Navigating skill atrophy in the AI age

00:00:00,080 --> 00:00:20,090 [Eric] [upbeat music] Welcome back to the Token Intelligent Show, where John and I try to navigate AI, business, and life, and you get to follow along [chuckles]- 00:00:20,090 --> 00:00:21,720 [John] ... in real time. 00:00:21,720 --> 00:00:27,340 [Eric] And hopefully learn from us. Today's episode is all about 00:00:27,340 --> 00:00:51,220 [Eric] whether or not, and how skills might atrophy in the age of AI, where we have machines doing a lot of work on our behalf. And there was a post on X from the CEO of Vercel that I thought was really interesting this week. It reminded me of, uh... It reminded me of a conversation that we had last week with a bunch of people- 00:00:51,220 --> 00:00:51,400 [John] Right 00:00:51,400 --> 00:00:52,120 [Eric] ... about skill atrophy. 00:00:52,120 --> 00:00:53,020 [John] Yeah. 00:00:53,020 --> 00:01:03,740 [Eric] And the post says, "If you don't use your body, it atrophies, so you go to the gym or on a run. If you don't use your brain, it atrophies." So what's your plan? 00:01:03,740 --> 00:01:04,879 [John] Wow! 00:01:04,879 --> 00:01:06,010 [Eric] Yeah, kind of a nice- 00:01:06,010 --> 00:01:06,060 [John] Yeah 00:01:06,060 --> 00:01:06,899 [Eric] ... spicy hot take. 00:01:06,900 --> 00:01:07,320 [John] Any, uh- 00:01:07,320 --> 00:01:07,660 [Eric] He's good at that 00:01:07,660 --> 00:01:10,200 [John] ... any good top responses on that one? 00:01:10,200 --> 00:01:13,860 [Eric] Let's see. I can look through- 00:01:13,860 --> 00:01:13,980 [John] Yeah. 00:01:13,980 --> 00:01:16,780 [Eric] Uh, 00:01:16,840 --> 00:01:21,600 [Eric] write shaders, play chess, talk philosophy, read books. 00:01:21,600 --> 00:01:22,900 [John] Okay. 00:01:22,900 --> 00:01:31,200 [Eric] Uh, oh, this is actually interesting. I was going [chuckles]... I, I thought about this: go to a mind gym, next billion-dollar idea. 00:01:31,200 --> 00:01:32,480 [John] [chuckles] A mind gym. 00:01:32,480 --> 00:01:33,220 [Eric] I actually thought about that. 00:01:33,220 --> 00:01:36,290 [John] Isn't that just Sudoku? 00:01:36,290 --> 00:01:36,500 [Eric] [chuckles] 00:01:36,580 --> 00:01:39,360 [John] And, uh, or no, or new- newspaper crossword puzzle? 00:01:39,360 --> 00:01:40,680 [Eric] Sudoku is a mind gym. 00:01:40,680 --> 00:01:42,120 [John] The New York Times crossword puzzle- 00:01:42,120 --> 00:01:42,310 [Eric] Yeah 00:01:42,310 --> 00:01:42,680 [John] ... or Wordle? 00:01:42,680 --> 00:01:43,840 [Eric] Yeah, Wordle, yeah. 00:01:43,840 --> 00:01:44,540 [John] Yeah. 00:01:44,540 --> 00:01:50,260 [Eric] Okay, what- here's, here's how I'll start the conversation: 00:01:50,260 --> 00:01:53,620 [Eric] What skills have you felt atrophy in the age of AI? 00:01:53,620 --> 00:01:53,630 [John] Mm. 00:01:53,630 --> 00:01:56,560 [Eric] 'Cause you and I use AI tools extensively- 00:01:56,560 --> 00:01:56,570 [John] Right 00:01:56,570 --> 00:01:57,450 [Eric] ... every day. 00:01:57,450 --> 00:01:59,240 [John] And probably differently, too. 00:01:59,240 --> 00:02:00,280 [Eric] Yes. 00:02:00,280 --> 00:02:09,759 [John] Yeah. I mean, the... Good question. I, I think, I mean, AI is best at coding, so I think it's natural, like, if you do any coding at all, um, like- 00:02:09,759 --> 00:02:09,979 [Eric] Yes 00:02:09,979 --> 00:02:18,320 [John] ... that's, you know, uh, it is better at coding than, than just about anything else. So that, that for sure, to some extent. 00:02:18,320 --> 00:02:20,460 [Eric] So your personal coding skills have atrophied? 00:02:20,460 --> 00:02:21,940 [John] Well, writing, writing code. 00:02:21,940 --> 00:02:23,560 [Eric] By hand-writing code. 00:02:23,560 --> 00:02:27,570 [John] But reading code, I pro- I read more code, I would say. 00:02:27,570 --> 00:02:28,440 [Eric] Right, that makes sense. 00:02:28,440 --> 00:02:30,859 [John] At this point. I mean, that's probably changing, too. 00:02:30,860 --> 00:02:31,120 [Eric] Yep. 00:02:31,120 --> 00:02:38,780 [John] So that... And those are separate things, right? I mean, just like if you were a writer versus, like, an, a reader, and they're related, but they're separate things. 00:02:38,780 --> 00:02:38,800 [Eric] Mm. 00:02:38,800 --> 00:02:38,950 [John] So- 00:02:38,950 --> 00:02:40,329 [Eric] They're definitely separate things 00:02:40,329 --> 00:02:45,720 [John] ... hand, so just the, like, actually hand-typing every line of code, like, that's super different. 00:02:45,720 --> 00:02:46,210 [Eric] Right. 00:02:46,210 --> 00:02:54,299 [John] Um, reading is a little bit different. Probably net, like, read more code because there's more code that to read. 00:02:54,300 --> 00:02:55,480 [Eric] There's more code to read, yeah. 00:02:55,480 --> 00:03:07,940 [John] Um, but even that, like, I think is definitely trending toward, like, um, being just completely results-based. Like, did it work? Having a good plan to, like, test edge cases to make sure it did really work. 00:03:07,940 --> 00:03:08,220 [Eric] Yep. 00:03:08,220 --> 00:03:21,880 [John] And like, that's gonna be way more results-driven. And it is, depending on the situation, 'cause we- 'cause I have, like, two worlds, where sometimes I'm working with a startup, and like, we can, like, go crazy with AI, and other times, it's, it's very legacy, been around for forever. 00:03:21,880 --> 00:03:22,120 [Eric] Right. 00:03:22,120 --> 00:03:24,940 [John] And you, and you, you can't for various reasons. 00:03:24,940 --> 00:03:25,120 [Eric] Right. 00:03:25,120 --> 00:03:26,220 [John] Um- 00:03:26,220 --> 00:03:29,549 [Eric] Have, have your Googling skills atrophied? 00:03:29,549 --> 00:03:29,560 [John] [chuckles] 00:03:29,560 --> 00:03:32,400 [Eric] Because that was a very, that was a very- 00:03:32,400 --> 00:03:32,410 [John] Yeah 00:03:32,410 --> 00:03:36,880 [Eric] ... very valuable skill, uh, in knowledge work. 00:03:36,880 --> 00:03:38,340 [John] Right. Agreed. 00:03:38,340 --> 00:03:41,960 [Eric] And I remember thinking, 00:03:41,960 --> 00:03:45,680 [Eric] "This is a very over," [chuckles] you're gonna love how oversimplified this is- 00:03:45,680 --> 00:03:46,130 [John] Okay, I can't wait 00:03:46,130 --> 00:03:48,460 [Eric] ... in terms of categorizing the world- 00:03:48,460 --> 00:03:48,470 [John] Right 00:03:48,470 --> 00:03:49,600 [Eric] ... of knowledge work, okay? 00:03:49,600 --> 00:03:52,100 [John] Okay. Are you gonna say Googlers and non-Googlers? [chuckles] 00:03:52,100 --> 00:03:54,940 [Eric] It's people who... Everyone used Google. 00:03:54,940 --> 00:03:55,410 [John] Yeah, for sure. 00:03:55,410 --> 00:03:57,280 [Eric] But there were people who were good at it- 00:03:57,280 --> 00:03:57,410 [John] Mm-hmm 00:03:57,410 --> 00:03:59,500 [Eric] ... and then there was everyone else, right? 00:03:59,500 --> 00:03:59,760 [John] Yeah. 00:03:59,760 --> 00:04:02,840 [Eric] And you could just find information way faster- 00:04:02,840 --> 00:04:02,870 [John] Yeah 00:04:02,870 --> 00:04:03,880 [Eric] ... than other people. 00:04:03,880 --> 00:04:10,060 [John] Yeah. Uh, I mean, and just to be specific about it, like, some of my favorites were 00:04:10,060 --> 00:04:11,520 [John] site colon- 00:04:11,520 --> 00:04:11,840 [Eric] Yes 00:04:11,840 --> 00:04:18,140 [John] ... Fill in a site name, and then search, and it, like, limits your search to that. Or filetype pdf, and then, like, you- 00:04:18,140 --> 00:04:18,420 [Eric] Yep 00:04:18,420 --> 00:04:18,979 [John] ... like, just- 00:04:18,980 --> 00:04:22,409 [Eric] Or using, using quotes strategically for different- 00:04:22,409 --> 00:04:22,560 [John] Yeah 00:04:22,560 --> 00:04:23,760 [Eric] ... parts of a query string- 00:04:23,760 --> 00:04:24,300 [John] Yep 00:04:24,300 --> 00:04:26,140 [Eric] ... you know, to force exact match. 00:04:26,140 --> 00:04:26,150 [John] Yep. 00:04:26,150 --> 00:04:27,680 [Eric] And there are tons of tricks like that- 00:04:27,680 --> 00:04:27,830 [John] Yeah, yeah 00:04:27,830 --> 00:04:29,020 [Eric] ... that are extremely helpful. 00:04:29,020 --> 00:04:36,770 [John] I remember reading a short, uh, I mean, I say short, but it was like a 75-page book on, like, Googling one time. [chuckles] 00:04:36,770 --> 00:04:37,820 [Eric] [chuckles] Yeah. I mean- 00:04:37,820 --> 00:04:41,160 [John] And it, and it had every single trick in the book, and it was extremely useful. 00:04:41,160 --> 00:04:41,880 [Eric] Yeah. 00:04:41,880 --> 00:04:46,660 [John] Um, okay, but the, the question is, have my Googling skills atrophied? 00:04:46,660 --> 00:04:47,290 [Eric] Because you- 00:04:47,290 --> 00:04:47,290 [John] Um, because- 00:04:47,290 --> 00:04:52,000 [Eric] ... primarily use LLMs as a, uh, for search, for finding information. 00:04:52,000 --> 00:05:08,360 [John] Um, I mean, I, I definitely don't apply my, you know, Googling skills. Like, I would do it differently. I still like, use Google sometimes, but I find myself more likely clicking the AI option- 00:05:08,360 --> 00:05:08,440 [Eric] Mm-hmm 00:05:08,440 --> 00:05:09,400 [John] ... natively with Google- 00:05:09,400 --> 00:05:09,580 [Eric] Mm-hmm 00:05:09,580 --> 00:05:22,520 [John] ... which I'm sure is what they want. Um, and part of that's probably just muscle memory. Um, and some of it is maybe wanting a little bit faster path to, like, verify information where I'm already in the browser. 00:05:22,520 --> 00:05:22,800 [Eric] Yep. 00:05:22,800 --> 00:05:23,440 [John] You know what I mean? 00:05:23,440 --> 00:05:23,900 [Eric] Yep. 00:05:23,900 --> 00:05:37,480 [John] But, I mean, Perpl- Perplexity has some cool things too, if you wanna, like, take that path. But I think because of that, um, the, the, obviously, like, the random, like, putting quotes around things, like, that's for sure atrophied, but- 00:05:37,480 --> 00:05:37,760 [Eric] Yep 00:05:37,760 --> 00:05:50,260 [John] ... I'm probably, like, less, um... And I, and I can- I search from LLM sometimes, too, depending on what it is. But, but if I wanna, like, quickly get to, like, source material, like, I would still start with probably Google's AI thing or Perplexity. 00:05:50,260 --> 00:05:50,640 [Eric] Mm-hmm. 00:05:50,640 --> 00:05:52,840 [John] So I can quickly get to, like, a source material. 00:05:52,840 --> 00:05:53,159 [Eric] Yep. 00:05:53,160 --> 00:06:08,600 [John] Which I'm finding is becoming more important. Like, say, you're, you're trying to find, like, prompts or, like, skills or something, like, it's bec- like, I pay way more attention now to, like, who made the thing or, like, where it came from than I probably used to. 00:06:08,600 --> 00:06:09,780 [Eric] When you say the thing, what do you mean? 00:06:09,780 --> 00:06:19,723 [John] That I'm looking for. So, like, if it's a bit of code or, like, an example of how to do something-... like, because there's just so much going on with, like, like, skills is a big thing right now- 00:06:19,724 --> 00:06:19,754 [Eric] Mm-hmm 00:06:19,754 --> 00:06:25,364 [John] ... in AI. Like, like, essentially, like, bits of code that can make your AI smarter or better at doing a thing. 