Fences, flagpoles, and the comeback of the generalist

00:00:00,100 --> 00:00:24,620 [Eric] [upbeat music] All right. Welcome back to Token Intelligence. John, something that's been on my mind is the way that AI and the use of AI, let's just say in knowledge work, since that's the, the realm that we operate in- 00:00:24,620 --> 00:00:25,360 [John] Right 00:00:25,360 --> 00:00:25,720 [Eric] ... uh, 00:00:27,160 --> 00:00:33,240 [Eric] is shifting the types of employees and skillsets that are valuable. 00:00:34,600 --> 00:00:46,760 [Eric] There are lots of things to talk about here. I think one of the major shifts that we've seen is the role of a software engineer has, has dramatically changed, right? 00:00:46,760 --> 00:00:46,840 [John] Right. 00:00:46,840 --> 00:00:57,400 [Eric] So, uh, a significant move from handwriting code to reviewing code generated by an AI agent. 00:00:57,400 --> 00:00:57,530 [John] Right. 00:00:57,530 --> 00:01:00,200 [Eric] Right? And it really changes the, the software- 00:01:00,200 --> 00:01:00,480 [John] If, if that 00:01:00,480 --> 00:01:01,140 [Eric] ... development lifecycle. [laughs] 00:01:01,140 --> 00:01:03,380 [John] Depe- depending on the company, right? 00:01:03,380 --> 00:01:04,500 [Eric] Yeah, depending on the company. 00:01:04,500 --> 00:01:04,730 [John] Um, right. 00:01:04,730 --> 00:01:08,180 [Eric] And, and it is important, I think, to, to acknowledge that 00:01:09,660 --> 00:01:17,960 [Eric] you and I work on the bleeding edge of this happening, and there are a lot of companies that are still trying to figure out how to adopt this, right? 00:01:17,960 --> 00:01:18,539 [John] Yeah. 00:01:18,540 --> 00:01:29,980 [Eric] Um, which I think it is a good thing, right? 'Cause it's not necessarily straightforward and it's changing so quickly that investing a lot of time into it can ultimately be wasted depending on, you know, what you're trying to do. 00:01:29,980 --> 00:01:31,560 [John] For sure. [laughs] 00:01:31,560 --> 00:01:48,420 [Eric] But one of the s- one of the, the, um, one of the types of employees or sort of skillsets that I think is really important and becoming even more valuable is a generalist. 00:01:49,560 --> 00:01:50,480 [John] Yes. 00:01:50,480 --> 00:01:56,330 [Eric] So let's define generalist first, though. And we didn't do any prep for this episode. 00:01:56,330 --> 00:01:56,380 [John] Right. 00:01:56,380 --> 00:01:57,740 [Eric] So give me your definition- 00:01:57,740 --> 00:01:58,870 [John] Generalist 00:01:58,870 --> 00:02:00,980 [Eric] ... of a generalist. [laughs] 00:02:00,980 --> 00:02:09,759 [John] So, uh, I think to start off, a little bit of commentary. It, it's engineering, but it's like that whole, if you're in SaaS, like product engineering design, like that whole org. 00:02:09,759 --> 00:02:09,799 [Eric] Mm. 00:02:09,800 --> 00:02:10,180 [John] Right? 00:02:10,180 --> 00:02:10,190 [Eric] Mm-hmm. 00:02:10,190 --> 00:02:14,580 [John] 'Cause you've got designers that all of a sudden can, like, do front-end stuff, and you've got- 00:02:14,580 --> 00:02:15,140 [Eric] Mm 00:02:15,140 --> 00:02:18,320 [John] ... front-end people that can, like, do ba- Like, it's just the whole- 00:02:18,320 --> 00:02:18,370 [Eric] Yep 00:02:18,370 --> 00:02:19,560 [John] ... org is messy, right? 00:02:19,560 --> 00:02:20,240 [Eric] Yep. 00:02:20,240 --> 00:02:23,500 [John] Um, but yeah. So the generalist, I, 00:02:24,680 --> 00:02:30,240 [John] I mean, it's, it's that broad, it's that broad skillset piece. Like, a, do you remember the T-shaped, like- 00:02:30,240 --> 00:02:30,600 [Eric] Yes 00:02:30,600 --> 00:02:31,640 [John] ... illustration? You know. 00:02:31,640 --> 00:02:31,960 [Eric] Yep. 00:02:31,960 --> 00:02:41,300 [John] Talking about having this, like, broad... So imagine the, like, letter T. So you've got, like, this broad, like, knowledge a- across a lot of disciplines- 00:02:41,300 --> 00:02:41,910 [Eric] Yep 00:02:41,910 --> 00:02:43,510 [John] ... in an organization, and then you're deep in one. 00:02:43,510 --> 00:02:43,540 [Eric] Yep. 00:02:43,540 --> 00:02:56,060 [John] So, like, that would be, you may be deep in engineering, but, like, you, you kind of get at least at a high level marketing and sales, and, like, you, you know enough to, like, be, to be really, really useful because you're deep in the one and broad in the other one. 00:02:56,060 --> 00:02:56,520 [Eric] Yep. 00:02:56,520 --> 00:03:03,440 [John] And people, like, I don't even know. I think people, like, I think the letter X is, like, a new cool, like, you wanna be X-shaped or, like- 00:03:03,440 --> 00:03:03,700 [Eric] Real- 00:03:03,700 --> 00:03:03,760 [John] So- 00:03:03,760 --> 00:03:05,000 [Eric] Wait, wait. I haven't seen this. 00:03:05,000 --> 00:03:05,100 [John] Yeah. [laughs] 00:03:05,100 --> 00:03:06,660 [Eric] Explain X-shaped to me. 00:03:06,660 --> 00:03:12,740 [John] I think the next axis is, like, leadership and people skills or something, and, like, I, to be honest- 00:03:12,740 --> 00:03:12,750 [Eric] Wow 00:03:12,750 --> 00:03:13,630 [John] ... I would have to look it up. 00:03:13,630 --> 00:03:13,640 [Eric] That's- 00:03:13,640 --> 00:03:15,160 [John] And then, and then there's some kind of, like- 00:03:15,160 --> 00:03:15,290 [Eric] [laughs] 00:03:15,290 --> 00:03:18,440 [John] ... three-legged shape that, that I've heard talked about- 00:03:18,440 --> 00:03:18,780 [Eric] Okay 00:03:18,780 --> 00:03:23,720 [John] ... where it's like you need to be deep in, like, three domains and then a generalist here. Like, it's getting a little wild. 00:03:23,720 --> 00:03:24,420 [Eric] Okay. Wow. 00:03:24,420 --> 00:03:28,460 [John] Like, I feel like the T-shaped thing was, uh, was it for a while, and then, like, with AI- 00:03:28,460 --> 00:03:28,630 [Eric] Wow 00:03:28,630 --> 00:03:29,920 [John] ... people are going, like- 00:03:29,920 --> 00:03:30,480 [Eric] It's gotten weird 00:03:30,480 --> 00:03:37,660 [John] ... I think the X thing was, like, prior to AI. I should probably just look this up. But, um, and then there's other shapes that have [laughs] come out with AI. Like you need- 00:03:37,660 --> 00:03:37,800 [Eric] Right 00:03:37,800 --> 00:03:52,880 [John] ... to be deep in three things. Um, but yeah. So to more tightly define the generalist piece, I th- I think what we're saying is if you look at a T-shape, we are emphasizing the top portion of the T more than the, like- 00:03:52,880 --> 00:03:53,280 [Eric] Yes 00:03:53,280 --> 00:03:55,400 [John] ... the, um, vertical portion of the T. 00:03:55,400 --> 00:03:55,820 [Eric] Yeah. Okay. 00:03:55,820 --> 00:03:57,020 [John] Horizontal over vertical- 00:03:57,020 --> 00:03:57,070 [Eric] Yeah 00:03:57,070 --> 00:03:58,940 [John] ... when it comes to the letter T. Um- 00:03:58,940 --> 00:04:01,700 [Eric] I, I really identify as a spiral, by the way- 00:04:01,700 --> 00:04:01,710 [John] [laughs] 00:04:01,710 --> 00:04:03,859 [Eric] ... if we're picking shapes. [laughs] 00:04:03,860 --> 00:04:06,829 [John] Upward spiral, I hope. [laughs] 00:04:06,829 --> 00:04:13,340 [Eric] Upward spiral, not a downward spiral. [laughs] Uh, y- I think that's right, is emphasizing- 00:04:13,340 --> 00:04:13,740 [John] Yeah 00:04:13,740 --> 00:04:20,740 [Eric] ... the, the horizontal, um, part of the T. And I think one of the things that is 00:04:21,920 --> 00:04:31,710 [Eric] interesting is, or, or even I would say probably a, a generalist that is, that creates asymmetric value within an- 00:04:31,710 --> 00:04:31,710 [John] Right 00:04:31,710 --> 00:04:40,340 [Eric] ... organization is actually someone, it's not even necessarily a T, but it's someone who, you know, so it's, like, horizontal- 00:04:40,340 --> 00:04:40,350 [John] Right 00:04:40,350 --> 00:04:42,140 [Eric] ... and then one, like, where it's really deep. 00:04:42,140 --> 00:04:42,440 [John] Yeah. 00:04:42,440 --> 00:04:48,060 [Eric] But actually is, like, a horizontal line with multiple lines down- 00:04:48,060 --> 00:04:48,070 [John] Yeah 00:04:48,070 --> 00:04:51,210 [Eric] ... that don't go as deep as a single line- 00:04:51,210 --> 00:04:51,210 [John] Right 00:04:51,210 --> 00:04:51,720 [Eric] ... on the T. 00:04:51,720 --> 00:04:52,060 [John] Right. 00:04:52,060 --> 00:04:52,150 [Eric] Right? 00:04:52,150 --> 00:04:53,860 [John] And I, and I think that's what, um... 00:04:55,160 --> 00:04:57,880 [John] Again, I don't know what the latest shape is, but that's the concept of what- 00:04:57,880 --> 00:04:58,640 [Eric] It's a fence. 00:05:00,000 --> 00:05:00,740 [John] [laughs] Fence shape. 00:05:00,740 --> 00:05:01,140 [Eric] [laughs] Fence shape. 00:05:01,140 --> 00:05:03,500 [John] I mean, pretty much, yeah. 00:05:03,500 --> 00:05:03,660 [Eric] So- 00:05:04,820 --> 00:05:07,800 [John] Or think about a platform. Like, you've got this, like, top part. 00:05:07,800 --> 00:05:07,900 [Eric] Yeah. 00:05:07,900 --> 00:05:09,880 [John] And you've got these, like, supporting pillars. 00:05:09,880 --> 00:05:10,180 [Eric] Right. 00:05:10,180 --> 00:05:14,800 [John] And we believe that the person is stronger that has more supporting pillars, essentially. 00:05:14,800 --> 00:05:17,300 [Eric] Yes. Oh, wow. That's much more elegant. 00:05:17,300 --> 00:05:18,130 [John] There you go. 00:05:18,130 --> 00:05:18,140 [Eric] Yeah. There you go. 00:05:18,140 --> 00:05:18,170 [John] [laughs] 00:05:18,170 --> 00:05:20,930 [Eric] Well, I, the reason I started thinking about this was 00:05:22,020 --> 00:05:27,680 [Eric] that a lot of times the per- the fence-shaped person 00:05:29,480 --> 00:05:29,950 [Eric] is, 00:05:31,400 --> 00:05:33,560 [Eric] um, their value 00:05:34,960 --> 00:05:39,620 [Eric] is, is very, very high in certain contexts- 00:05:39,620 --> 00:05:39,690 [John] Yeah 00:05:39,690 --> 00:05:42,510 [Eric] ... or in certain phases, right? 00:05:42,510 --> 00:05:42,560 [John] Right. 00:05:42,560 --> 00:05:45,560 [Eric] So the fence-shaped person, I'm just gonna continue to use that. 00:05:45,560 --> 00:05:46,120 [John] Do it. Yeah. 00:05:46,120 --> 00:05:46,380 [Eric] Um, 00:05:47,540 --> 00:05:48,520 [Eric] the fence-shaped person, 00:05:49,560 --> 00:05:54,420 [Eric] uh, is generally really good in early stage- 00:05:54,420 --> 00:05:54,860 [John] Okay 00:05:54,860 --> 00:06:08,896 [Eric] ... because, you know, or earlier stage companies because-There are fewer resources and sort of everything needs to be done, and you need to wear a different, you know, a bunch of different hats, as the cliche phrase goes, right? 00:06:08,896 --> 00:06:09,755 [John] Right. 00:06:09,756 --> 00:06:15,255 [Eric] And so the fence shaped person is pretty dang good at a lot of different things. 00:06:15,256 --> 00:06:15,716 [John] Right. 00:06:15,716 --> 00:06:22,456 [Eric] And so, you know, in an early-stage company, especially in tech startups, there's just always a fire to put out, right? 00:06:22,456 --> 00:06:22,476 [John] Yeah. 00:06:22,476 --> 00:06:23,956 [Eric] Like you said, there's a marketing thing- 00:06:23,956 --> 00:06:23,966 [John] Right 00:06:23,966 --> 00:06:24,816 [Eric] ... there's a product thing- 00:06:24,816 --> 00:06:24,826 [John] Right 00:06:24,826 --> 00:06:26,776 [Eric] ... there's a customer thing, there's a whatever, right? 00:06:26,776 --> 00:06:27,536 [John] Yeah. 00:06:27,536 --> 00:06:28,036 [Eric] And so 00:06:29,076 --> 00:06:33,156 [Eric] being pretty good at a lot of those different things, you create a ton of value. 00:06:33,156 --> 00:06:41,356 [John] Yeah. I think there's two ways to think. There, there's always a fire to put out, on the flip side is there's always a moving bottleneck, right? 00:06:41,356 --> 00:06:41,856 [Eric] Hmm. 00:06:41,856 --> 00:06:46,076 [John] Because if... I mean, think about product engineering design- 00:06:46,076 --> 00:06:46,376 [Eric] Yep 00:06:46,376 --> 00:06:47,366 [John] ... like revenue, 00:06:48,476 --> 00:06:55,416 [John] like de- depending... Like if you're, if revenue's going great, you probably have like a, you may have an engineering bottleneck if it's SaaS. 00:06:55,416 --> 00:06:55,856 [Eric] Yep. 00:06:55,856 --> 00:07:02,116 [John] If [chuckles] revenue's not going so great, then like that's the one. And it doesn't matter how good the product is, you're gonna have problems. 00:07:02,116 --> 00:07:02,376 [Eric] Yep. 00:07:02,376 --> 00:07:03,156 [John] So. 00:07:03,156 --> 00:07:15,616 [Eric] Totally. A, a good example of this is someone that I worked with who was, uh, it was at a company, and I actually hired this person as, they were a software engineer. 00:07:15,616 --> 00:07:16,356 [John] Okay. 00:07:16,356 --> 00:07:24,126 [Eric] But I hired them, and it was a very technical product. I hired them as a customer success manager, right? 00:07:24,126 --> 00:07:24,135 [John] Okay. 00:07:24,136 --> 00:07:26,936 [Eric] But we were early enough to where they were the first dedicated- 00:07:26,936 --> 00:07:27,096 [John] Right 00:07:27,096 --> 00:07:31,396 [Eric] ... you know, customer facing hire that was not pre-sales. 00:07:31,396 --> 00:07:32,036 [John] Right. 00:07:32,036 --> 00:07:38,426 [Eric] And which means they basically did everything, you know, customer support, you know, account renewal, all, all the things, right? 