Until next time, keep on rocking it out there. And I'm looking forward to enjoying another round of the Super Data Science Podcast with you very soon.