00:06:25,364 --> 00:06:25,624 [Eric] Yep. 00:06:25,624 --> 00:06:35,284 [John] Like, there's a ton of, like, security problems with, with random people, like, putting out like a Gmail skill, and it has all this like prompt injection, like, security nightmare stuff in it. 00:06:35,284 --> 00:06:35,783 [Eric] Yep. 00:06:35,784 --> 00:06:41,134 [John] So because of that, like, whereas before I probably would've been less likely to, like, care where it came from- 00:06:41,134 --> 00:06:41,134 [Eric] Yep 00:06:41,134 --> 00:06:53,064 [John] ... and it was just more like, "Eh, like, seems like a lot of people are using this." Like, I'll go... Like, one I found the other day, I went and looked up, like, who created it and where they work, and like, you know, like I actually vetted it, and I would not have gone to that trouble- 00:06:53,064 --> 00:06:53,214 [Eric] Yeah 00:06:53,214 --> 00:06:53,944 [John] ... like in the past. 00:06:53,944 --> 00:06:55,984 [Eric] You're having to be more discerning. 00:06:55,984 --> 00:06:56,704 [John] Yeah. Right. 00:06:56,704 --> 00:06:56,804 [Eric] Interesting. 00:06:56,804 --> 00:07:06,524 [John] So that's a change. But then the other change is like, w- you can be way lazier as to, like, not having to, like, exactly what type or like get the wording perfect. 00:07:06,524 --> 00:07:06,544 [Eric] Mm-hmm. 00:07:06,544 --> 00:07:08,244 [John] Like, you can be lazier and get good results. 00:07:08,244 --> 00:07:10,044 [Eric] Yes. Yeah, I agree. I agree. 00:07:10,044 --> 00:07:14,584 [John] Yeah. What about you? On, on searching for information, let's keep it, like, broad. 00:07:14,584 --> 00:07:26,274 [Eric] I- I've probably moved, I was gonna say 75%, but it's probably more to using LLMs for finding information. 00:07:26,274 --> 00:07:26,344 [John] Right. 00:07:26,344 --> 00:07:35,624 [Eric] And usually, I... Usually, I would say I, I go to Google in two situations. 00:07:35,624 --> 00:07:36,804 [John] Okay. 00:07:36,804 --> 00:07:39,024 [Eric] One is 00:07:39,024 --> 00:07:41,484 [Eric] if I can't find what I'm looking for- 00:07:41,484 --> 00:07:41,724 [John] Hmm 00:07:41,724 --> 00:07:42,844 [Eric] ... in the LLM search- 00:07:42,844 --> 00:07:42,904 [John] Mm 00:07:42,904 --> 00:07:53,644 [Eric] ... which happens sometimes, right? It's like I- okay, I'm not... There's something that I'm not... Something in my prompt is not getting me the type of results that I want- 00:07:53,644 --> 00:07:53,654 [John] Right 00:07:53,654 --> 00:08:03,124 [Eric] ... and it's not citing the right type of information, et cetera. And I'll give you an example of that. I was looking for 00:08:03,124 --> 00:08:09,304 [Eric] a- I was looking for a quote for a blog post, 00:08:09,304 --> 00:08:20,544 [Eric] and I re- I remembered pieces of the quote or the theme of the quote, and but I could not remember the exact quote, and I couldn't remember who... I think it- I thought- 00:08:20,544 --> 00:08:20,554 [John] Mm 00:08:20,554 --> 00:08:24,484 [Eric] ... it was Mark Twain. It actually ended up, he's misquoted as saying it. He actually didn't say it. 00:08:24,484 --> 00:08:25,174 [John] Yep. 00:08:25,174 --> 00:08:25,264 [Eric] But- 00:08:25,264 --> 00:08:27,473 [John] Super common with Mark Twain, [clears throat] I've found, to be honest. 00:08:27,473 --> 00:08:28,734 [Eric] He's so quotable. 00:08:28,734 --> 00:08:28,744 [John] Yep. 00:08:28,744 --> 00:08:29,864 [Eric] He's so quotable. 00:08:29,864 --> 00:08:30,084 [John] Yep. 00:08:30,084 --> 00:08:31,844 [Eric] Um, 00:08:31,844 --> 00:08:33,404 [Eric] so 00:08:33,404 --> 00:08:42,684 [Eric] it- the, the quote is, I'm paraphrasing slightly: "History... " You know, "People say history repeats itself- 00:08:42,684 --> 00:08:43,364 [John] Right 00:08:43,364 --> 00:08:46,244 [Eric] ... That's not exactly true, but it, but history rhymes." 00:08:46,244 --> 00:08:47,193 [John] Yeah, I've heard that- 00:08:47,193 --> 00:08:47,193 [Eric] Yep 00:08:47,193 --> 00:08:48,584 [John] ... attributed to him before. 00:08:48,584 --> 00:08:51,784 [Eric] Um, n- not Mark Twain, but 00:08:51,784 --> 00:08:51,824 [Eric] I- 00:08:51,824 --> 00:08:53,894 [John] But it sounds like it could be. [chuckles] 00:08:53,894 --> 00:08:54,224 [Eric] It... Oh, totally. 00:08:54,224 --> 00:08:54,764 [John] Which is the classic- 00:08:54,764 --> 00:08:56,464 [Eric] Well, he did have a very similar quote 00:08:56,464 --> 00:08:56,484 [John] ... LLM problem. 00:08:56,484 --> 00:08:57,404 [Eric] He had a very similar quote. 00:08:57,404 --> 00:08:58,184 [John] Okay. 00:08:58,184 --> 00:09:00,044 [Eric] Um, 00:09:00,044 --> 00:09:07,454 [Eric] anyways, I just couldn't get the LLM to return results or source material that was helpful. 00:09:07,454 --> 00:09:07,544 [John] Yeah. 00:09:07,604 --> 00:09:10,904 [Eric] And so I ended up going to Google, and I found it, um- 00:09:10,904 --> 00:09:11,544 [John] Okay 00:09:11,544 --> 00:09:13,334 [Eric] ... you know, so that's one example of when I go- 00:09:13,334 --> 00:09:18,884 [John] Do, do you go to Google? Do you reach for Google or for a different LLM when you have that problem typically? 00:09:18,884 --> 00:09:20,084 [Eric] I usually just go to Google. 00:09:20,084 --> 00:09:21,054 [John] Yeah, same. 00:09:21,054 --> 00:09:21,244 [Eric] Uh- 00:09:21,244 --> 00:09:36,124 [John] I, and I realized that I could... 'Cause I was thinking while you were t- like, well, like, back to the original question to me, um, personally, like, I almost never use Google. It's the, just the coding thing, where I'm like, eh, like I really need to, like, make sure like- 00:09:36,124 --> 00:09:36,193 [Eric] Yep 00:09:36,193 --> 00:09:53,144 [John] ... this is not, not gonna be a problem, um, just 'cause of what I do professionally. But personally, like, use, would use some various LLM, and, and honestly, like, Google, uh... What is it? The Google Flash is what my- 00:09:53,144 --> 00:09:53,154 [Eric] Mm-hmm 00:09:53,154 --> 00:10:00,364 [John] ... We both use this software called Raycast on our Macs. Like, my default for Raycast is set to Google Flash- 00:10:00,364 --> 00:10:00,374 [Eric] Mm 00:10:00,374 --> 00:10:03,114 [John] ... like, whatever their thing is, which is great. It's searching, like, as you- 00:10:03,114 --> 00:10:03,114 [Eric] Yep 00:10:03,114 --> 00:10:03,784 [John] ... might imagine. 00:10:03,784 --> 00:10:04,624 [Eric] Yeah. 00:10:04,624 --> 00:10:06,684 [John] And then Perplexity is good- 00:10:06,684 --> 00:10:06,904 [Eric] Mm-hmm 00:10:06,904 --> 00:10:12,124 [John] ... too. And um, and then there's another company called Exa, like E-X-A- 00:10:12,124 --> 00:10:12,334 [Eric] Mm-hmm 00:10:12,334 --> 00:10:31,263 [John] ... that like does some neat stuff. And I haven't done in-depth. Like, I couldn't tell you, like, exactly what I think is better for what. Um, and but anyways, all that to say, especially personally or just if I want some quick information, like I actually do use, just use Google's, like, LLM thing directly, like, from my computer. 00:10:31,263 --> 00:10:32,164 [Eric] Yeah, yeah. 00:10:32,164 --> 00:10:32,784 [John] Um- 00:10:32,784 --> 00:10:37,764 [Eric] I- Google is more efficient for f- if you have a pretty good idea of what you're looking for- 00:10:37,764 --> 00:10:37,894 [John] Mm 00:10:37,894 --> 00:10:39,574 [Eric] ... but you can't remember the exact URL or whatever. 00:10:39,574 --> 00:10:40,164 [John] Exactly. Right. 00:10:40,164 --> 00:10:42,074 [Eric] It's way more efficient than an LLM- 00:10:42,074 --> 00:10:42,074 [John] Right 00:10:42,074 --> 00:10:45,444 [Eric] ... 'cause it has to do the search, it issues the queries, it summarizes the information. 00:10:45,444 --> 00:10:45,844 [John] Yep. 00:10:45,844 --> 00:10:47,364 [Eric] The form factor of the search results- 00:10:47,364 --> 00:10:47,784 [John] Mm-hmm 00:10:47,784 --> 00:10:50,244 [Eric] ... you know, is, is not great in a lot of cases. 00:10:50,244 --> 00:10:50,264 [John] Right. 00:10:50,264 --> 00:10:52,504 [Eric] You have to kind of dig around, you know, for the source material. 00:10:52,504 --> 00:10:53,384 [John] Right. 00:10:53,384 --> 00:10:54,014 [Eric] So yeah- 00:10:54,014 --> 00:10:54,014 [John] Yeah 00:10:54,014 --> 00:10:57,444 [Eric] ... I would definitely say advanced Google skills are atrophying, but- 00:10:57,444 --> 00:10:58,584 [John] Right 00:10:58,584 --> 00:11:09,664 [Eric] ... I would argue, and this is really where I wanted to start digging deep in the conversation. I would argue that 00:11:09,664 --> 00:11:15,744 [Eric] there are n- the way that we find information has fundamentally changed, and so you're building new skills, and so- 00:11:15,744 --> 00:11:15,884 [John] Yes 00:11:15,884 --> 00:11:21,624 [Eric] ... I don't think that the atrophy of those advanced Google skills is a bad thing- 00:11:21,624 --> 00:11:21,664 [John] Right 00:11:21,664 --> 00:11:24,484 [Eric] ... because I don't think they're going to be necessary anymore. 00:11:24,484 --> 00:11:29,023 [John] Well, and, and I think I'm starting to think about the skill atrophy in buckets. 00:11:29,024 --> 00:11:29,444 [Eric] Hmm. 00:11:29,444 --> 00:11:41,084 [John] One is, was the skill always just a means to a mean anyways? Or like, bucket n- that's bucket number one. Bucket number two for me, is this like a core fundamental skill? 00:11:41,084 --> 00:11:42,324 [Eric] Okay, give me an example- 00:11:42,324 --> 00:11:42,374 [John] Um 00:11:42,374 --> 00:11:42,984 [Eric] ... of each of those. 00:11:42,984 --> 00:11:43,224 [John] Yeah. 00:11:43,224 --> 00:11:44,504 [Eric] So a means to an end, and then core skill. 00:11:44,504 --> 00:11:49,324 [John] Means to an end, um, the perfect example is what we've been talking about is Google. 00:11:49,324 --> 00:11:49,664 [Eric] Yep. 00:11:49,664 --> 00:11:51,784 [John] Like, let me Google something. Like, it's super means to an end. 00:11:51,784 --> 00:11:52,434 [Eric] Mm-hmm. 00:11:52,434 --> 00:11:54,994 [John] Like, I just need the end result. I do not care how I get there. 00:11:54,994 --> 00:11:55,004 [Eric] Yep. 00:11:55,004 --> 00:11:56,564 [John] I just want it to be efficient and fast and- 00:11:56,564 --> 00:11:56,784 [Eric] Yep 00:11:56,784 --> 00:12:04,264 [John] ... whatever. Um, one that's more in the middle, and I think moving to a means to an end, is coding in general. 