00:07:38,426 --> 00:07:40,576 [John] Right. Troubleshooting customer issues, [chuckles] like- 00:07:40,576 --> 00:07:41,316 [Eric] Yeah, 100%. 00:07:41,316 --> 00:07:41,336 [John] Yeah. Right. 00:07:41,336 --> 00:07:43,176 [Eric] 100%, right? Gave product feedback. 00:07:44,236 --> 00:07:49,256 [Eric] And this person went on to, uh, actually become a product manager. 00:07:49,256 --> 00:07:49,756 [John] Okay. 00:07:49,756 --> 00:07:52,876 [Eric] Which is not a, that's not an uncommon path, you know, to sort of- 00:07:52,876 --> 00:07:52,996 [John] Sure 00:07:52,996 --> 00:08:07,156 [Eric] ... have a technical background, have some customer facing role that's highly technical, you know, which means that you're very, very close to the product problems. So they go on to become a product manager. Um, and then they actually went on to become a sales engineer. 00:08:07,156 --> 00:08:07,576 [John] Okay. 00:08:07,576 --> 00:08:13,046 [Eric] Um, you know, pre-sales, and so they did a bunch of different things within the organization. 00:08:13,046 --> 00:08:13,276 [John] Right. 00:08:13,276 --> 00:08:21,776 [Eric] And, and so that, in that early stage, it's ama- and that's a great example of, okay, we're bringing on our first customers, they're getting larger. We need someone to make sure those customers are okay, great. 00:08:21,776 --> 00:08:22,616 [John] Right. 00:08:22,616 --> 00:08:31,356 [Eric] They build a team that stabilizes. It's like, okay, great. Well, we actually need, based on all that we're uncovering, to focus on the product, you know? So can we move this person into a product role? Um- 00:08:31,356 --> 00:08:31,536 [John] Right. 00:08:32,656 --> 00:08:42,925 [Eric] I think in larger organizations, a lot of time this manifests in zero to one projects, where someone can get a lot done without a lot of dependencies, right? 00:08:42,925 --> 00:08:42,955 [John] Yes. 00:08:42,956 --> 00:09:08,196 [Eric] So in an early stage, you know, you're sort of, you can move around, you can do a lot of different things, you can manage a lot of responsibility. Uh, but in, you know, a, in a much larger organization that's, you know, where there are more resources, it's less agile. Usually those people are, "Hey, can you go, like, get this project off the ground?" And you just have very few dependencies 'cause you can do so many things yourself. 00:09:08,196 --> 00:09:08,306 [John] Right. 00:09:08,306 --> 00:09:09,416 [Eric] Right? Um, 00:09:10,696 --> 00:09:24,016 [Eric] but it can also be difficult for the fence shaped person to sort of move up traditional paths of management or, you know, within an organizational hierarchy because they don't necessarily have the T piece. 00:09:24,016 --> 00:09:24,616 [John] Hmm. 00:09:24,616 --> 00:09:26,776 [Eric] If that makes sense. Would you agree with that or no? 00:09:26,776 --> 00:09:28,496 [John] I don't know, because, 00:09:30,536 --> 00:09:34,516 [John] uh, tr- I mean, you said traditional path, so maybe. But it, it depends on what the roles are. 00:09:34,516 --> 00:09:35,016 [Eric] Hmm. 00:09:35,016 --> 00:09:35,236 [John] So, 00:09:36,616 --> 00:09:41,376 [John] like, I'll think about technology, like if you're wanna be like a CTO or CIO. Like 00:09:42,436 --> 00:09:46,366 [John] I, I don't know the exact stats, but I think the ranked order is, 00:09:47,676 --> 00:09:51,976 [John] especially like CTO, most likely to come from some kind of infrastructure- 00:09:51,976 --> 00:09:52,276 [Eric] Mm-hmm 00:09:52,276 --> 00:09:52,686 [John] ... type role. 00:09:53,736 --> 00:09:55,506 [John] Infrastructure reliability, like second- 00:09:55,506 --> 00:09:55,506 [Eric] Yep 00:09:55,506 --> 00:10:00,736 [John] ... most likely. And this is probably just by like number of people that have jobs, like some kind of dev type background. 00:10:00,736 --> 00:10:00,756 [Eric] Yep. 00:10:00,756 --> 00:10:05,036 [John] And then you do like management and that. And then third would be some- like something else- 00:10:05,036 --> 00:10:05,166 [Eric] Yep 00:10:05,166 --> 00:10:07,616 [John] ... which is like data or- 00:10:07,616 --> 00:10:07,696 [Eric] Yep 00:10:07,696 --> 00:10:11,316 [John] ... like some that were like analyst types, like, you know. 00:10:11,316 --> 00:10:11,736 [Eric] Yep. 00:10:11,736 --> 00:10:12,936 [John] Um, so 00:10:14,196 --> 00:10:25,076 [John] from that definition, like if you think about it, like yes, like the specific like infrastructure, like management, like CTO, like that's kind of like a, was really deep in this piece and- 00:10:25,076 --> 00:10:25,456 [Eric] Yep 00:10:25,456 --> 00:10:28,056 [John] ... um, but I think the depth is like 00:10:29,376 --> 00:10:32,676 [John] at the end of the day might just be a proxy for like trust- 00:10:32,676 --> 00:10:32,746 [Eric] Hmm 00:10:32,746 --> 00:10:37,816 [John] ... of expertise. Like establishing yourself as like a trusted expert in a thing. 00:10:37,816 --> 00:10:38,186 [Eric] Yeah. 00:10:38,186 --> 00:10:41,036 [John] And like might not matter what it is. 00:10:41,036 --> 00:10:41,656 [Eric] Yep. 00:10:41,656 --> 00:10:46,796 [John] And then like the second level is like establishing your trust a little more specifically with 00:10:47,936 --> 00:10:53,636 [John] some form of like management leadership, so that's not like a hard thing. That's like more of a soft thing. 00:10:53,636 --> 00:10:54,096 [Eric] Yep. 00:10:54,096 --> 00:10:59,356 [John] And then beyond that, like maybe like vision or something like is maybe a third, third tier. 00:10:59,356 --> 00:10:59,656 [Eric] Yeah. 00:10:59,656 --> 00:11:06,056 [John] But I think it's just like trust and like decision-making a- across like a couple of different vectors. 00:11:06,056 --> 00:11:13,136 [Eric] Yep. I agree. I think the other, I think the other angle on this, if we think about a software engineer or, 00:11:14,236 --> 00:11:17,536 [Eric] you know, let's say someone in marketing or someone in data, 00:11:18,676 --> 00:11:28,196 [Eric] I think the other thing about a true fence shaped person is that you're, you have a bit of a ceiling on, um, 00:11:29,596 --> 00:11:35,136 [Eric] sort of a, like moving up the local hierarchy within a particular discipline. 00:11:35,136 --> 00:11:35,456 [John] Okay. 00:11:35,456 --> 00:11:35,636 [Eric] Right? 00:11:35,636 --> 00:11:35,996 [John] Sure. 00:11:35,996 --> 00:11:38,236 [Eric] So for example, you know, uh, 00:11:39,536 --> 00:11:42,216 [Eric] let's say the fence shaped person, 00:11:43,796 --> 00:11:51,596 [Eric] you know, has a decent understanding of data and analytics. You know, they know enough SQL to get the right answers. Again, we sort of go back to like- 00:11:51,596 --> 00:11:51,656 [John] Right 00:11:51,656 --> 00:11:53,996 [Eric] ... they can do a lot of things without dependencies. 