The Invisible Work AI Can't Replace—with Paul Traficanti (Nextworld)

If AI can produce the deck, write the PRD, and draft the battle card—what's left for you to own?

In this episode of Futureproof, Xano CEO Prakash Chandran sits down with Paul Traficanti, Director of Product Marketing at Nextworld and writer behind the Substack newsletter Somebody Manage Me. Paul is a founder turned operator who has worked across gaming, direct-to-consumer, and enterprise SaaS—and he's been writing about something most AI conversations skip over: as AI takes over the visible work, the invisible human layer underneath becomes even more invisible. Together, he and Prakash explore what taste, judgment, and critical thinking actually look like in daily work, why AI-generated documents are creating a new kind of organizational noise, and how leaders can start making the invisible contributions of their teams visible again. They also dig into how hiring is changing when the floor has been raised, why meetings are shifting from document reviews to alignment conversations, and what it means that an LLM-written document is now being summarized by another LLM so no one has to read it.

Topics covered include:

• Taste as the new competitive advantage: Why the human decisions behind a prompt—what to say, to whom, and why—matter more than the polished output AI produces.
• The document graveyard problem: How AI-generated PRDs, battle cards, and decks are creating more content but less understanding, and why less is about to become more again.
• "Show me the prompt": A practical framework for cutting through AI slop by asking teammates what context they gave the model and why—and the water cooler test for whether they truly understand their own output.
• Hiring when the floor is raised: Why interview homework assignments no longer reveal much, and how companies are rethinking evaluation to test for critical thinking, not production skill.
• AI literacy gaps across functions: Why developers adopted AI fastest but sales and marketing teams are still catching up, and the emerging need for internal AI enablement roles.

Subscribe to Futureproof wherever you get your podcasts.

From Xano - The fastest way to create a production-ready backend for any app or agent. Xano unifies AI speed, code control, and visual clarity, so you never trade reliability for velocity. Sign up for free today.

Transcript

SPEAKER_01 0:00

There's taste, uh, judgment, critical thinking. Those are the types of things that AI can't really replicate. You give it the decisions that you made, the critical thinking, the taste, and those types of things, and it can work off of that. It can give you 10 different versions of how to say something, but what you tell it to say is what comes from you, right? And that's the part that you have to own.

SPEAKER_00 0:33

I'm Prakash Chandran, CEO of Xano. Today I'm joined by Paul Trafficanti, director of product marketing at Nextworld, and the writer behind the Substack newsletter, Somebody Manage Me. Paul is a founder turned operator who spent his career at the intersection of product and go-to-market, from co-founding Deirdrop, backed by A16Z, to leading growth companies across gaming, direct-to-consumer, and enterprise SaaS. He also builds with AI tools like Lovable and Claude, so he's not just theorizing about this from the sidelines. Paul has been writing about something that I don't think enough people are talking about. As AI takes over more of the visible work, the production, the formatting, the first drafts, the invisible work underneath becomes even more invisible. The taste, the judgment, the ability to look at a polished AI output and know it's saying the wrong thing to the wrong audience. That is what this conversation is about. The human layer that AI can't replace and what happens to teams and organizations that forget it's there. Paul, I'm really excited to talk about this today. Welcome.

SPEAKER_01 1:37

Thanks so much for having me, Prakash. I'm excited.

SPEAKER_00 1:40

Of course. So I thought we'd get started with just uh a little bit of background. Tell us a little bit about yourself and what you do today at Nextworld.

SPEAKER_01 1:48

Yeah, for sure. So, so as you mentioned, uh I've I've kind of blended my career with a little bit of founder, um, and and most recently I'm an operator at a company called Nextworld. Um, but before I joined Nextworld, um, I largely was working in the gaming space. Um, and that really kind of opened my perspective towards the whole AI debate. Um, you know, this is never evolving thing, but um, developers are probably in a generalization slightly, but developers are probably the most like or one of the most anti-AI industries that I have come across. Um, and I think that's because of a mix, it's a very much a mix of art and science, um, or it's science in this instance being code. But um, so spending years around that has made me think a lot about what really is human, like what is creative, what are those types of things. Um and um, as you mentioned most recently, um, I am the director of product marketing at Nextworld. Um, we are an enterprise application platform, like no code enterprise application platform, and we also have a financial ERP.