00:12:04,264 --> 00:12:04,504 [Eric] Mm-hmm. 00:12:04,504 --> 00:12:15,384 [John] Like, like, every- like, on a normal... A- and you know what's, uh, helpful actually to, in thinking about this? Is thinking about how you work profess- how you work professionally versus live like your personal life. 00:12:15,384 --> 00:12:16,104 [Eric] Hmm. 00:12:16,488 --> 00:12:25,048 [John] ... if there's no crossover. So, for example, like, I speak in English in my personal life and professionally. Like, maybe I use slightly different words, but, like, it's kinda there. 00:12:25,048 --> 00:12:25,528 [Eric] Mm-hmm. 00:12:25,528 --> 00:12:28,388 [John] Like, very core skill. But, 00:12:28,388 --> 00:12:37,408 [John] and, and coding is kinda one of those things where, like, I don't personally, I, I probably do more than I used to, but, like, I don't personally do a lot of, like, coding for fun. It's just kinda it's my job. 00:12:37,408 --> 00:12:38,048 [Eric] Right. 00:12:38,048 --> 00:12:40,568 [John] Um, and that one for me 00:12:40,568 --> 00:12:42,008 [John] is more of a means to an end thing- 00:12:42,008 --> 00:12:42,418 [Eric] Mm 00:12:42,418 --> 00:12:47,128 [John] ... of like, if I could just, like, say English and get what I wanted, then, like- 00:12:47,128 --> 00:12:47,228 [Eric] Yeah 00:12:47,228 --> 00:12:48,468 [John] ... I would prefer to do that. 00:12:48,468 --> 00:12:48,587 [Eric] Yeah. 00:12:48,587 --> 00:12:54,728 [John] I think there's at least, there are some people that are split on that, that, like, really just love the process, and like, that's not their preference- 00:12:54,728 --> 00:12:54,948 [Eric] Yep 00:12:54,948 --> 00:13:01,428 [John] ... for coding. Um, but from a business mindset, like, every business owner, like, would, would prefer- 00:13:01,428 --> 00:13:01,478 [Eric] Totally 00:13:01,478 --> 00:13:02,668 [John] ... for them to say English and code to be written. 00:13:02,668 --> 00:13:03,268 [Eric] Totally. Yep. 00:13:03,268 --> 00:13:05,998 [John] And, and whether there's step- there's still steps in between, but maybe- 00:13:05,998 --> 00:13:05,998 [Eric] Yep 00:13:05,998 --> 00:13:07,318 [John] ... like, long-term it's not. 00:13:07,318 --> 00:13:07,348 [Eric] Yep. 00:13:07,348 --> 00:13:14,028 [John] So I think that one's, like, nuanced, but long-term, like, it was pretty much an ends to a mean, at least from, like, the highest level of an organisation- 00:13:14,028 --> 00:13:14,038 [Eric] Yep 00:13:14,038 --> 00:13:24,087 [John] ... always an ends to a mean. I think, so those are my bucket one. And then bucket two of, like, core skills for me, I think, I think 00:13:24,088 --> 00:13:38,768 [John] communication stays as a core skill. Writing, speaking, um, w- having organised thoughts, you know, all of that, like, stays as a core skill, and but could atrophy depending on how you're using AI. 00:13:38,768 --> 00:13:38,788 [Eric] Mm-hmm. 00:13:38,788 --> 00:13:44,428 [John] Like, that's my... And we'll talk more about that, I think, in a minute. So that's one of my, like, big buckets of core skills. 00:13:44,428 --> 00:13:45,188 [Eric] Yep. 00:13:45,188 --> 00:13:46,848 [John] And then 00:13:46,848 --> 00:13:52,228 [John] closely related to that would be critical thinking, strategic thinking, like that- 00:13:52,228 --> 00:13:52,238 [Eric] Yeah 00:13:52,238 --> 00:13:53,108 [John] ... that skill. 00:13:53,108 --> 00:13:53,518 [Eric] Yep. 00:13:53,518 --> 00:13:58,628 [John] Which could also atrophy because more and more you could outsource that to AI. 00:13:58,628 --> 00:13:59,188 [Eric] Yep. 00:13:59,188 --> 00:14:05,118 [John] Um, and then, and then my last one would be something around, like, creativity. Like- 00:14:05,118 --> 00:14:05,118 [Eric] Mm 00:14:05,118 --> 00:14:14,928 [John] ... creative problem-solving, creative thinking is another one where you could outsource to AI, and, and that one can be used for good or for bad, right? AI can kind of expand your creativity- 00:14:14,928 --> 00:14:14,978 [Eric] Mm-hmm 00:14:14,978 --> 00:14:19,348 [John] ... or it can, it can limit it 'cause you're just like, I don't know, like, whatever it says, we'll do that. 00:14:19,348 --> 00:14:19,648 [Eric] Yep. Yep. 00:14:19,648 --> 00:14:20,068 [John] So. 00:14:20,068 --> 00:14:20,348 [Eric] Yeah. 00:14:20,348 --> 00:14:28,668 [John] What, what... Yeah, what would you add to that, or how would you modify the buckets for you personally? 00:14:28,668 --> 00:14:32,248 [Eric] I think that 00:14:32,248 --> 00:14:36,948 [Eric] I, I would probably s- I, I agree on fundamental skills 00:14:36,948 --> 00:14:40,828 [Eric] and just light commentary on the, on the core fundamentals. 00:14:40,828 --> 00:14:42,028 [John] Mm-hmm. 00:14:42,028 --> 00:14:48,408 [Eric] I think that there's way more danger in letting those atrophy. 00:14:48,408 --> 00:14:49,228 [John] Okay. Yeah. 00:14:49,228 --> 00:14:51,288 [Eric] Um, 00:14:51,288 --> 00:14:56,238 [Eric] because I think, like, the lived human experience- 00:14:56,238 --> 00:14:56,238 [John] Right 00:14:56,238 --> 00:15:02,288 [Eric] ... is a really, it, it is the most valuable input to those core skills. 00:15:02,288 --> 00:15:03,068 [John] Okay. 00:15:03,068 --> 00:15:09,268 [Eric] Right? And so, um, but that takes critical thinking and reflection and processing. 00:15:09,268 --> 00:15:09,968 [John] Right. 00:15:09,968 --> 00:15:19,108 [Eric] And I think that if you may-- if you don't let those atrophy, your use of AI will be, um, you'll get more leverage out of AI. 00:15:19,108 --> 00:15:19,668 [John] Right. 00:15:19,668 --> 00:15:23,148 [Eric] Um, because it's an augmentation, not a replacement. 00:15:23,148 --> 00:15:24,048 [John] Right. 00:15:24,048 --> 00:15:29,808 [Eric] The means to an end bucket, I would, um, 00:15:29,808 --> 00:15:36,508 [Eric] I would probably split that bucket, and 00:15:36,508 --> 00:15:41,588 [Eric] there are, um, 00:15:41,588 --> 00:15:43,928 [Eric] there are 00:15:43,928 --> 00:15:56,268 [Eric] sort of... Yeah, I would split the bucket into two, and the way that I'm thinking about it off the cuff here is that you have sort of automation of, um, 00:15:56,268 --> 00:16:01,028 [Eric] you have automation of things that are not super high-value work, right? So- 00:16:01,028 --> 00:16:01,308 [John] Sure 00:16:01,308 --> 00:16:05,108 [Eric] ... for example, um, 00:16:05,108 --> 00:16:10,398 [Eric] I'll give you an example of a, of a skill that is, that is, I'm very thankful is atrophying for me. 00:16:10,398 --> 00:16:10,698 [John] Okay. All right. 00:16:10,698 --> 00:16:12,108 [Eric] Um, 00:16:12,108 --> 00:16:13,318 [Eric] formatting documents- [chuckles] 00:16:13,318 --> 00:16:13,628 [John] [chuckles] 00:16:13,628 --> 00:16:15,108 [Eric] ... very quickly. So I mean- 00:16:15,108 --> 00:16:15,208 [John] Yeah 00:16:15,208 --> 00:16:16,068 [Eric] ... I have- 00:16:16,068 --> 00:16:16,388 [John] Sure 00:16:16,388 --> 00:16:28,668 [Eric] ... I literally have keys, keyboard shortcuts and keys remapped on my keyboard. Pr- like, for one of the primary purposes is to navigate a text document, copy, cut- 00:16:28,668 --> 00:16:28,678 [John] Okay 00:16:28,678 --> 00:16:30,498 [Eric] ... replace text without touching the mouse- 00:16:30,498 --> 00:16:30,498 [John] Okay 00:16:30,498 --> 00:16:33,688 [Eric] ... and just doing it all. My key, my key repeat rate- 00:16:33,688 --> 00:16:34,048 [John] Yes 00:16:34,048 --> 00:16:37,468 [Eric] ... is, I think, the fastest that you can configure it from the command line. 00:16:37,468 --> 00:16:37,708 [John] Okay. 00:16:37,708 --> 00:16:38,347 [Eric] Like, [chuckles] 00:16:38,347 --> 00:16:40,618 [John] But, but you're saying that's atrophy? 00:16:40,618 --> 00:16:42,208 [Eric] I just don't use it as much anymore- 00:16:42,208 --> 00:16:42,218 [John] Okay, sure 00:16:42,218 --> 00:16:45,558 [Eric] ... because formatting a text document is per- 00:16:45,558 --> 00:16:45,558 [John] Trivial, yeah 00:16:45,558 --> 00:16:46,928 [Eric] ... it's such a great job for- 00:16:46,928 --> 00:16:47,188 [John] Right 00:16:47,188 --> 00:16:48,348 [Eric] ... for an LLM, right? 00:16:48,348 --> 00:16:48,708 [John] Right. 00:16:48,708 --> 00:16:52,888 [Eric] And so there's that sort of thing, where you sort of build up these... It's a means to an end, right? 00:16:52,888 --> 00:16:52,928 [John] Yeah. 00:16:52,928 --> 00:16:54,878 [Eric] I just want to get a document, like, reformatted, right? 00:16:54,878 --> 00:16:54,888 [John] Yep. 00:16:54,888 --> 00:17:03,148 [Eric] And so you can sort of throw that away. But then I do think there are other, um... And I'm trying to think of a good example. 00:17:03,148 --> 00:17:03,648 [John] Yeah. 00:17:03,648 --> 00:17:07,048 [Eric] But there are other, there are other 00:17:07,048 --> 00:17:11,988 [Eric] skills where the process, I think, is an important part of it. 00:17:11,988 --> 00:17:12,768 [John] Right. 00:17:12,768 --> 00:17:13,848 [Eric] Um- 00:17:13,848 --> 00:17:19,308 [John] Okay, yeah, I see, I see what you mean. Like, writing, writing is, like, an obvious example, at least to me, of that. 00:17:19,308 --> 00:17:19,528 [Eric] Yep. 00:17:19,528 --> 00:17:21,608 [John] But there's, there's others too. 00:17:21,608 --> 00:17:22,068 [Eric] Yes. 00:17:22,068 --> 00:17:32,268 [John] Which I'll- the other side of the coding piece is, I mean, there is an argument that understanding how the code works is part of the process- 00:17:32,268 --> 00:17:32,498 [Eric] Mm-hmm 00:17:32,498 --> 00:17:58,188 [John] ... um, and I think should be and will be for a while. Like, actually, we, we talked about this with a group, um, last week, and there's a just really expert-level developer, oper- like, um, technical, um, DevOps person, IT operations person that, that was saying how, you know, they're writing less code, but, um, they're, they're still reading a lot of code, finding things that they would architect differently. 00:17:58,188 --> 00:17:58,228 [Eric] Yep. 00:17:58,228 --> 00:18:13,068 [John] So they're essentially, like, kind of have an architect hat on. And the most interesting thing that he brought up was that the particular thing he was working on, code was fine, code worked, architecture was fine, no, like, scalability or bad problems- 00:18:13,068 --> 00:18:13,078 [Eric] Mm-hmm 00:18:13,078 --> 00:18:22,712 [John] ... or even security problems. But he said, "I, um... " All of that being the case-... I knew how to do it 10 times cheaper than it was planning to do it. 00:18:22,712 --> 00:18:23,531 [Eric] That's right, yeah. 00:18:23,532 --> 00:18:24,212 [John] Which is interesting- 00:18:24,212 --> 00:18:24,392 [Eric] Yeah 00:18:24,392 --> 00:18:24,982 [John] ... right? 'Cause like- 00:18:24,982 --> 00:18:27,312 [Eric] Claude Code, it was gonna be... I remember that. He said- 00:18:27,312 --> 00:18:28,122 [John] Yeah, whatever the future thing was. 00:18:28,122 --> 00:18:30,502 [Eric] It was gonna be five grand a month or 10 grand a month- 00:18:30,502 --> 00:18:30,502 [John] Yeah 00:18:30,502 --> 00:18:30,772 [Eric] ... and he's like- 00:18:30,772 --> 00:18:32,002 [John] Which is, like, a couple hundred bucks whatever. 