00:11:53,996 --> 00:11:54,496 [John] Yes. 00:11:54,496 --> 00:11:55,376 [Eric] Um- 00:11:55,376 --> 00:11:55,556 [John] Yep. 00:11:56,696 --> 00:11:59,416 [Eric] But they're not an expert on statistical analysis- 00:11:59,416 --> 00:11:59,776 [John] Sure 00:11:59,776 --> 00:12:02,976 [Eric] ... you know, or, you know, sort of building a regression model. 00:12:02,976 --> 00:12:03,316 [John] Yep. 00:12:03,316 --> 00:12:05,176 [Eric] Right? I mean, they may understand those things. 00:12:05,176 --> 00:12:06,056 [John] Right. 00:12:06,056 --> 00:12:25,032 [Eric] But they are not an expert in that, right? And so-That's why I think there's this interesting limitation, because in generally moving up in hierarchy within a particular, like division of the company or department, requires a high level of specialization. 00:12:25,032 --> 00:12:28,011 [John] Yeah, sure. That makes sense. Well, and it's... 00:12:30,052 --> 00:12:31,662 [John] I think it depends on the company. 00:12:31,662 --> 00:12:31,662 [Eric] Yeah. 00:12:31,662 --> 00:12:33,542 [John] 'Cause like there, there's, there are definitely some 00:12:34,932 --> 00:12:38,351 [John] not, like, typically non-startup, non-software companies- 00:12:38,352 --> 00:12:38,502 [Eric] Yep 00:12:38,502 --> 00:12:40,392 [John] ... that it's actually ends up being 00:12:41,532 --> 00:12:44,272 [John] not at all the people that are, like, best at the jobs. 00:12:44,272 --> 00:12:45,032 [Eric] Yes. Yeah, yeah, yeah. 00:12:45,032 --> 00:12:48,102 [John] Like, the negative term is, like, failing upward, right? 00:12:48,102 --> 00:12:48,112 [Eric] Sure. 00:12:48,112 --> 00:12:53,892 [John] Like, essentially, like, these people that are like, "Oh, like, we'll move you into management. Like, we actually don't want you doing the work." [chuckles] 00:12:53,892 --> 00:12:54,452 [Eric] Yeah, yeah, yeah. 00:12:54,452 --> 00:12:56,512 [John] Um, and there's that track, which I don't think is ideal. 00:12:56,512 --> 00:12:56,931 [Eric] Yeah. 00:12:56,932 --> 00:12:57,552 [John] Um- 00:12:57,552 --> 00:12:58,732 [Eric] But it is interesting because- 00:12:58,732 --> 00:12:58,802 [John] It is interesting 00:12:58,802 --> 00:13:03,272 [Eric] ... you see this, you see this happen all the time, and then I wanna talk about how I think that AI is- 00:13:03,272 --> 00:13:03,352 [John] Yeah 00:13:03,352 --> 00:13:04,052 [Eric] ... changing this. 00:13:04,052 --> 00:13:05,032 [John] Right. 00:13:05,032 --> 00:13:20,272 [Eric] Uh, but I think this resonates with a lot of people's experience where there's someone who is extremely valuable in the... Well, let's actually think about two different people. Let's think about a T-shaped person, right? 00:13:21,282 --> 00:13:31,212 [Eric] And they, they create most of their value based on the, like, depth of the vertical part of the T. 00:13:31,212 --> 00:13:31,842 [John] Right. 00:13:31,842 --> 00:13:33,212 [Eric] Right? So let's say a data analyst who's- 00:13:33,212 --> 00:13:33,572 [John] Sure 00:13:33,572 --> 00:13:34,332 [Eric] ... unbelievable- 00:13:34,332 --> 00:13:34,632 [John] Yeah 00:13:34,632 --> 00:13:50,772 [Eric] ... at statistical modeling, and so they're just incredible at forecasting, right? But it can actually be difficult for that person to... They... A better way to say it would be they have to build a bunch of additional skills in order to break out of that local maximum- 00:13:50,772 --> 00:13:50,782 [John] Yeah 00:13:50,782 --> 00:13:51,792 [Eric] ... if that makes sense, right? 00:13:51,792 --> 00:13:52,711 [John] Yeah. 00:13:52,712 --> 00:13:53,172 [Eric] Um, 00:13:54,472 --> 00:14:00,732 [Eric] and so a lot of times you can see these really talented people who don't ha- And maybe they don't want it. I'm not saying that, you know- 00:14:00,732 --> 00:14:01,052 [John] Mm-hmm 00:14:01,052 --> 00:14:04,842 [Eric] ... it's, I'm not, it's not inevitable that everyone has to move up the organization. 00:14:04,842 --> 00:14:05,131 [John] Right. Yeah. 00:14:05,132 --> 00:14:08,562 [Eric] But if you think about career mobility, that person has to add a bunch of skill, right? 00:14:08,562 --> 00:14:08,572 [John] Right. 00:14:08,572 --> 00:14:10,112 [Eric] And there's sort of a local maximum there, right? 00:14:11,212 --> 00:14:26,432 [Eric] And I think with the fence shape person, you see that a lot where in earlier phases of a project or even a company, they have a lot of responsibility. Then as things grow and more experience is needed or more expertise is needed- 00:14:26,432 --> 00:14:26,992 [John] Right 00:14:26,992 --> 00:14:29,361 [Eric] ... they sort of top out, right? And a lot of times those- 00:14:29,361 --> 00:14:29,361 [John] Yes 00:14:29,361 --> 00:14:32,081 [Eric] ... people will leave the company or there are things that- 00:14:32,081 --> 00:14:32,912 [John] They get burned out. 00:14:32,912 --> 00:14:40,861 [Eric] Yeah, they get burned out, or there are things that are, uh, objectively make sense but culturally seem counterintuitive- 00:14:40,861 --> 00:14:40,861 [John] Sure 00:14:40,861 --> 00:14:43,272 [Eric] ... based on the amount of equity that this person has built up. 00:14:43,272 --> 00:14:43,602 [John] Right. 00:14:43,602 --> 00:14:47,872 [Eric] Right? And so a classic example would be an early sales leader- 00:14:47,872 --> 00:14:48,312 [John] Right 00:14:48,312 --> 00:14:51,192 [Eric] ... who helps a company pass, like, multiple milestones- 00:14:51,192 --> 00:14:51,262 [John] Yeah 00:14:51,262 --> 00:14:53,372 [Eric] ... and it's sort of like they're the GOAT, you know? 00:14:53,372 --> 00:14:53,892 [John] Yeah. Right. 00:14:53,892 --> 00:14:57,112 [Eric] And then they get layered and, you know. 00:14:57,112 --> 00:14:57,172 [John] Yeah. 00:14:57,172 --> 00:15:01,402 [Eric] Or they're essentially, you know, it's like, "Okay, you, you either need to go do something else or go find- 00:15:01,402 --> 00:15:01,402 [John] Right 00:15:01,402 --> 00:15:05,852 [Eric] ... another role or whatever." And it's because, like, you just don't have the experience to do this at scale. 00:15:05,852 --> 00:15:06,412 [John] Yep. 00:15:06,412 --> 00:15:07,792 [Eric] Um- 00:15:07,792 --> 00:15:20,172 [John] Here, here's what I think the danger of the generalist is, 'cause I think there's a lot of upsides. I think there's two. One is, like, what you're describing, [clears throat] is, like, the, the burnout, get layered, like, don't, 00:15:21,212 --> 00:15:22,232 [John] aren't deep enough- 00:15:22,232 --> 00:15:22,242 [Eric] Mm-hmm 00:15:22,242 --> 00:15:23,212 [John] ... and, like, maybe some pieces. 