00:18:32,002 --> 00:18:32,992 [Eric] Yeah, a couple hundred dollars, yeah. 00:18:32,992 --> 00:18:37,732 [John] Yeah, so that, that's a, that is a thing where, like, 00:18:37,792 --> 00:18:43,432 [John] uh, probably will, like, over time, like, get better, but a very interesting use case where 00:18:43,432 --> 00:18:44,572 [John] nobody would know. 00:18:44,572 --> 00:18:44,592 [Eric] Mm-hmm. 00:18:44,592 --> 00:18:53,512 [John] Like, there's not a security problem. There's... A- And from Claude's perspective, like, it didn't... You didn't say to optimize for cost. Like, it, like, it- no security problem, no this, no that. 00:18:53,512 --> 00:18:53,702 [Eric] Yep. 00:18:53,702 --> 00:18:56,112 [John] Like, it's fine. It's expensive. Like, who cares? 00:18:56,112 --> 00:18:57,041 [Eric] Yeah. [chuckles] Yeah, yeah. 00:18:57,041 --> 00:18:57,552 [John] You know, from, from their- 00:18:57,552 --> 00:18:57,732 [Eric] Yeah 00:18:57,732 --> 00:19:00,392 [John] ... perspective. So it's interesting. 00:19:00,392 --> 00:19:05,032 [Eric] You mentioned, you mentioned writing in... I mean, I write for my job. 00:19:05,032 --> 00:19:05,312 [John] Right. Right. 00:19:05,312 --> 00:19:08,792 [Eric] Uh, and I mean, I have for a long time. It just now it's- 00:19:08,792 --> 00:19:09,012 [John] Yeah 00:19:09,012 --> 00:19:09,692 [Eric] ... a job title. 00:19:09,692 --> 00:19:09,872 [John] Right. 00:19:09,872 --> 00:19:10,552 [Eric] Uh, I guess it's how I earn a living. 00:19:10,552 --> 00:19:16,172 [John] Full circle for you, right? I think early on in your career, you, you, you, um, did- 00:19:16,172 --> 00:19:17,032 [Eric] I did a lot of writing- 00:19:17,032 --> 00:19:17,062 [John] Yeah 00:19:17,062 --> 00:19:17,792 [Eric] ... in an agency, yeah. 00:19:17,792 --> 00:19:18,632 [John] Yeah, that's what I mean. Yeah. 00:19:18,632 --> 00:19:18,792 [Eric] Yep. 00:19:18,792 --> 00:19:19,192 [John] Right. 00:19:19,192 --> 00:19:19,772 [Eric] Um- 00:19:19,772 --> 00:19:20,832 [John] So that's kinda, kinda interesting. 00:19:20,832 --> 00:19:21,712 [Eric] I haven't thought about that. 00:19:21,712 --> 00:19:21,812 [John] Mm. 00:19:21,812 --> 00:19:22,872 [Eric] That is very cool. 00:19:22,872 --> 00:19:22,952 [John] Yeah. 00:19:22,952 --> 00:19:34,792 [Eric] Full circle. Uh, the... What, this is what's interesting about that because, uh, I- it's hard to put writing in those two buckets that you mentioned. 00:19:34,792 --> 00:19:35,242 [John] Okay. 00:19:35,242 --> 00:19:36,732 [Eric] And I'll tell you why for me- 00:19:36,732 --> 00:19:36,982 [John] Yeah 00:19:36,982 --> 00:19:38,892 [Eric] ... personally, and how I think about atrophy. 00:19:38,892 --> 00:19:39,492 [John] Mm-hmm. 00:19:39,492 --> 00:19:43,122 [Eric] Because I, I deal with this every single day. 00:19:43,122 --> 00:19:43,132 [John] Right. 00:19:43,132 --> 00:19:52,412 [Eric] Deal with is probably not, not the right word. I, I, I experience the impact of AI on writing as a skill- 00:19:52,412 --> 00:19:52,492 [John] Right 00:19:52,492 --> 00:19:53,312 [Eric] ... every single day. 00:19:53,312 --> 00:19:54,292 [John] Right. 00:19:54,292 --> 00:19:56,492 [Eric] So 00:19:56,552 --> 00:20:05,692 [Eric] the, the main thing that I would say is I actually think that good writing is r- is good thinking. 00:20:05,692 --> 00:20:06,612 [John] Sure. 00:20:06,612 --> 00:20:16,692 [Eric] Uh, I mean, y- you know, you can... Uh, y- someone could have... Someone could write very elegant prose that's 00:20:16,692 --> 00:20:17,712 [Eric] really bad thinking, right? 00:20:17,712 --> 00:20:17,852 [John] Yeah. 00:20:17,852 --> 00:20:18,812 [Eric] It's possible, right? 00:20:18,812 --> 00:20:19,072 [John] Yeah. 00:20:19,072 --> 00:20:24,492 [Eric] So when I say good writing, I sort of mean objectively, objectively good writing- 00:20:24,492 --> 00:20:24,932 [John] Right 00:20:24,932 --> 00:20:29,472 [Eric] ... um, is ultimately just the manifestation of good thinking. 00:20:29,472 --> 00:20:30,451 [John] Yeah. Right. 00:20:30,451 --> 00:20:32,212 [Eric] And so 00:20:32,212 --> 00:20:35,732 [Eric] I can't separate those two in my- 00:20:35,732 --> 00:20:35,741 [John] Sure 00:20:35,741 --> 00:20:37,232 [Eric] ... like, mental model for writing. 00:20:37,232 --> 00:20:37,482 [John] Right. 00:20:37,482 --> 00:20:38,892 [Eric] Um, 00:20:38,892 --> 00:20:41,892 [Eric] so I view writing as a core skill. 00:20:41,892 --> 00:20:41,991 [John] Right. 00:20:41,992 --> 00:20:52,292 [Eric] Um, the thinking part is most important, but then there are, let's say, like, the tactical components of style- 00:20:52,292 --> 00:20:52,512 [John] Yeah 00:20:52,512 --> 00:20:53,392 [Eric] ... grammar- 00:20:53,392 --> 00:20:54,082 [John] Mm-hmm 00:20:54,082 --> 00:20:54,832 [Eric] ... narrative. 00:20:54,832 --> 00:20:54,882 [John] Right 00:20:54,882 --> 00:20:59,032 [Eric] Like, the, you know, sort of taking those thoughts and actually, you know, materializing them on a page. 00:20:59,032 --> 00:20:59,401 [John] Right. 00:20:59,401 --> 00:21:05,392 [Eric] And there's skill there as well. Um, and 00:21:05,392 --> 00:21:12,412 [Eric] w- what's been interesting for me is that... So I actually recently made a decision 00:21:12,412 --> 00:21:16,972 [Eric] on where I will use AI in writing, and this is- 00:21:16,972 --> 00:21:17,012 [John] Okay 00:21:17,012 --> 00:21:18,592 [Eric] ... somewhat of a human experiment, so- 00:21:18,592 --> 00:21:19,112 [John] Right 00:21:19,112 --> 00:21:20,632 [Eric] ... we'll have to check back in on this- 00:21:20,632 --> 00:21:21,312 [John] Nice, yeah 00:21:21,312 --> 00:21:33,681 [Eric] ... uh, you know, in a couple weeks or months. So I use AI for writing at Vercel extremely heavily. I mean, I don't- 00:21:33,681 --> 00:21:33,681 [John] Sure 00:21:33,681 --> 00:21:41,972 [Eric] ... I rarely create any content without using AI as part of the process. As, like, a core part of the process. 00:21:41,972 --> 00:21:42,852 [John] Right. 00:21:42,852 --> 00:21:45,992 [Eric] And that's true for most people on the team, and people have different methodologies. I mean- 00:21:45,992 --> 00:21:46,172 [John] Right 00:21:46,172 --> 00:21:53,832 [Eric] ... one of the, uh, writers on the team I work with actually y- exclusively uses Claude Code with a local file system. 00:21:53,832 --> 00:21:54,152 [John] Sure. 00:21:54,152 --> 00:21:54,532 [Eric] You know, and- 00:21:54,532 --> 00:21:54,892 [John] Yeah 00:21:54,892 --> 00:21:56,252 [Eric] ... [chuckles] it's pretty awesome. 00:21:56,252 --> 00:21:56,852 [John] Yeah, yeah. 00:21:56,852 --> 00:22:11,442 [Eric] Uh, and I've tried a bunch of different methodologies, which is probably, you know, fun to talk about on another episode. But, you know, the, the writing team is a very, like... It's a very skilled, experienced group of people. 00:22:11,442 --> 00:22:11,452 [John] Right. 00:22:11,452 --> 00:22:13,452 [Eric] And so we all use AI like crazy- 00:22:13,452 --> 00:22:13,672 [John] Right 00:22:13,672 --> 00:22:15,872 [Eric] ... but there's a final human pass on it- 00:22:15,872 --> 00:22:15,952 [John] Right 00:22:15,952 --> 00:22:18,312 [Eric] ... um, which is very critical. I mean- 00:22:18,312 --> 00:22:18,741 [John] Yeah, right 00:22:18,741 --> 00:22:21,392 [Eric] ... we don't publish anything without that. 00:22:21,392 --> 00:22:21,432 [John] Right. 00:22:21,432 --> 00:22:22,412 [Eric] Um- 00:22:22,412 --> 00:22:26,652 [John] Well, and you say final human pass, like, a lot of humans- 00:22:26,652 --> 00:22:26,912 [Eric] Yes 00:22:26,912 --> 00:22:27,612 [John] ... going over- 00:22:27,612 --> 00:22:28,672 [Eric] Multiple humans. 00:22:28,672 --> 00:22:28,892 [John] Yeah. 00:22:28,892 --> 00:22:29,112 [Eric] Yeah, yeah, yeah. 00:22:29,112 --> 00:22:29,122 [John] Right. 00:22:29,122 --> 00:22:30,552 [Eric] It goes... The, the review process- 00:22:30,552 --> 00:22:30,652 [John] Yeah 00:22:30,652 --> 00:22:31,241 [Eric] ... is very extensive. 00:22:31,241 --> 00:22:32,112 [John] Different, yeah. 00:22:32,112 --> 00:22:36,532 [Eric] Um, but the rate at which you can 00:22:36,532 --> 00:22:41,712 [Eric] translate your thoughts into drafts and iterate- 00:22:41,712 --> 00:22:41,742 [John] Right 00:22:41,742 --> 00:22:46,992 [Eric] ... especially in a group, AI is such an extreme accelerant to that. 00:22:46,992 --> 00:22:47,152 [John] Right. 00:22:47,152 --> 00:22:49,632 [Eric] Um, or find the right information or- 00:22:49,632 --> 00:22:50,092 [John] Mm-hmm 00:22:50,152 --> 00:22:52,722 [Eric] ... double-check code samples that you're putting into a post. 00:22:52,722 --> 00:22:52,732 [John] Yeah. 00:22:52,732 --> 00:22:54,212 [Eric] I mean, all that sort of stuff. 00:22:54,212 --> 00:22:54,792 [John] Yeah. 00:22:54,792 --> 00:22:54,972 [Eric] Um- 00:22:54,972 --> 00:22:59,032 [John] I hadn't even thought of that, 'cause there's so many, like, bugs and docs out there- 00:22:59,032 --> 00:22:59,272 [Eric] Mm-hmm 00:22:59,332 --> 00:23:04,712 [John] ... with, like, code samples, and [chuckles] having an LLM be like, "Hey, actually, can you grab this and compile it and see if it works?" 00:23:04,712 --> 00:23:05,432 [Eric] Totally. I mean- 00:23:05,432 --> 00:23:06,032 [John] It's super cool. 00:23:06,032 --> 00:23:06,802 [Eric] It's awesome. 00:23:06,802 --> 00:23:06,891 [John] Yeah. 00:23:06,892 --> 00:23:16,902 [Eric] Uh, but I... But there is... And, I mean, it's a commercial, it's a bus- it's a commercial business, right? 00:23:16,902 --> 00:23:16,912 [John] Right. 00:23:16,912 --> 00:23:17,832 [Eric] And so- 00:23:17,832 --> 00:23:17,952 [John] Right 00:23:17,952 --> 00:23:20,152 [Eric] ... you know, it's not, you're not writing a novel. 00:23:20,152 --> 00:23:20,202 [John] Yeah. 00:23:20,202 --> 00:23:20,852 [Eric] Um- 00:23:20,852 --> 00:23:21,592 [John] Right 00:23:21,592 --> 00:23:26,532 [Eric] ... you're producing content as part of a, like, a commercial entity. But 00:23:26,532 --> 00:23:36,712 [Eric] I noticed that my tendency, which is a, I think a very common thing for humans, is to follow the path of least resistance. And when you run into a- 00:23:36,712 --> 00:23:36,982 [John] Right 00:23:36,982 --> 00:23:38,432 [Eric] ... a friction in writing- 00:23:38,432 --> 00:23:38,492 [John] Right 00:23:38,492 --> 00:23:43,032 [Eric] ... which happens all the time, you just use AI to break through the barrier. 00:23:43,032 --> 00:23:43,472 [John] Yep. 00:23:43,472 --> 00:23:48,032 [Eric] And that is really, really helpful in a lot of situations. But- 00:23:48,032 --> 00:23:53,812 [John] You're talking like what traditionally, like, called writer's block, or, or do you mean, like, something else? 00:23:53,812 --> 00:24:03,951 [Eric] So let's just... Here's a very specific example. So I wrote a post, I think it was published this week. We'll put it in the show notes. 00:24:03,952 --> 00:24:04,732 [John] Sure. 00:24:04,732 --> 00:24:09,872 [Eric] And the, um, 00:24:09,872 --> 00:24:24,000 [Eric] one of the tricky things about the post was, you sort of define a narrative arc for it, and sometimes it's really difficult to determine how d-... important pieces fit into the narrative arc. 00:24:24,000 --> 00:24:24,040 [John] Okay. 00:24:24,040 --> 00:24:27,640 [Eric] So I have data, or I have examples- 00:24:27,640 --> 00:24:27,660 [John] Mm. 00:24:27,660 --> 00:24:28,140 [Eric] -that really- 00:24:28,140 --> 00:24:28,210 [John] Right 00:24:28,210 --> 00:24:36,700 [Eric] ... prove the point, but sometimes you need to actually do major surgery and restructure- 00:24:36,700 --> 00:24:37,700 [John] Mm-hmm 00:24:37,700 --> 00:24:46,250 [Eric] ... restructure the post in order to include information that's going to make the, make the narrative more compelling overall. 00:24:46,250 --> 00:24:46,360 [John] Yes. Sure. 00:24:46,360 --> 00:24:49,720 [Eric] Um, that's a very tricky thing. 00:24:49,720 --> 00:24:50,120 [John] Yeah. 00:24:50,120 --> 00:24:54,090 [Eric] Um, that's a very tricky thing to solve in a concise way- 00:24:54,090 --> 00:24:54,090 [John] Right 00:24:54,090 --> 00:24:59,280 [Eric] ... 'cause you can just make something long and sort of like, you know, solve it through brute force- 00:24:59,280 --> 00:24:59,290 [John] Yeah 00:24:59,290 --> 00:25:04,380 [Eric] ... you know, or whatever. But doing that in a concise way is very difficult. And 00:25:04,380 --> 00:25:12,840 [Eric] on the one hand, the... On the one hand, it's very helpful to use AI to, like, rapidly iterate through- 00:25:12,840 --> 00:25:12,970 [John] Right 00:25:12,970 --> 00:25:18,460 [Eric] ... multiple types of restructuring, et cetera. But 00:25:18,460 --> 00:25:25,460 [Eric] a- and, and I definitely do that sometimes, but I thought that's actually a skill that because 00:25:25,460 --> 00:25:28,480 [Eric] th- that particular content is written for humans to read- 00:25:28,480 --> 00:25:28,760 [John] Mm-hmm 00:25:28,760 --> 00:25:31,219 [Eric] ... which is different than, I would say, documentation- 00:25:31,220 --> 00:25:31,460 [John] Right 00:25:31,460 --> 00:25:32,750 [Eric] ... and, you know, content is changing, right? 00:25:32,750 --> 00:25:32,760 [John] Right. 00:25:32,760 --> 00:25:38,020 [Eric] Where a lot of it's going to be read by agents, right? But this content was specifically for humans to read. 00:25:38,020 --> 00:25:38,900 [John] Right. 00:25:38,900 --> 00:25:40,520 [Eric] I thought, okay, 00:25:40,520 --> 00:25:47,300 [Eric] the narrat- understanding the audience and the narrative and the interpretation of this 00:25:47,300 --> 00:25:50,340 [Eric] is a really critical writing skill, and I don't want- 00:25:50,340 --> 00:25:50,430 [John] Yeah 00:25:50,430 --> 00:25:52,080 [Eric] ... that to atrophy. 00:25:52,080 --> 00:25:52,680 [John] Okay. 00:25:52,680 --> 00:25:59,079 [Eric] But, uh, it's tricky because AI is very, very helpful in sort of rapidly iterating through that. 00:25:59,080 --> 00:25:59,200 [John] Right. 00:25:59,200 --> 00:26:08,680 [Eric] So my solve, that was a very long way of saying my solve for this is that I don't use AI at all, writing on my personal blog. 00:26:08,680 --> 00:26:09,280 [John] Nice. Yeah. 00:26:09,280 --> 00:26:21,860 [Eric] So end-to-end, my posts start in a markdown document, and I don't use AI at all. The only way that I use AI is to search for links or other things like that. 00:26:21,860 --> 00:26:22,160 [John] Sure. 00:26:22,160 --> 00:26:28,680 [Eric] But... And that's actually, I mean, earlier this summer, I wrote several posts that were like, that were augmented by AI, and I- 00:26:28,680 --> 00:26:28,690 [John] Sure 00:26:28,690 --> 00:26:33,300 [Eric] ... sort of used it, you know, pretty heavily in the process for editing and other things like that. 00:26:33,300 --> 00:26:33,760 [John] Right. 00:26:33,760 --> 00:26:39,860 [Eric] Um, but it is just, let's call it bare metal, you know, where it's just me and a markdown editor- 00:26:39,860 --> 00:26:40,200 [John] Right 00:26:40,200 --> 00:26:40,869 [Eric] ... and that's it- 00:26:40,869 --> 00:26:40,869 [John] Right 00:26:40,869 --> 00:26:53,320 [Eric] ... end to end. And what I have found, at least initially, is that it has made me, it, it has made my use of AI far more effective at work. 00:26:53,320 --> 00:26:53,820 [John] Hmm. 00:26:53,820 --> 00:26:55,340 [Eric] Like, I think I'm way faster at- 00:26:55,340 --> 00:26:55,400 [John] Yeah 00:26:55,400 --> 00:26:57,380 [Eric] ... using AI to edit documents. 00:26:57,380 --> 00:26:58,060 [John] I believe that. 00:26:58,060 --> 00:27:01,420 [Eric] Because I think that 00:27:01,420 --> 00:27:03,600 [Eric] I can 00:27:03,600 --> 00:27:17,089 [Eric] more... W- w- working those muscles, I can sort of see what needs to happen more quickly. Um, and so I can be more specific with AI as opposed to going through, like, multiple rounds of iteration and sort of offloading that, if that makes sense. 00:27:17,089 --> 00:27:27,420 [John] Right. That's super interesting. I think I've experienced something a little bit similar but different with, um, AI's amazing ability to take audio to text. 00:27:27,420 --> 00:27:28,780 [Eric] Mm. 00:27:28,780 --> 00:27:32,320 [John] Which we talked about this earlier today. We talked about dragon dictation. 00:27:32,320 --> 00:27:33,470 [Eric] Dragon dictation- 00:27:33,470 --> 00:27:33,540 [John] Been around- 00:27:33,540 --> 00:27:35,140 [Eric] That's definitely going in the show notes 00:27:35,140 --> 00:27:37,880 [John] ... 15, 20 years. [laughing] Um, and they, they tried- 00:27:37,880 --> 00:27:38,000 [Eric] Yeah 00:27:38,000 --> 00:27:42,500 [John] ... to solve this problem and, and honestly, like, I didn't realise this, but 00:27:42,500 --> 00:27:49,590 [John] was, like, had gotten pretty good even before all the AI stuff. And then a lot of these companies popped up again solving the same problem- 00:27:49,590 --> 00:27:49,600 [Eric] Mm-hmm 00:27:49,600 --> 00:28:01,600 [John] ... but with AI, and it's even better now. And, um, and a, a lot of the pitch from the companies, if you go to, like, WhisperFlow's website, um, is, yeah, whatever the average typing speed is of a person- 00:28:01,600 --> 00:28:01,610 [Eric] Yep 00:28:01,610 --> 00:28:04,020 [John] ... like, you can talk four to three times faster- 00:28:04,020 --> 00:28:04,050 [Eric] Yep 00:28:04,050 --> 00:28:04,880 [John] ... two times faster. 00:28:04,880 --> 00:28:05,680 [Eric] Yep. 00:28:05,680 --> 00:28:17,440 [John] Um, true for most, you know, most people. Um, so I was s- so I played around with them and was kind of like, eh, like, this is fine, but really, like, more recently have, have realised that 00:28:17,440 --> 00:28:22,760 [John] at l- at least for me, and, and part of this is, like, job-wise, 00:28:22,760 --> 00:28:30,080 [John] for the last, like, however many years, I've been in, like, way more meetings talking, and then when I'm typing, it's, like, more likely, like, short form- 00:28:30,080 --> 00:28:30,180 [Eric] Mm-hmm 00:28:30,180 --> 00:28:37,660 [John] ... Slack or maybe, like, coding. Like, do... Like, I don't write long-form emails, like, as part of my job, and I really haven't in a long- 00:28:37,660 --> 00:28:38,040 [Eric] Neither do I 00:28:38,040 --> 00:28:38,559 [John] ... a long time. 00:28:38,559 --> 00:28:39,720 [Eric] And the data shows that. 00:28:39,720 --> 00:28:40,340 [John] Yeah. [chuckles] 00:28:40,340 --> 00:28:41,929 [Eric] Right. Yeah. We'll link to the show- 00:28:41,929 --> 00:28:41,929 [John] Right 00:28:41,929 --> 00:28:44,190 [Eric] ... where I analyse 20 years of email. [chuckles] 00:28:44,190 --> 00:28:53,250 [John] Yeah. That's a great point. Um, so I just got to thinking, like, oh, I've actually spent 15 years refining this skill through being in meetings and conversing- 00:28:53,250 --> 00:28:53,250 [Eric] Yes 00:28:53,250 --> 00:28:53,800 [John] ... with people- 00:28:53,800 --> 00:28:54,400 [Eric] Yep 00:28:54,400 --> 00:29:03,680 [John] ... of thinking in this, like, auditory way, and then, but then I have this other thing where I, I think I should sit down and write. 00:29:03,680 --> 00:29:04,180 [Eric] Mm-hmm. 00:29:04,180 --> 00:29:07,300 [John] Like, like, I don't know that that's gonna make sense for me. 00:29:07,300 --> 00:29:07,500 [Eric] Yeah. 00:29:07,500 --> 00:29:07,880 [John] Right? 00:29:07,880 --> 00:29:08,580 [Eric] Sure. 00:29:08,580 --> 00:29:18,120 [John] So because of that, I've just found that, like, it's, it's so good. Like, again, we've got to have Raycast on the show. Raycast has an awesome, like, app- 00:29:18,120 --> 00:29:18,210 [Eric] Yep 00:29:18,210 --> 00:29:34,250 [John] ... like, on your phone where you can, where you can do audio, um, audio to text, and it uses, like, a really good API, and it's very accurate. Um, but the, but you can go on a walk, or you can put it on a table or whatever you want to do, and then just get a ton of information out on a subject- 00:29:34,250 --> 00:29:34,250 [Eric] Yep 00:29:34,250 --> 00:29:48,140 [John] ... whether it's, like, uh, anything, a to-do list or, like, planning an architecture for something technical or whatever it is, and, um, it doesn't have to be perfectly organised. It doesn't have... But, but if you get the content- 00:29:48,140 --> 00:29:48,500 [Eric] Yes 00:29:48,500 --> 00:29:59,210 [John] ... AI is great at organising it post, like, conversation and then great at, like, trying to fact-check some things or, like, if you had open questions, like, go research questions. 00:29:59,210 --> 00:29:59,240 [Eric] Yep. 00:29:59,240 --> 00:30:08,760 [John] It's just, it's just such an interesting, like, path that was always available and probably always could have taken advantage of, but it is exponentially more useful. 00:30:08,760 --> 00:30:09,319 [Eric] Yes. 00:30:09,320 --> 00:30:26,945 [John] And, and, [chuckles] and I was... I think I shared this, like, this morning actually. Um, I was actually in a meeting, and this felt like the- every few weeks, I have these surreal mo- moments. I was in a meeting with somebody, one other person-... and on the table, we had two tables on the phone, like face up. One of them was an- 00:30:26,945 --> 00:30:27,756 [Eric] Two- yeah, two phones on the table. 00:30:27,756 --> 00:30:32,816 [John] Yeah, two phones on the table, yeah. One of them was a, was a phone with a AI note taker- 00:30:32,816 --> 00:30:33,376 [Eric] Mm-hmm 00:30:33,376 --> 00:30:48,076 [John] ... um, Granola, which is an awesome note taker. The other one was a phone with ChatGPT on voice mode, and we had a m- three, like, two people and the AI, like, in the meeting. [chuckles] And we had a whiteboard. 