00:15:23,212 --> 00:15:23,832 [Eric] Mm-hmm. 00:15:23,832 --> 00:15:35,652 [John] But I think there's a reason why. And the, I think the superpower of the generalist is, like, I can come with this, like, arsenal, like a wide arsenal, 00:15:36,672 --> 00:15:46,732 [John] and then the superpower is if I figure out how to align the pieces of my arsenal that are best aligned with the current bottlenecks, problems, or, like, ways I can add value, like, that's the superpower. 00:15:46,732 --> 00:15:47,232 [Eric] Mm. Yes. 00:15:47,232 --> 00:15:48,302 [John] 'Cause I'm wide. 00:15:48,302 --> 00:15:48,312 [Eric] Yes. 00:15:48,312 --> 00:15:48,982 [John] Like, that's- 00:15:48,982 --> 00:15:49,182 [Eric] Yes. Yes 00:15:49,182 --> 00:15:50,392 [John] ... that's the advantage. So, 00:15:51,492 --> 00:15:56,192 [John] so the, the mistakes are I do too much because I can. Like, that's- 00:15:56,192 --> 00:15:56,201 [Eric] Yes 00:15:56,201 --> 00:15:57,352 [John] ... like, is the number one mistake. 00:15:57,352 --> 00:15:57,822 [Eric] Yep. Agreed. 00:15:57,822 --> 00:16:00,772 [John] 'Cause I, I am wide, and I can do all these things, and I do too much. 00:16:00,772 --> 00:16:00,792 [Eric] Yep. 00:16:00,792 --> 00:16:13,512 [John] My value is not focused enough, and, and it's not... And even though it's, like, maybe, like, shallower than some people, because I have a, like, innate ability to align value with what's needed at the moment- 00:16:13,512 --> 00:16:13,912 [Eric] Yes. Agreed 00:16:13,912 --> 00:16:33,872 [John] ... and to ignore the other things that I could do 'cause I have the ability but, but I know aren't, like, the bottleneck movers. That to, that to me is the magic of, like, the best generalist and how they can continue to elevate is that focused... And, and then it moves, and then they, like, recognize, like, "Okay, I gotta give this away- 00:16:33,872 --> 00:16:34,012 [Eric] Yep 00:16:34,012 --> 00:16:44,702 [John] ... because it's not the bottleneck anymore. It's not the most high-value thing." Figure out how to give it away and then shift and then focus their, like, their, their expertise there. 00:16:44,702 --> 00:16:44,712 [Eric] Totally agree. 00:16:44,712 --> 00:16:45,832 [John] And do it again and again and again. 00:16:45,832 --> 00:16:46,792 [Eric] That is the superpower. Yep. 00:16:46,792 --> 00:16:51,972 [John] And deal with the, like, uncomfort of every time you have to do that 'cause you're letting go of something that, like- 00:16:51,972 --> 00:16:52,112 [Eric] Yes 00:16:52,112 --> 00:16:54,012 [John] ... you succeeded in and that you're good at and whatever. 00:16:54,012 --> 00:16:54,692 [Eric] Yep. 00:16:54,692 --> 00:16:59,751 [John] Like, that, that's it. Like, i- if you can do that, and regardless of AI, that's always been the superpower. 00:16:59,752 --> 00:17:00,052 [Eric] Yep. 00:17:00,052 --> 00:17:01,392 [John] And I don't think AI changes that. 00:17:02,972 --> 00:17:03,952 [Eric] I agree. 00:17:03,952 --> 00:17:05,742 [John] Like, high level, I think the practical changes. 00:17:05,742 --> 00:17:24,552 [Eric] I, I totally agree. I think practically what's really interesting about AI is that it makes... It, I think it diminishes the barriers that traditionally are created around the depth of the vertical part of a T. 00:17:24,552 --> 00:17:25,412 [John] Yeah. 00:17:25,412 --> 00:17:29,002 [Eric] So let's think about the fence shape person in the example that we used earlier- 00:17:29,002 --> 00:17:29,002 [John] Right 00:17:29,002 --> 00:17:29,832 [Eric] ... around data analysis. 00:17:29,832 --> 00:17:35,992 [John] So you're saying the fence posts are, like, deeper than they would have been before. [laughs] 00:17:35,992 --> 00:17:37,892 [Eric] [laughs] Flagpole and fence post. 00:17:37,892 --> 00:17:37,972 [John] Okay. 00:17:37,972 --> 00:17:39,332 [Eric] There we go. We even get alliteration. 00:17:39,332 --> 00:17:39,692 [John] There you go. Yeah, yeah. 00:17:39,692 --> 00:17:39,892 [Eric] Right? 00:17:39,892 --> 00:17:40,992 [John] I like it. 00:17:40,992 --> 00:17:42,592 [Eric] Well, the... 00:17:44,172 --> 00:17:53,712 [Eric] Yeah. S- So if you think about the fence shape person, you have, you know, a sort of relatively similar level of depth, um, 00:17:54,812 --> 00:17:57,032 [Eric] you know, in all of these fence posts that go into the ground, right? 00:17:58,352 --> 00:18:03,792 [Eric] And what AI does is allows you to achieve, 00:18:05,092 --> 00:18:07,252 [Eric] not in everything and not in every situation- 00:18:07,252 --> 00:18:07,572 [John] Right 00:18:07,572 --> 00:18:15,848 [Eric] ... but far more than was previously possible, it allows you to achieve depth... where you previously didn't have it. 00:18:15,848 --> 00:18:16,068 [John] Yes. 00:18:16,068 --> 00:18:22,258 [Eric] Right? So, and I would say that I'm a good example of this specifically in the data space, right? 00:18:22,258 --> 00:18:22,478 [John] Okay. Yeah. 00:18:22,478 --> 00:18:23,688 [Eric] I've done a lot of analytics. 00:18:23,688 --> 00:18:23,698 [John] Right. 00:18:23,698 --> 00:18:25,048 [Eric] I worked for a company that did- 00:18:25,048 --> 00:18:25,118 [John] Yeah 00:18:25,118 --> 00:18:26,838 [Eric] ... you know, data for analytics and whatever- 00:18:26,838 --> 00:18:26,838 [John] Yeah 00:18:26,838 --> 00:18:26,998 [Eric] ... right? 00:18:28,328 --> 00:18:34,898 [Eric] And I have a really good understanding of a regression analysis and how to use it- 00:18:34,898 --> 00:18:34,898 [John] Uh-huh 00:18:34,898 --> 00:18:39,678 [Eric] ... and have used regression analyses to, to make decisions, right? 00:18:39,678 --> 00:18:39,728 [John] Right. 00:18:39,728 --> 00:18:46,968 [Eric] But I, you know, previously could not have, you know, written Python- 00:18:46,968 --> 00:18:47,248 [John] Yeah 00:18:47,248 --> 00:18:48,558 [Eric] ... you know, over a dataset- 00:18:48,558 --> 00:18:48,578 [John] Looked one up. Right 00:18:48,578 --> 00:18:52,418 [Eric] ... and like, you know, used Jupyter Notebooks to do a regression- 00:18:52,418 --> 00:18:52,468 [John] Right 00:18:52,468 --> 00:18:54,808 [Eric] ... analysis that was, you know- 00:18:54,868 --> 00:18:54,878 [John] Right 00:18:54,878 --> 00:18:56,328 [Eric] ... [laughs] was legitimate- 00:18:56,328 --> 00:18:56,378 [John] Right 00:18:56,378 --> 00:18:57,208 [Eric] ... and accurate, right? 00:18:57,208 --> 00:18:58,008 [John] Right. 