00:30:48,076 --> 00:30:48,556 [Eric] Two people- 00:30:48,556 --> 00:30:48,836 [John] So- 00:30:48,836 --> 00:30:49,856 [Eric] Two AIs in the meeting. 00:30:49,856 --> 00:30:50,506 [John] Yeah, technically. 00:30:50,506 --> 00:30:50,516 [Eric] [chuckles] 00:30:50,516 --> 00:30:52,256 [John] Yeah, one is listening, the other is talking. 00:30:52,256 --> 00:30:52,436 [Eric] Yeah. 00:30:52,436 --> 00:30:54,946 [John] But, I mean, you can imagine, you're in this room, and there's a whiteboard- 00:30:54,946 --> 00:30:54,946 [Eric] Yeah 00:30:54,946 --> 00:30:55,856 [John] ... with another person. 00:30:55,856 --> 00:30:56,116 [Eric] Yeah. 00:30:56,116 --> 00:31:01,066 [John] And we're, like, architecting something, and you get to a spot, and we're like: What tool should we use for that? 00:31:01,066 --> 00:31:01,085 [Eric] Yeah. 00:31:01,085 --> 00:31:03,216 [John] And we literally, like, just talk to the AI. 00:31:03,216 --> 00:31:03,286 [Eric] Yeah. 00:31:03,286 --> 00:31:06,386 [John] And be like, "Do a web search," like, "What are the tools?" and like, uh- 00:31:06,386 --> 00:31:06,386 [Eric] For sure 00:31:06,386 --> 00:31:18,396 [John] ... "for this thing," and you keep writing, and then you keep going, and the efficiency is, is wild. 'Cause that would've been a bunch of takeaways that somebody had to go do and research and come back, but we basically got almost all the options real time- 00:31:18,396 --> 00:31:18,636 [Eric] Amazing 00:31:18,636 --> 00:31:21,876 [John] ... and then could refine the options and be like, "Well, is this one better at this- 00:31:21,876 --> 00:31:21,886 [Eric] Right 00:31:21,886 --> 00:31:24,956 [John] ... or this?" And, and it's just doing web searches. It's not like- 00:31:24,956 --> 00:31:25,136 [Eric] Right 00:31:25,136 --> 00:31:27,975 [John] ... anything crazy. But it was a surreal experience. 00:31:27,975 --> 00:31:36,636 [Eric] That is wild. I, I totally agree with you on the voice-to-text and summarization. 00:31:36,636 --> 00:31:36,656 [John] Mm. 00:31:36,656 --> 00:31:42,316 [Eric] I've used it to distill a lot of ideas, 'cause external processing is very helpful. 00:31:42,316 --> 00:31:42,616 [John] Yeah. 00:31:42,616 --> 00:31:44,276 [Eric] Like speaking out an idea out loud- 00:31:44,276 --> 00:31:44,286 [John] Mm-hmm 00:31:44,286 --> 00:31:46,446 [Eric] ... 'cause you realize, like, oh, that actually is dumb. [chuckles] 00:31:46,446 --> 00:31:46,476 [John] [laughs] 00:31:46,476 --> 00:31:48,076 [Eric] It seemed good in my head. 00:31:48,076 --> 00:31:48,316 [John] Right. 00:31:48,316 --> 00:31:48,996 [Eric] Um- 00:31:48,996 --> 00:31:49,516 [John] Right. 00:31:49,516 --> 00:31:51,476 [Eric] You know, and writing is the same. You know- 00:31:51,476 --> 00:31:51,536 [John] Yeah 00:31:51,536 --> 00:31:53,286 [Eric] ... writing text or typing out text- 00:31:53,286 --> 00:31:53,286 [John] Yeah 00:31:53,286 --> 00:31:54,566 [Eric] ... can have the same effect. They're- 00:31:54,566 --> 00:31:54,566 [John] Yeah 00:31:54,566 --> 00:31:56,166 [Eric] ... but they're different. Um- 00:31:56,166 --> 00:32:10,036 [John] Do, when you, when you blog personally, is that audio part of your process, or do you just start with the writing and the markdown? Like, do you just sit down with the markdown file open and just type, or, or do you often have, like, some, like, audio notes that maybe you'd reference the text to? 00:32:10,036 --> 00:32:18,976 [Eric] I don't use, I don't use audio. Um, I don't use audio generally because I 00:32:18,976 --> 00:32:21,466 [Eric] take notes as I think about things. 00:32:21,466 --> 00:32:21,496 [John] Oh, okay. 00:32:21,496 --> 00:32:24,916 [Eric] I will jot notes down in a draft. So I just, I, I use a- 00:32:24,916 --> 00:32:25,006 [John] Oh 00:32:25,006 --> 00:32:25,796 [Eric] ... I use a tool called- 00:32:25,796 --> 00:32:26,616 [John] So you kind of already have a starting point 00:32:26,616 --> 00:32:27,716 [Eric] ... iA Writer. 00:32:27,716 --> 00:32:28,376 [John] Okay. 00:32:28,376 --> 00:32:30,196 [Eric] And it's a markdown editor. 00:32:30,196 --> 00:32:30,995 [John] Mm-hmm. 00:32:30,995 --> 00:32:32,636 [Eric] They have a mobile app, and so- 00:32:32,636 --> 00:32:32,646 [John] Hmm 00:32:32,646 --> 00:32:34,936 [Eric] ... there's a folder in there called Blog Drafts. 00:32:34,936 --> 00:32:35,436 [John] Okay. 00:32:35,436 --> 00:32:41,096 [Eric] And I have a bunch of ideas in there, and as I think through things, I will actually 00:32:41,096 --> 00:32:48,056 [Eric] just write down notes in there. And so what's interesting about that is, 00:32:48,056 --> 00:32:52,896 [Eric] the posts will come together over time- 00:32:52,896 --> 00:32:52,926 [John] Mm 00:32:52,926 --> 00:32:58,285 [Eric] ... and sort of as my thoughts develop on it, you know, a narrative arc will start to appear. 00:32:58,285 --> 00:32:58,296 [John] Yeah. 00:32:58,296 --> 00:33:02,386 [Eric] And then that helps me think about it even more critically, and then- 00:33:02,386 --> 00:33:02,386 [John] Right 00:33:02,386 --> 00:33:04,376 [Eric] ... actually writing the post from there- 00:33:04,376 --> 00:33:04,856 [John] It's just 00:33:04,856 --> 00:33:07,596 [Eric] ... is not as difficult and goes pretty quickly. 00:33:07,596 --> 00:33:07,656 [John] Yeah. 00:33:07,656 --> 00:33:09,816 [Eric] So, um... 00:33:09,816 --> 00:33:11,596 [John] Yeah. That, I th- that's cool. 00:33:11,596 --> 00:33:11,706 [Eric] Yeah. 00:33:11,706 --> 00:33:18,636 [John] That, I've heard that, that that is a common process for, like, people that, that like, do blogging, like, and in a, in a good process. 00:33:18,636 --> 00:33:19,026 [Eric] Yeah. 00:33:19,026 --> 00:33:19,036 [John] Uh... 00:33:19,036 --> 00:33:33,296 [Eric] It's been helpful for me. But I, I have used the audio before, um, but I, I haven't used it recently, but it's usually because of the note... You know, it's usually because of the notes thing. Um, but 00:33:33,296 --> 00:33:37,336 [Eric] I think there's three things emerging here, which 00:33:37,336 --> 00:33:39,046 [Eric] let's, let's close out- 00:33:39,046 --> 00:33:39,216 [John] Yeah 00:33:39,216 --> 00:33:45,476 [Eric] ... by breaking down these, these three things. Um, 00:33:45,476 --> 00:33:49,636 [Eric] there... Or, really, we'll talk through the three things, and then one of them is a question. 00:33:49,636 --> 00:33:49,936 [John] Okay. 00:33:49,936 --> 00:33:54,796 [Eric] So, there are skills that atrophy. 00:33:54,796 --> 00:33:55,656 [John] Yep. 00:33:55,656 --> 00:34:01,505 [Eric] Right? There are, um, but... Or skills are, are atrophying. 00:34:01,505 --> 00:34:01,535 [John] Yes. 00:34:01,536 --> 00:34:02,755 [Eric] Is that the right way to say that? 00:34:02,756 --> 00:34:03,716 [John] I don't know. 00:34:03,716 --> 00:34:05,696 [Eric] Skills are undergoing atrophy. [laughs] 00:34:05,696 --> 00:34:07,696 [John] Sure. [laughs] 00:34:07,696 --> 00:34:09,916 [Eric] That's so cl- that's so clinical. 00:34:09,916 --> 00:34:09,956 [John] Yeah. 00:34:09,956 --> 00:34:11,256 [Eric] Skills are undergoing atrophy. 00:34:11,256 --> 00:34:12,036 [John] Right. 00:34:12,036 --> 00:34:15,626 [Eric] Uh, we won't name the episode. That will not be the episode title. [chuckles] 00:34:15,626 --> 00:34:16,315 [John] [chuckles] Yeah. 00:34:16,316 --> 00:34:18,496 [Eric] Uh, okay, so that's happening. 00:34:18,496 --> 00:34:19,736 [John] Mm-hmm. 00:34:19,736 --> 00:34:22,416 [Eric] New skills are being formed. 00:34:22,416 --> 00:34:22,996 [John] Yes, definitely. 00:34:22,996 --> 00:34:26,076 [Eric] So, for example, 00:34:26,136 --> 00:34:32,285 [Eric] your process of speaking things out or having AI as a collaborator in a meeting- 00:34:32,285 --> 00:34:32,296 [John] Yeah 00:34:32,296 --> 00:34:34,796 [Eric] ... solving a problem, those are new skills- 00:34:34,796 --> 00:34:34,876 [John] Yeah 00:34:34,876 --> 00:34:36,376 [Eric] ... new ways of doing things. 00:34:36,376 --> 00:34:36,896 [John] Yeah. 00:34:36,896 --> 00:34:36,916 [Eric] Um- 00:34:36,916 --> 00:34:59,656 [John] Like, a, just super quick example of that, helping somebody with something literally, like, an hour ago, and, um, they were typing, and I'm, like, kind of, like, walking with them, I was like, "I would love..." Like, I was like, "Do you have such and such installed on your computer?" He's like, "No." I was like: " Wouldn't it be so great if I could sit here and show you this, and I was just speaking, and the words, like, typed on your computer, and it, like, did the thing?" 00:34:59,656 --> 00:34:59,876 [Eric] Yeah. 00:34:59,876 --> 00:35:01,736 [John] Like, that can happen. Like, we can do that now. 00:35:01,736 --> 00:35:02,456 [Eric] We can do it now. 00:35:02,456 --> 00:35:02,716 [John] Yeah. 00:35:02,716 --> 00:35:03,115 [Eric] Yes. 00:35:03,116 --> 00:35:03,175 [John] But- 00:35:03,176 --> 00:35:03,716 [Eric] It's amazing, yeah. 00:35:03,716 --> 00:35:04,436 [John] Yeah. Anyways. 00:35:04,436 --> 00:35:07,656 [Eric] So there's these, you know... And even in write, like, at writing at Vercel- 00:35:07,656 --> 00:35:07,906 [John] Yes 00:35:07,906 --> 00:35:10,506 [Eric] ... they're completely new skill sets that I've learned using AI- 00:35:10,506 --> 00:35:10,766 [John] Yes, exactly 00:35:10,766 --> 00:35:11,495 [Eric] ... as part of the writing process. 00:35:11,496 --> 00:35:12,076 [John] Yeah. 00:35:12,076 --> 00:35:16,115 [Eric] Okay, so skills are atrophying. New skills are being formed- 00:35:16,116 --> 00:35:16,146 [John] New skills are emerging 00:35:16,146 --> 00:35:17,056 [Eric] ... new ways of doing, right? 00:35:17,056 --> 00:35:17,096 [John] Yeah. 00:35:17,096 --> 00:35:18,506 [Eric] So, like, retrieving information- 00:35:18,506 --> 00:35:18,506 [John] Yep 00:35:18,506 --> 00:35:19,206 [Eric] ... a la Google, right? 00:35:19,206 --> 00:35:19,216 [John] Yeah. 00:35:19,216 --> 00:35:25,216 [Eric] That's changing. So the question, the core question is: 00:35:25,216 --> 00:35:28,216 [Eric] How do you decide 00:35:28,216 --> 00:35:34,626 [Eric] which skills to, uh, work on so that they do not atrophy? 00:35:34,626 --> 00:35:34,696 [John] Right. 00:35:34,696 --> 00:35:36,446 [Eric] Which muscles do you need to- 00:35:36,446 --> 00:35:36,816 [John] Yeah, right 00:35:36,816 --> 00:35:38,546 [Eric] ... maintain? Like, how are you- 00:35:38,546 --> 00:35:38,676 [John] Right 00:35:38,676 --> 00:35:40,296 [Eric] ... how are you thinking about that in your world? 00:35:40,296 --> 00:35:44,495 [John] Yeah. Yeah, I mean, 00:35:44,556 --> 00:35:51,066 [John] to me, it's actually maybe, and we kinda covered this at the beginning, of, like, which ones, the, with the two buckets. 00:35:51,066 --> 00:35:51,096 [Eric] Mm-hmm. 00:35:51,096 --> 00:35:53,925 [John] So the, I mean, the short answer is, like, well, the fundamental ones, and then- 00:35:53,925 --> 00:35:53,925 [Eric] Yep 00:35:53,925 --> 00:35:55,956 [John] ... the real question is, like, well, what's fundamental? 00:35:55,956 --> 00:35:56,056 [Eric] Yes. 00:35:56,056 --> 00:36:18,996 [John] And that might... I think that depends a little bit person to person, as far as like, what do you want? Like, 'cause if you're in two levels, like, one, let's just, like, say this was... Let's use the, like, strength training analogy. Like, cool, what do you want? Like, you can strength train, and the goal is, like, longevity and overall health, and like, you can strength train to be a bodybuilder, right? 00:36:18,996 --> 00:36:19,196 [Eric] Mm-hmm. 00:36:19,196 --> 00:36:21,576 [John] So there's that, like, intensity and like- 00:36:21,576 --> 00:36:21,606 [Eric] Oh, yeah 00:36:21,606 --> 00:36:22,336 [John] ... what's your goal- 00:36:22,336 --> 00:36:22,536 [Eric] Yep 00:36:22,536 --> 00:36:31,626 [John] ... piece.... and you can also, like, kind of train for, like, flexibility versus, like, overall, like, how much, like, weight can I move? 00:36:31,626 --> 00:36:31,676 [Eric] Yep. 00:36:31,676 --> 00:36:38,976 [John] Which is kind of another thing. Um, so I think that, like, applied to, to skill atrophy 00:36:38,976 --> 00:36:45,756 [John] is, like, the, uh, s- there's big buckets, like communication, like we talked about. 00:36:45,756 --> 00:36:46,335 [Eric] Mm-hmm. 00:36:46,336 --> 00:36:53,835 [John] And you may really try... And I think you need to pick one inside of communication to keep sharp- 00:36:53,836 --> 00:36:53,856 [Eric] Mm-hmm 00:36:53,856 --> 00:37:12,776 [John] ... but you don't necessarily need to, like, keep all of them sharp. Like, for example, 'cause I think there's gonna be so much more fluidity, and the perfect example is, like, writing versus, like, communication and speaking. If you have one of those sharp, A, like, i- it's ordered, it's, like, properly ordered thinking is, like, kind of the, like- 00:37:12,776 --> 00:37:12,836 [Eric] Yes 00:37:12,836 --> 00:37:14,025 [John] ... the tied together, so that's the most- 00:37:14,025 --> 00:37:14,025 [Eric] Yep 00:37:14,025 --> 00:37:14,576 [John] ... important one. 00:37:14,576 --> 00:37:15,236 [Eric] Yep. 00:37:15,236 --> 00:37:21,345 [John] But then, like, the outcome of, like, whether you do it through writing or through speaking matters a little bit less because- 00:37:21,345 --> 00:37:21,345 [Eric] Yep 00:37:21,345 --> 00:37:23,876 [John] ... transitioning between the two is gonna become really easy. 00:37:23,876 --> 00:37:24,636 [Eric] Agreed. 00:37:24,636 --> 00:37:39,416 [John] So that to me, for me, the way I think about it, is I've got to keep the, like, critical thinking, ordered thinking, like, piece sharp. How it, like, comes out actually matters less because the transition between the, like, either- 00:37:39,416 --> 00:37:39,426 [Eric] Mm 00:37:39,426 --> 00:37:41,276 [John] ... writing or speaking is, like- 00:37:41,276 --> 00:37:41,286 [Eric] What a great point 00:37:41,286 --> 00:37:42,456 [John] ... almost seamless. 00:37:42,456 --> 00:37:43,195 [Eric] Yeah. What a great point. 00:37:43,196 --> 00:37:48,576 [John] Um, and that can come to some practicality of, like, who you are as a person, what are your natural skills- 00:37:48,576 --> 00:37:48,586 [Eric] Yep 00:37:48,586 --> 00:37:55,116 [John] ... et cetera. Um, but that's how I think about it on that, and then since we haven't touched on it much, the creativity piece- 00:37:55,116 --> 00:37:55,896 [Eric] Mm 00:37:55,896 --> 00:38:06,316 [John] ... is, um, is really interesting, and I think that's the one that I probably have the most question marks about. Because there's an aspect of AI 00:38:06,316 --> 00:38:11,136 [John] where... Of, you know, where, like, I didn't think about it that way. Like, happens all the time, right? 00:38:11,136 --> 00:38:11,156 [Eric] Yep. 00:38:11,156 --> 00:38:24,766 [John] But there's another aspect where I've, I said this to a friend the other day, is, like, imagine, and this is in the software world, imagine every single app that existed was written by the same developer with the same brain. [chuckles] 00:38:24,766 --> 00:38:24,796 [Eric] Right. 00:38:24,796 --> 00:38:26,056 [John] I mean, that's weird. 00:38:26,056 --> 00:38:26,176 [Eric] Yeah. 00:38:26,176 --> 00:38:31,236 [John] And essentially, like, anybody can talk to that developer and ask it to do things. 00:38:31,236 --> 00:38:31,856 [Eric] Right. 00:38:31,856 --> 00:38:38,516 [John] And then real world is like, well, imagine there's about five of them, and they, like, basically are the same, but they're a little bit different. 00:38:38,516 --> 00:38:39,176 [Eric] Mm-hmm. Mm-hmm. 00:38:39,176 --> 00:38:41,905 [John] So and the, and the bigger question is, like, how does that fundamentally- 00:38:41,905 --> 00:38:41,956 [Eric] Yeah 00:38:41,956 --> 00:38:55,536 [John] ... change things? But specific to this, in the skill atrophy, um, it just really got me thinking of, far as, like, any amount of creativity that is, like, different than in, like, again, in coding land- 00:38:55,536 --> 00:38:55,796 [Eric] Mm-hmm 00:38:55,796 --> 00:38:59,836 [John] ... than the one software brain that's doing [chuckles] all the things- 00:38:59,836 --> 00:39:00,136 [Eric] Yeah 00:39:00,136 --> 00:39:01,456 [John] ... becomes extremely valuable. 00:39:01,456 --> 00:39:02,466 [Eric] That's super interesting. 00:39:02,466 --> 00:39:08,056 [John] And then the question is, like, how do you get there? Because if it's essentially, like, trained on, like, all knowledge- 00:39:08,056 --> 00:39:08,216 [Eric] Mm 00:39:08,216 --> 00:39:12,816 [John] ... like, or is trending toward training on, like, all, like, digital knowledge- 00:39:12,816 --> 00:39:12,866 [Eric] Yep 00:39:12,866 --> 00:39:19,156 [John] ... let's call it that. Like, what would be the things that would make it so you have some kind of, like, creativity piece- 00:39:19,156 --> 00:39:19,166 [Eric] Mm 00:39:19,166 --> 00:39:21,526 [John] ... that maybe it doesn't have? And I don't know the answer to that question. 00:39:21,526 --> 00:39:21,586 [Eric] That's really interesting. 00:39:21,586 --> 00:39:22,975 [John] I think that's a very difficult- 00:39:22,976 --> 00:39:23,516 [Eric] Yeah 00:39:23,516 --> 00:39:24,316 [John] ... question. 00:39:24,316 --> 00:39:30,736 [Eric] That's an interesting example. I, I ran into that recently. I was writing a video script, and actually, we were doing 00:39:30,796 --> 00:39:33,976 [Eric] production on the video, so actually recording it. It was a demo. 00:39:33,976 --> 00:39:34,236 [John] Okay, yeah. 00:39:34,236 --> 00:39:35,096 [Eric] Sort of a long-form- 00:39:35,096 --> 00:39:35,185 [John] Right 00:39:35,185 --> 00:39:43,336 [Eric] ... narrative demo, and which was very fun, by the way, um, to sort of write a script end to end and then, like- 00:39:43,336 --> 00:39:43,526 [John] Yeah 00:39:43,526 --> 00:39:44,196 [Eric] ... do production on it. 00:39:44,196 --> 00:39:44,435 [John] Yeah. 00:39:44,436 --> 00:39:45,196 [Eric] Um- 00:39:45,196 --> 00:39:51,796 [John] And, like, such a quick turnaround too, right? Like, I know if you're in, you know, a more traditional media role, I mean, that's, like, a long process. 00:39:51,796 --> 00:39:53,086 [Eric] Yeah, yeah, yeah. It was quick. 00:39:53,086 --> 00:39:53,096 [John] Yeah. 00:39:53,096 --> 00:39:57,076 [Eric] But the... Yeah, the production team at Vercel is, uh, outstanding. 00:39:57,076 --> 00:39:57,266 [John] Yeah, for sure. 00:39:57,266 --> 00:40:04,405 [Eric] Uh, and the guy who was actually doing the demo, who I wrote the script for, was outstanding, and so it was, it was very, very- 00:40:04,405 --> 00:40:04,405 [John] Yeah 00:40:04,405 --> 00:40:06,056 [Eric] ... it was, like, shockingly quick. Um- 00:40:06,056 --> 00:40:06,616 [John] Super cool. 00:40:06,616 --> 00:40:06,896 [Eric] But- 00:40:06,896 --> 00:40:07,776 [John] Yep 00:40:07,776 --> 00:40:20,176 [Eric] ... it was very interesting because I, I heavily used AI to help me think through... What's very difficult with a demo with screen sharing, where the person is also being videoed, is how do you do cuts 00:40:20,176 --> 00:40:24,076 [Eric] at various points? Especially if you're using AI, and it's thinking- 00:40:24,076 --> 00:40:24,155 [John] Right. 00:40:24,156 --> 00:40:25,836 [Eric] You want it to be authentic- 00:40:25,836 --> 00:40:25,845 [John] Sure 00:40:25,845 --> 00:40:26,936 [Eric] ... but you don't wanna, like- 00:40:26,936 --> 00:40:27,676 [John] Bore people. Yeah 00:40:27,676 --> 00:40:28,685 [Eric] ... bore people. 00:40:28,685 --> 00:40:28,716 [John] Right. 00:40:28,716 --> 00:40:31,856 [Eric] There are just a number of... It, it's 00:40:31,856 --> 00:40:36,296 [Eric] very tricky, actually, to do that in a way that feels very natural, um- 00:40:36,296 --> 00:40:37,296 [John] Right 00:40:37,296 --> 00:40:40,855 [Eric] ... and to write in a way that someone can speak as opposed to- 00:40:40,856 --> 00:40:41,076 [John] Yeah 00:40:41,076 --> 00:40:49,896 [Eric] ... like, read. And, and then you have these aspects of, like, you know, I had to determine an example application 00:40:49,896 --> 00:40:53,485 [Eric] that made narrative sense, like- 00:40:53,485 --> 00:40:53,485 [John] Yeah 00:40:53,485 --> 00:40:59,416 [Eric] ... could be built in the sequence that the demo needed to happen in to show these different products in a very particular sequence- 00:40:59,416 --> 00:40:59,996 [John] Right 00:40:59,996 --> 00:41:03,896 [Eric] ... and seem real and legitimate, like, be an actual use case. 00:41:03,896 --> 00:41:04,496 [John] Yeah. Yeah. 00:41:04,496 --> 00:41:16,045 [Eric] Um, and it was pr- uh, AI was very helpful, but it also, we, we live changed some things because they're like, "You know, there's a ve- better way to do this. It would be really cool if we did this or this- 00:41:16,045 --> 00:41:16,045 [John] Right 00:41:16,045 --> 00:41:18,846 [Eric] ... you know?" Anyways, that creativity piece, it was, it was- 00:41:18,846 --> 00:41:18,846 [John] Yeah 00:41:18,846 --> 00:41:30,856 [Eric] ... I had that very experience where I was like, okay, it, the, it was so high context, like, in the moment, that it was cool to see human creativity, like, flourishing in that moment of multiple people- 00:41:30,856 --> 00:41:31,076 [John] Right 00:41:31,076 --> 00:41:34,956 [Eric] ... exchanging ideas and coming up with, you know, a better- 00:41:34,956 --> 00:41:34,966 [John] Right 00:41:34,966 --> 00:41:39,216 [Eric] ... implementation of, you know, a decent idea, but a much better way to do it. 00:41:39,216 --> 00:41:51,516 [John] Yeah, that's super interesting. And, and then there's the other, like, you know, if you're in software today, and you're innovating and doing a lot of new stuff, like, the models, the training still lags. 00:41:51,516 --> 00:41:51,616 [Eric] Mm-hmm. 00:41:51,616 --> 00:41:52,086 [John] So it's not like- 00:41:52,086 --> 00:41:52,096 [Eric] Mm-hmm 00:41:52,096 --> 00:41:54,776 [John] ... it's been trained on whatever, like, brand-new thing you're doing. 