00:18:58,008 --> 00:19:01,608 [Eric] And so what I was doing was sort of taking my knowledge of like I, 00:19:02,668 --> 00:19:06,468 [Eric] I have a question that needs to be answered and decisions that need to be made. 00:19:06,468 --> 00:19:07,048 [John] Right. 00:19:07,048 --> 00:19:11,788 [Eric] I understand the tools that need to be used, but I am not a T shape in that regard, so I have- 00:19:11,788 --> 00:19:11,908 [John] Yeah, yeah 00:19:11,908 --> 00:19:13,108 [Eric] ... to go work with someone who is- 00:19:13,108 --> 00:19:13,608 [John] Exactly 00:19:13,608 --> 00:19:14,748 [Eric] ... in order for them to do the work, right? 00:19:14,748 --> 00:19:18,268 [John] Or decide not to do it 'cause you like don't have the resource or don't think it's worth the time. 00:19:18,268 --> 00:19:18,788 [Eric] Exactly. 00:19:18,788 --> 00:19:18,928 [John] Yeah. Right. 00:19:18,928 --> 00:19:20,648 [Eric] Which is probably way more common than the other- 00:19:20,648 --> 00:19:20,888 [John] Yeah. Sure, sure 00:19:20,888 --> 00:19:22,308 [Eric] ... right? Unless it's really critical. 00:19:22,308 --> 00:19:22,808 [John] Right. 00:19:22,808 --> 00:19:27,808 [Eric] But what's interesting now is that I can do that very easily. 00:19:27,808 --> 00:19:28,008 [John] Mm-hmm. 00:19:28,008 --> 00:19:31,088 [Eric] Right? Like, I can perform regression analyses, um- 00:19:31,088 --> 00:19:31,448 [John] Right 00:19:31,448 --> 00:19:33,148 [Eric] ... you know, using Claude Code. 00:19:33,148 --> 00:19:33,208 [John] Yeah. 00:19:33,208 --> 00:19:35,008 [Eric] And it's extremely quick. 00:19:35,008 --> 00:19:35,528 [John] Right. 00:19:35,528 --> 00:19:39,958 [Eric] Um, and it's not that difficult to validate. And honestly- 00:19:39,958 --> 00:19:39,958 [John] Yeah 00:19:39,958 --> 00:19:49,708 [Eric] ... even if I have questions about it or I want validation, it's f- it is so much faster for an actual data analyst to, to review it quickly- 00:19:49,708 --> 00:19:49,718 [John] Sure 00:19:49,718 --> 00:19:51,808 [Eric] ... and give it a thumbs up or not, right? 00:19:51,808 --> 00:19:51,838 [John] Yeah. 00:19:51,838 --> 00:19:59,648 [Eric] And so this cycle that used to either prohibit people from doing work or take a really long time- 00:19:59,648 --> 00:19:59,788 [John] Yeah 00:19:59,788 --> 00:20:02,688 [Eric] ... because a fence shape person had to go to a T shape person- 00:20:02,688 --> 00:20:03,308 [John] Right 00:20:03,308 --> 00:20:07,478 [Eric] ... that is, those, um, those barriers are being lowered. 00:20:07,478 --> 00:20:07,508 [John] Yeah. 00:20:07,508 --> 00:20:13,788 [Eric] And what's interesting is that it's not just with things like data or writing code, it's a lot of different things, right? Like research. 00:20:13,788 --> 00:20:14,388 [John] Yeah. 00:20:14,388 --> 00:20:17,768 [Eric] Um, I mean, it's, it's pretty incredible actually. 00:20:17,768 --> 00:20:24,728 [John] Yeah. You, you know what's funny about that? Because this is my initial reaction, which I know is not fair. I think to myself, "Yeah, it wasn't that hard before." 00:20:24,728 --> 00:20:24,998 [Eric] [laughs] 00:20:24,998 --> 00:20:29,117 [John] Um, but, but you know, that's just 'cause of my background. And, and, and I'd say- 00:20:29,117 --> 00:20:29,168 [Eric] Yeah 00:20:29,168 --> 00:20:36,207 [John] ... not that hard, like if you like grab a like off-the-shelf library, it, it, it's not that many lines of code. Like, it's not that hard. 00:20:36,208 --> 00:20:36,608 [Eric] Yep. 00:20:36,608 --> 00:20:50,268 [John] But, like, it would be for somebody without the context or without the knowledge. And, and I think the most interesting thing about AI that's not talked about enough is there's a ton of things that were, quote, like, actually not that hard before, and the main problem was the, like, 00:20:51,368 --> 00:20:59,707 [John] starting point problem and the like... Like, it-- Like, I-- We could've sat down, I could've showed you like how to do a basic regression analysis and give you a template in Python- 00:20:59,708 --> 00:20:59,897 [Eric] Yep 00:20:59,897 --> 00:21:03,608 [John] ... in like 20 minutes and told you like, "Next time you wanna do it, like fill in these things," and you would've had- 00:21:03,648 --> 00:21:03,658 [Eric] Mm-hmm 00:21:03,658 --> 00:21:04,188 [John] ... a tool. 00:21:04,188 --> 00:21:05,188 [Eric] Mm-hmm. Like, yeah. 00:21:05,188 --> 00:21:11,668 [John] That was possible before. But now... But you didn't know that, and there's like a million things out there- 00:21:11,668 --> 00:21:12,948 [Eric] Oh, that's such a good point 00:21:12,948 --> 00:21:15,418 [John] ... that, that are like that, that like I don't know- 00:21:15,418 --> 00:21:15,418 [Eric] Yep 00:21:15,418 --> 00:21:22,288 [John] ... how easy it would've been for an expert to show me in 20 minutes how to do X. And now, like, it's a thing. 00:21:22,288 --> 00:21:23,158 [Eric] That's such a good point. 00:21:23,158 --> 00:21:23,698 [John] And, and I think- 00:21:23,698 --> 00:21:26,368 [Eric] Yeah. It wasn't that... Yeah. There are so many things where it wasn't that hard before. 00:21:26,368 --> 00:21:26,568 [John] Yeah. 00:21:26,568 --> 00:21:26,788 [Eric] Yep. 00:21:26,788 --> 00:21:41,218 [John] But, but none of us knew except in our like little areas. So like the... So that, that's like a really undersold piece of AI. And then there are some things, and I think these are where people get really impressed, that were hard before and it's still like really good at. 00:21:41,218 --> 00:21:41,228 [Eric] Yes. 00:21:41,228 --> 00:21:45,368 [John] But I think we skip over this like really long breadth of things that were... They actually weren't that hard before. 00:21:45,368 --> 00:21:49,897 [Eric] I agree. And, uh, that is really interesting be- to me because 00:21:50,908 --> 00:22:01,108 [Eric] that, in those situations, I think many times if we give ourselves and other people the benefit of the doubt, they were making a good decision on... Well, let's take me for example- 00:22:01,108 --> 00:22:01,117 [John] Yeah 00:22:01,117 --> 00:22:02,328 [Eric] ... with the regression analysis, right? 00:22:02,328 --> 00:22:02,338 [John] Mm-hmm. 00:22:04,328 --> 00:22:15,888 [Eric] It wasn't that hard, and I had even thought multiple times, you know, over the years like, "I could just, I could sit down and learn this," but the question- 00:22:15,888 --> 00:22:15,897 [John] Right 00:22:15,897 --> 00:22:18,018 [Eric] ... is, is that the best use of my time? 00:22:18,018 --> 00:22:18,668 [John] Yeah. Right. 