00:41:54,776 --> 00:41:54,916 [Eric] Right. 00:41:54,916 --> 00:41:58,576 [John] Like, it has a lot of, like, information that would be relevant, but there's that, too. 00:41:58,576 --> 00:41:59,936 [Eric] Yep. Yeah. 00:41:59,936 --> 00:42:00,436 [John] Yeah. 00:42:00,436 --> 00:42:02,196 [Eric] I think 00:42:02,196 --> 00:42:07,476 [Eric] my, my closing thoughts on, on how do you decide what to let atrophy- 00:42:07,476 --> 00:42:08,276 [John] Yep 00:42:08,336 --> 00:42:14,045 [Eric] ... I have two major, two sort of big points on that. Major points? I don't know. 00:42:14,045 --> 00:42:14,116 [John] Yeah. 00:42:14,116 --> 00:42:15,076 [Eric] I have two thoughts on that. 00:42:15,076 --> 00:42:15,946 [John] Yeah. [chuckles] 00:42:15,946 --> 00:42:16,476 [Eric] They're major. 00:42:16,476 --> 00:42:17,856 [John] Theseses? [chuckles] 00:42:17,856 --> 00:42:18,516 [Eric] The- theses. 00:42:18,516 --> 00:42:19,516 [John] Theses. [chuckles] 00:42:19,516 --> 00:42:20,356 [Eric] Uh- 00:42:20,356 --> 00:42:20,896 [John] Wow. 00:42:20,896 --> 00:42:21,236 [Eric] Yes. 00:42:21,236 --> 00:42:22,156 [John] It's been a long week. 00:42:22,156 --> 00:42:25,456 [Eric] Uh, it has been a long week. Um, 00:42:25,456 --> 00:42:41,372 [Eric] the first one is that I, I think-... it is, it, I think reframing the way that I think about AI is a big part of 00:42:41,372 --> 00:42:43,692 [Eric] approaching the atrophy question- 00:42:43,692 --> 00:42:44,072 [John] Right 00:42:44,072 --> 00:42:47,132 [Eric] ... well, because 00:42:47,132 --> 00:42:54,912 [Eric] w- this is interesting. If you think about almost all of our conversation, it's in the context of using AI to automate something- 00:42:54,912 --> 00:42:55,012 [John] Yeah 00:42:55,012 --> 00:42:56,172 [Eric] ... that we used to do. 00:42:56,172 --> 00:42:57,732 [John] Yeah, definitely. 00:42:57,732 --> 00:43:03,172 [Eric] But that is a fundamentally limiting framing of the tool. 00:43:03,172 --> 00:43:03,592 [John] Yeah. 00:43:03,592 --> 00:43:12,932 [Eric] Right? What's very interesting, if we think about writing, is that... Because the default is, okay, it can crank out a blog post for you really quickly, right? 00:43:12,932 --> 00:43:13,132 [John] Right. 00:43:13,132 --> 00:43:14,732 [Eric] You know, or it can get you an outline really quickly- 00:43:14,732 --> 00:43:14,742 [John] Right 00:43:14,742 --> 00:43:20,152 [Eric] ... or whatever, right? I- that's the wrong way to think about it. Uh, 00:43:20,152 --> 00:43:23,542 [Eric] you can use the tool to become a better writer. 00:43:23,542 --> 00:43:24,552 [John] Right. Yeah. 00:43:24,552 --> 00:43:25,902 [Eric] It can teach you. 00:43:25,902 --> 00:43:26,032 [John] Yeah. 00:43:26,032 --> 00:43:28,252 [Eric] Um, right? And I'm, I'm not saying- 00:43:28,252 --> 00:43:28,852 [John] Yeah 00:43:28,852 --> 00:43:31,712 [Eric] ... it's a, it's a full replacement for, like, experience, practice- 00:43:31,712 --> 00:43:31,722 [John] Right 00:43:31,722 --> 00:43:33,112 [Eric] ... you know, you know, whatever. 00:43:33,112 --> 00:43:37,952 [John] Deeper, well-researched, better documented, literally have it critique- 00:43:37,952 --> 00:43:38,532 [Eric] Exactly. 00:43:38,532 --> 00:43:38,972 [John] Well, yeah- 00:43:38,972 --> 00:43:39,242 [Eric] Exactly 00:43:39,242 --> 00:43:40,202 [John] ... all of those things. 00:43:40,202 --> 00:43:40,202 [Eric] You know? 00:43:40,202 --> 00:43:40,232 [John] Yeah. 00:43:40,232 --> 00:43:46,692 [Eric] So, uh, so that I think is, it- it's almost an inversion, saying, like, okay, skills are gonna atrophy- 00:43:46,692 --> 00:43:46,702 [John] Right 00:43:46,702 --> 00:43:54,292 [Eric] ... but ironically, the things that y- you can fight atrophy using the thing that, using the threat- [chuckles] 00:43:54,292 --> 00:43:54,492 [John] Yeah 00:43:54,492 --> 00:43:55,512 [Eric] ... to atrophy, right? 00:43:55,512 --> 00:43:56,272 [John] Yeah. 00:43:56,272 --> 00:43:57,211 [Eric] Which is really interesting. 00:43:57,212 --> 00:43:57,912 [John] That's- 00:43:57,912 --> 00:43:59,492 [Eric] Um. 00:43:59,492 --> 00:44:02,252 [John] Which means there's gonna be a huge divergent in, in people- 00:44:02,252 --> 00:44:03,442 [Eric] Yes. I agree 00:44:03,442 --> 00:44:05,612 [John] ... and how they use AI and how they think about AI. 00:44:05,612 --> 00:44:17,532 [Eric] Yep. So I think that's a really, I think that's a really big one. A- another example would be, you know, I use... A- AI is a great way to learn things, right? 00:44:17,532 --> 00:44:17,572 [John] Yeah. 00:44:17,572 --> 00:44:18,142 [Eric] In voice mode. 00:44:18,142 --> 00:44:18,172 [John] Right. 00:44:18,172 --> 00:44:22,432 [Eric] One thing that I love to do if I'm driving around by myself, 00:44:22,432 --> 00:44:24,322 [Eric] is to just pick a topic that I'm really interested in. 00:44:24,322 --> 00:44:24,372 [John] Yeah. 00:44:24,372 --> 00:44:25,612 [Eric] I did this with economics. 00:44:25,612 --> 00:44:25,672 [John] Okay. 00:44:25,672 --> 00:44:26,652 [Eric] So I turned Jabby- 00:44:26,652 --> 00:44:26,662 [John] Right 00:44:26,662 --> 00:44:27,772 [Eric] ... ChatGPT voice mode on. 00:44:27,772 --> 00:44:28,472 [John] Yeah. 00:44:28,472 --> 00:44:31,292 [Eric] I asked for historical examples of commoditization- 00:44:31,292 --> 00:44:31,342 [John] Oh, cool 00:44:31,342 --> 00:44:32,332 [Eric] ... you know, throughout history. 00:44:32,332 --> 00:44:32,862 [John] Right. 00:44:32,862 --> 00:44:35,452 [Eric] You know, what happened during the Industrial Revolution? 00:44:35,452 --> 00:44:35,952 [John] Yep. 00:44:35,952 --> 00:44:40,172 [Eric] You know, what were the different impacts of that? Just a l- asking it questions- 00:44:40,172 --> 00:44:40,182 [John] Yeah 00:44:40,182 --> 00:44:43,952 [Eric] ... right? And you have this, an unbelievable repository of information. 00:44:43,952 --> 00:44:44,252 [John] Right. 00:44:44,252 --> 00:44:47,052 [Eric] You know, and so you can sort of do advanced study on economics- 00:44:47,052 --> 00:44:47,062 [John] Yeah 00:44:47,062 --> 00:44:48,062 [Eric] ... while you're driving to the grocery store. 00:44:48,062 --> 00:44:51,312 [John] I did this with, um... I've been kind of obsessed with the Gilded Age- 00:44:51,372 --> 00:44:51,772 [Eric] Mm 00:44:51,772 --> 00:44:52,432 [John] ... recently. 00:44:52,432 --> 00:44:53,812 [Eric] That's Mark Twain. He coined that. 00:44:53,812 --> 00:44:59,452 [John] Yeah. Yep. Oh, oh, that's right, yeah! So which is, like, um, Rock- like, Johnny Rockefeller- 00:44:59,452 --> 00:44:59,772 [Eric] Mm-hmm 00:44:59,772 --> 00:45:00,852 [John] ... and, and, um, JP Morgan- 00:45:00,852 --> 00:45:00,861 [Eric] Mm-hmm 00:45:00,861 --> 00:45:02,521 [John] ... like, Vanderbilt, those guys. 00:45:02,521 --> 00:45:03,132 [Eric] Mm-hmm. 00:45:03,132 --> 00:45:10,912 [John] Um, 'cause a lot of people have said, like, the closest, like, thing in history of an innovation-driven, like, boom... 'Cause World War II doesn't, 'cause that's war-driven- 00:45:10,912 --> 00:45:11,192 [Eric] Mm-hmm 00:45:11,192 --> 00:45:12,392 [John] ... has been that. So we- 00:45:12,392 --> 00:45:12,402 [Eric] Mm 00:45:12,402 --> 00:45:13,092 [John] ... should talk about that. 00:45:13,092 --> 00:45:13,372 [Eric] Okay. 00:45:13,372 --> 00:45:13,712 [John] On a separate episode. 00:45:13,712 --> 00:45:14,712 [Eric] There's an episode on that. 00:45:14,712 --> 00:45:15,391 [John] Yeah. 00:45:15,392 --> 00:45:20,211 [Eric] Okay, the second thought on atrophy 00:45:20,212 --> 00:45:22,292 [Eric] is that 00:45:22,292 --> 00:45:32,052 [Eric] i- it's, again, a kind of a, a reframing of the way you think about AI. Instead of thinking, "Okay, what can I outsource to AI?- 00:45:32,052 --> 00:45:32,252 [John] Right 00:45:32,252 --> 00:45:43,532 [Eric] ... And what should I not outsource to AI?" I think a better way to frame it is: What skills will make me more effective in using AI to do the things- 00:45:43,532 --> 00:45:43,652 [John] Sure 00:45:43,652 --> 00:45:44,632 [Eric] ... that I want to do? 00:45:44,632 --> 00:45:44,972 [John] Yeah. 00:45:44,972 --> 00:45:53,812 [Eric] Right? So for example, working with data. You don't need to 00:45:53,872 --> 00:45:57,972 [Eric] continue to progress in your ability to handwrite, you know- 00:45:57,972 --> 00:45:58,772 [John] SQL. Yeah. 00:45:58,772 --> 00:45:59,732 [Eric] You can handwrite SQL- 00:45:59,732 --> 00:45:59,801 [John] Right 00:45:59,801 --> 00:46:01,762 [Eric] ... right? Window functions. [chuckles] 00:46:01,762 --> 00:46:01,792 [John] [chuckles] 00:46:01,792 --> 00:46:03,502 [Eric] Okay, just let AI do that for you. 00:46:03,502 --> 00:46:03,512 [John] Yeah. Sure. 00:46:03,512 --> 00:46:04,032 [Eric] Right? 00:46:04,032 --> 00:46:05,412 [John] Sure. 00:46:05,412 --> 00:46:14,152 [Eric] The, the right question, I believe, is: What, how, what are the skills and muscles that you need to either maintain or build- 00:46:14,152 --> 00:46:14,162 [John] Right 00:46:14,162 --> 00:46:16,472 [Eric] ... in order to use AI for data? 00:46:16,472 --> 00:46:16,912 [John] Yeah. 00:46:16,912 --> 00:46:24,081 [Eric] Right? And some of those are gonna be existing skills that are at threat of atrophy because of the path of least resistance, that will eventually diminish your ability- 00:46:24,081 --> 00:46:24,092 [John] Right 00:46:24,092 --> 00:46:25,312 [Eric] ... to use AI well for data. 00:46:25,312 --> 00:46:25,692 [John] Right. 00:46:25,692 --> 00:46:27,452 [Eric] And some of those, I think, are gonna be new skills. 00:46:27,452 --> 00:46:33,632 [John] Yeah, and I think one of the, the new skills, again, we could do a whole show on this, is, is the meta work- 00:46:33,632 --> 00:46:33,772 [Eric] Yes 00:46:33,772 --> 00:46:36,352 [John] ... of, like, w- the tools that build the tools. 00:46:36,352 --> 00:46:36,592 [Eric] Yes. 00:46:36,592 --> 00:46:38,912 [John] Like, that's where I think a lot of work is going. 00:46:38,912 --> 00:46:39,212 [Eric] Mm-hmm. 00:46:39,212 --> 00:46:41,632 [John] It's, it's you're working on the tools that build the tools. 00:46:41,632 --> 00:46:46,681 [Eric] Yes, I agree. All right, I'm gonna go write a blog post without AI. [chuckles] 00:46:46,681 --> 00:46:48,092 [John] [chuckles] All right. 00:46:48,092 --> 00:46:48,912 [Eric] Thanks for joining us. 00:46:48,912 --> 00:46:48,922 [John] Yeah. 00:46:48,922 --> 00:46:52,032 [Eric] I'll catch you on the next one.