00:22:18,668 --> 00:22:18,768 [Eric] You know? 00:22:18,768 --> 00:22:20,658 [John] Which is back to the value alignment thing we were talking about. 00:22:20,658 --> 00:22:25,008 [Eric] Right. It is like making that t- that, that fence post- 00:22:25,008 --> 00:22:25,088 [John] Right 00:22:25,088 --> 00:22:29,258 [Eric] ... is like, is making that deeper the best use of my time? 00:22:29,258 --> 00:22:29,288 [John] Right. 00:22:29,288 --> 00:22:32,388 [Eric] And a lot of times the answer w- is no, right? 00:22:32,388 --> 00:22:32,408 [John] Right. 00:22:32,408 --> 00:22:38,328 [Eric] But that was with the mindset of like, okay, well I don't... And part of that is because it's an unknown- 00:22:38,328 --> 00:22:38,618 [John] Right 00:22:38,618 --> 00:22:42,178 [Eric] ... and I don't know how involved that would be, you know, et cetera, right? 00:22:42,178 --> 00:22:42,188 [John] Right. 00:22:42,188 --> 00:22:46,968 [Eric] And I mean, of course there are situations where you say like, "Well, maybe it is actually valuable for me to do this," and like- 00:22:46,968 --> 00:22:47,228 [John] Sure 00:22:47,228 --> 00:22:48,468 [Eric] ... we've all done things like that. 00:22:48,468 --> 00:22:48,887 [John] Yeah. Right. 00:22:48,888 --> 00:22:55,808 [Eric] But that is really interesting on the un- like the, you don't know how... You don't know that it wasn't actually that hard. 00:22:55,808 --> 00:22:56,248 [John] Yeah. 00:22:56,248 --> 00:23:00,468 [Eric] And so you make probably a wise judgment call on it's not worth my time- 00:23:00,468 --> 00:23:00,608 [John] Right 00:23:00,608 --> 00:23:02,668 [Eric] ... uh, because I'm just not gonna use it that often. 00:23:02,668 --> 00:23:02,888 [John] Right. 00:23:02,888 --> 00:23:04,548 [Eric] Right? Or whatever it is. 00:23:04,548 --> 00:23:12,368 [John] Yeah. And there's just so many, like, vectors of that. Like, I've never been like a real, like, front-end guy. I d- like I definitely could've learned it. 00:23:12,368 --> 00:23:13,317 [Eric] Mm-hmm. Yeah. 00:23:13,317 --> 00:23:15,817 [John] Um, and then like with the AI stuff, it's like, oh. 00:23:15,817 --> 00:23:16,108 [Eric] It's, yeah. 00:23:16,108 --> 00:23:20,548 [John] And even as you like see code being generated, you're like, "Oh, actually." Like, I... You know what I mean? [laughs] Like now- 00:23:20,548 --> 00:23:20,948 [Eric] Yeah 00:23:20,948 --> 00:23:22,348 [John] ... and you're like, "Okay, like I could've done this." 00:23:22,348 --> 00:23:27,848 [Eric] That's a great... That's a really good example of going the other way, where sort of back-end database- 00:23:27,848 --> 00:23:27,968 [John] Yeah 00:23:27,968 --> 00:23:28,828 [Eric] ... and data stuff- 00:23:28,828 --> 00:23:29,007 [John] Right 00:23:29,007 --> 00:23:30,598 [Eric] ... has been your realm for- 00:23:30,598 --> 00:23:30,598 [John] Right 00:23:30,598 --> 00:23:32,317 [Eric] ... you know, 15 years. 00:23:32,317 --> 00:23:32,317 [John] Yeah. 00:23:32,317 --> 00:23:36,578 [Eric] And there wasn't a need for you to sit down and learn how to write React. 00:23:36,578 --> 00:23:36,588 [John] Yeah, exactly. 00:23:36,588 --> 00:23:37,618 [Eric] Like, it just wasn't- 00:23:37,618 --> 00:23:37,618 [John] Right 00:23:37,618 --> 00:23:38,968 [Eric] ... it wasn't that valuable. 00:23:38,968 --> 00:23:39,048 [John] Right. 00:23:39,048 --> 00:23:40,228 [Eric] But it also wasn't that- 00:23:40,228 --> 00:23:41,778 [John] And I hired people like when I needed it- 00:23:41,778 --> 00:23:41,778 [Eric] Yeah 00:23:41,778 --> 00:23:43,028 [John] ... and, uh, yeah. Right. 00:23:43,028 --> 00:23:44,918 [Eric] But it also wasn't that hard, right? Like you were- 00:23:44,918 --> 00:23:44,918 [John] Right 00:23:44,918 --> 00:23:46,328 [Eric] ... totally capable of doing that. 00:23:46,328 --> 00:23:47,348 [John] Right. Right. 00:23:47,348 --> 00:23:51,988 [Eric] But what's so fascinating to me is that's just a solved problem now, largely. 00:23:51,988 --> 00:23:52,108 [John] Yeah. Yeah. 00:23:52,108 --> 00:23:53,558 [Eric] Right? And I'm not saying that AI- 00:23:53,558 --> 00:23:53,558 [John] Well- 00:23:53,558 --> 00:23:54,468 [Eric] ... doesn't hallucinate- 00:23:54,468 --> 00:23:54,478 [John] Right 00:23:54,478 --> 00:23:54,868 [Eric] ... and there aren't, you know- 00:23:54,868 --> 00:23:55,317 [John] Well, and the hard- 00:23:55,317 --> 00:23:55,928 [Eric] ... problems, but- 00:23:55,928 --> 00:23:58,288 [John] ... and the, and the hard is relative like anything. Like, 00:23:59,368 --> 00:24:01,868 [John] like SQL is my favorite example of this. Like, 00:24:03,118 --> 00:24:07,148 [John] m- there's a lot of people out there that could learn SQL without AI in a weekend. 00:24:07,148 --> 00:24:07,548 [Eric] Mm-hmm. 00:24:07,548 --> 00:24:13,488 [John] I've like personally taught maybe a half a dozen people in very short amounts of time to write basic SQL. 00:24:13,488 --> 00:24:14,068 [Eric] Mm-hmm. 00:24:14,068 --> 00:24:23,072 [John] It is like one of the easiest, quote, "programming languages." To be an absolute like master of SQL, like-I only know a few. 00:24:23,072 --> 00:24:23,672 [Speaker 2] Mm-hmm. 00:24:23,672 --> 00:24:27,892 [John] And, and that's, like, really deep, like, optimizations of like- 00:24:27,892 --> 00:24:27,901 [Speaker 2] Yep 00:24:27,901 --> 00:24:31,762 [John] ... y- you know, I mean, there, there's a bunch of, like, directions you can go, like analytics or production optimization- 00:24:31,762 --> 00:24:31,772 [Speaker 2] Mm-hmm 00:24:31,772 --> 00:24:35,592 [John] ... whatever. So that's my favorite example of like- 00:24:35,592 --> 00:24:35,602 [Speaker 2] Yeah 00:24:35,602 --> 00:24:39,471 [John] ... there's a surprisingly low barrier even without AI to, like, learn how to- 00:24:39,472 --> 00:24:39,522 [Speaker 2] Yep 00:24:39,522 --> 00:24:43,252 [John] ... do a thing, and then there's a surprisingly high barrier to be, like, really, really good. 00:24:43,252 --> 00:24:43,792 [Speaker 2] Yep. 00:24:43,792 --> 00:24:44,082 [John] Um, and- 00:24:44,082 --> 00:24:49,912 [Speaker 2] And so is AI removing some of that distinction because it can write SQL really well? 00:24:49,912 --> 00:24:53,272 [John] Um, yeah. I mean, yes and, yes and no. 00:24:54,892 --> 00:25:08,652 [John] It, I mean, like, w- so, like, let's say this is, like, and, like, for the people not watching. Like, visually, let, let's do it on a scale of numbers. Let's say, like, in a weekend, if, like, 100 is, like, the absolute best ex- expert at SQL- 00:25:08,652 --> 00:25:08,772 [Speaker 2] Mm-hmm 00:25:08,772 --> 00:25:13,292 [John] ... and, like, one is, like... Let's just start at zero. Zero is, you know, nothing. Like, in a weekend, 00:25:14,712 --> 00:25:17,432 [John] um... And let, and let's, let's say the scale is kind of like 00:25:18,692 --> 00:25:21,032 [John] not expertise, but, like, value. 00:25:21,032 --> 00:25:21,072 [Speaker 2] Yeah. 00:25:21,072 --> 00:25:23,272 [John] So, like, in a weekend, you can get- 00:25:23,272 --> 00:25:23,872 [Speaker 2] Not, yeah 00:25:23,872 --> 00:25:23,981 [John] ... to, let's call it, like, 30- 00:25:23,981 --> 00:25:25,252 [Speaker 2] It's not academic knowledge- 00:25:25,252 --> 00:25:25,912 [John] Yeah, yeah, yeah 00:25:25,912 --> 00:25:27,192 [Speaker 2] ... but it's like I can wield this tool- 00:25:27,192 --> 00:25:27,352 [John] Yeah 00:25:27,352 --> 00:25:29,092 [Speaker 2] ... in order to cr- you know, get valuable- 00:25:29,092 --> 00:25:29,652 [John] Yeah 00:25:29,652 --> 00:25:30,112 [Speaker 2] ... you know, valuable insights. 00:25:30,112 --> 00:25:36,122 [John] So the business, like, from a business value, like, you can get, like, pretty high on that scale, 00:25:37,172 --> 00:25:39,932 [John] like, in a weekend. Let's call it, like, maybe 30 or 40 or something. 00:25:39,932 --> 00:25:40,652 [Speaker 2] Okay, wow. 00:25:40,652 --> 00:25:44,912 [John] Um, let's call it 30, um, as far as, like, I can do all the basics to get data out- 00:25:44,912 --> 00:25:44,922 [Speaker 2] Mm-hmm 00:25:44,922 --> 00:25:46,332 [John] ... to, like, make business decisions. 00:25:46,332 --> 00:25:46,471 [Speaker 2] Mm-hmm. 00:25:46,472 --> 00:25:49,552 [John] Which that's the most valuable piece in most situations. 00:25:49,552 --> 00:25:49,932 [Speaker 2] Yep. 00:25:49,932 --> 00:25:55,492 [John] But the other valuable piece is like, "Oh, I'm trying to scale app to, like, hundreds of thousands of users." 00:25:55,492 --> 00:25:56,062 [Speaker 2] Mm. Mm-hmm. 00:25:56,062 --> 00:26:00,502 [John] "And, like, I need a relational database, and, like, I need to optimize some things." 00:26:00,502 --> 00:26:00,552 [Speaker 2] Mm. 00:26:00,552 --> 00:26:01,732 [John] Like, that's, like- 00:26:01,732 --> 00:26:02,092 [Speaker 2] Mm-hmm 00:26:02,092 --> 00:26:04,232 [John] ... a different... It's still SQL. Like- 00:26:04,232 --> 00:26:04,342 [Speaker 2] Yep 00:26:04,342 --> 00:26:13,302 [John] ... but that's a very different, like, skill set. And if we... And it's even higher business value [laughs] if you're trying to scale a, you know, company, and, like- 00:26:13,302 --> 00:26:13,332 [Speaker 2] Yep 00:26:13,332 --> 00:26:14,412 [John] ... that's your bottleneck. 00:26:14,412 --> 00:26:14,852 [Speaker 2] Yep. 00:26:14,852 --> 00:26:20,202 [John] Now, to be fair, a lot of times people just spend money now, right? Like, you can just be like, "I don't know, make the database bigger." 00:26:20,202 --> 00:26:20,572 [Speaker 2] Mm-hmm. 00:26:20,572 --> 00:26:22,532 [John] Um, [laughs] like, people just, like, you can use money- 00:26:22,532 --> 00:26:22,632 [Speaker 2] Right 00:26:22,632 --> 00:26:23,292 [John] ... to bail yourself out- 00:26:23,292 --> 00:26:23,582 [Speaker 2] Right. Yeah 00:26:23,582 --> 00:26:25,192 [John] ... nowadays, 'cause that's pretty easy to do. 00:26:25,192 --> 00:26:25,812 [Speaker 2] Yeah. 00:26:25,812 --> 00:26:38,592 [John] But, like, let's say that wasn't an option. Like, then that, like, the value is way closer to that 100 mark because it's, like, a bottleneck for the entire company, the entire org. Like, and you have constraints or whatever. Like, it's there. 00:26:38,592 --> 00:26:38,692 [Speaker 2] Yep. 00:26:38,692 --> 00:26:42,842 [John] But, like, getting to the, like, if it's about getting the data to make decisions, 00:26:43,872 --> 00:26:51,512 [John] um, manually, like s- you know, let's call it, like, 30 out, out of 100. And then, like, with, with AI, like, maybe you get to, like, I don't know, 50 or 60. Like, maybe you, like, double- 00:26:51,512 --> 00:26:51,522 [Speaker 2] Yeah 00:26:51,522 --> 00:26:52,832 [John] ... or more than that. 00:26:52,832 --> 00:26:53,312 [Speaker 2] Right. 00:26:53,312 --> 00:26:54,102 [John] Like, and make business- 00:26:54,102 --> 00:26:54,452 [Speaker 2] Which is significant 00:26:54,452 --> 00:26:55,332 [John] ... which is super significant. 00:26:55,332 --> 00:26:57,212 [Speaker 2] I mean, that's 100% and- 00:26:57,272 --> 00:26:57,412 [John] Yeah 00:26:57,412 --> 00:26:58,472 [Speaker 2] ... 100% improvement. 00:26:58,472 --> 00:26:59,912 [John] And I think that's about right. 00:26:59,912 --> 00:27:00,572 [Speaker 2] Yep. 00:27:00,572 --> 00:27:04,132 [John] And, and, and the only, like, things we've run into, um... 00:27:05,352 --> 00:27:12,172 [John] The, the majority of the things we've run into have nothing to do with SQL syntax. They have to do with the data. 00:27:12,172 --> 00:27:12,382 [Speaker 2] Yes. 00:27:12,382 --> 00:27:15,162 [John] Like, people don't actually understand how... 00:27:15,162 --> 00:27:15,162 [Speaker 2] Yep 00:27:15,162 --> 00:27:22,972 [John] ... which has always been a lot of the value in an analyst, is, like, lit- it's as simple as this, is like you don't know what filters to put on. 00:27:22,972 --> 00:27:23,692 [Speaker 2] Yeah. 00:27:23,692 --> 00:27:26,832 [John] And, and there's hundreds of, like, options to filter this. 00:27:26,832 --> 00:27:27,592 [Speaker 2] Mm-hmm. 00:27:27,592 --> 00:27:32,692 [John] And it's not abundantly clear, like, which ones to turn on and off to get to the result you want. 00:27:32,692 --> 00:27:33,912 [Speaker 2] Yep. Yep. 00:27:33,912 --> 00:27:35,332 [John] Which is, like, cleaning, right? 00:27:35,332 --> 00:27:36,012 [Speaker 2] Yeah. 00:27:36,012 --> 00:27:43,532 [John] Um, that, that's the value. And, like, if you don't have the context, then, then yeah, you can, like, get to the data all day long, but it's gonna be wrong 'cause you don't know how to- 00:27:43,532 --> 00:27:43,592 [Speaker 2] Yeah 00:27:43,592 --> 00:27:44,791 [John] ... clean it, filter it, sort it. 00:27:44,792 --> 00:27:50,012 [Speaker 2] Yeah. All right. There we have it. The fence shape person- 00:27:50,012 --> 00:27:50,162 [John] I like it 00:27:50,162 --> 00:27:51,692 [Speaker 2] ... the fence and the flagpole. 00:27:51,692 --> 00:27:53,442 [John] The fence [laughs] and the flagpole. 00:27:53,442 --> 00:27:55,452 [Speaker 2] [laughs] Is that the episode title? 00:27:55,452 --> 00:27:56,902 [John] Maybe. Maybe. 00:27:56,902 --> 00:27:59,212 [Speaker 2] [laughs] All right. Well, thanks for joining. 00:27:59,212 --> 00:27:59,242 [John] Awesome. 00:27:59,242 --> 00:28:09,102 [Speaker 2] And we'll catch you on the next show. [outro music]