SPEAKER_00 2:52

One of the reasons I came across you and wanted to reach out was your Substack. And you started writing uh about a lot of these thoughts. Um before we dive in, what was the impetus to saying, you know what, I need to create a space to start talking about this?

SPEAKER_01 3:06

There were so many things, so many newsletters, so many things across my feed about how to tactically do this thing, how to make this cloud scale for X, Y, and Z, and how to do all these things. And, you know, I consume all that and I use it a great deal, but it still felt to me like I was showing up to work on a day-to-day basis. And the hard part of my day-to-day wasn't how am I going to build this thing with AI. Um, that was easy to just get your hands in. It was more like, what are the parts of work that um maybe haven't changed or maybe more emphasized that people aren't necessarily talking about as a result of AI? And I just felt like there's not enough of a narrative talking about that side of work right now.

SPEAKER_00 3:50

Was there like a key example or something that happened that made you that kind of like highlights this? I think when we were talking before, I think you had also written this. You know, people make slide decks in Claude and they present it, but they don't ever stop to think about wait, should I have made that slide deck in the first place? So that's kind of one example. Maybe expand a little bit more on that and kind of some examples that you've seen in the organization.

SPEAKER_01 4:16

Yeah. Um, that was a big, that was a big moment for me. So um basically, the one of our lead content marketers who really touches every deck that historically has come out of the org started to realize that a lot of people were making PowerPoints mostly for internal use cases, but through Claude. And um, it caused like a moment of sort of concern and question around like, well, what is the value that I I bring? It was a little bit of like she had a bit of an existential existential crisis. Um, and that we were talking about. And um, that really was a big moment for me where I realized like her value is so much greater than the output of a deck. Like that is like a trade, perhaps. But um, that really she was able to, you know, enter a customer conversation and come out of it with three different ways that can change our messaging or change an ad campaign. Um, because something from that, you know, triggered a light bulb. And, you know, she's also able to ask that same customer questions that help guide her to that. And those are all like people skills, those are all strategic thinking and critical thinking skills that, like, sure, if you provide that to Claude or to Chat GBT, it'll give you a better output. But without you providing it that and without you utilizing those types of critical thinking skills, we're missing something. And we've been so focused on the tactics with AI that I just think we've been overlooking kind of that blurred line. So that was that was a big moment for me. Um, but I think I think to the second part of your question, I think we see this every day. I mean, I I get Teams responses sometimes of people that like clearly copied and pasted something from Claude. And I asked myself, did we critically think about this? I I use Claude for everything, like I'm not against it, but it's just uh it means a lot more when it feels like that person has kind of connected the dots in a way that an LLM can and uses the LLM as a tool, you know.

SPEAKER_00 6:10

Um Yeah. When we were talking before, I I was chuckling because you were saying, like, you know, people are leveraging Claude all the time, but it feels like the output is very mid. Yeah. As and I was just like, you know, and I think what that means is there's like a human taste component that is missing when you can so like acutely tell that a machine wrote it, right? Yeah, exactly. One of the things that we were talking about before was what a PRD used to be. And I was kind of giving uh giving an example of like back in the day, a PRD was a distillation of an individual or product manager's experience, uh, opinion, and outlook on how a certain product or feature should be built, right? And when you read it, you were like, okay, this is something that this person has deeply thought about. And now when you uh read a PRD, it's just so very clearly uh outputted by AI and it it almost kind of um commoditizes it, right? It like makes it inaccessible in a way, like where you kind of remove that human layer from it. Um I find that really interesting these days because it almost feels like now there's a proliferation of documents, like a document graveyard that AI is outputting. Are you seeing that too? And how do you separate out the signal from the noise?

SPEAKER_01 7:31

Yeah. Um, absolutely seeing that. Um, I'd be curious about places that aren't, but they're probably just not using AI if if you're not seeing that. As I said in my article, I think like there's a few things that still very clearly matter and they come through in a document or not, right? There's there's taste, uh, judgment, critical thinking. Those are the types of things that like AI can't really replicate. Um it can't tell you um what decisions it it what decisions uh you it can't you give it the decisions that you made, the critical thinking, the taste, and those types of things, and it can work off of that. It can give you 10 different versions of how to say something, but what you tell it to say is what comes from you, right? And that's the part that you have to own. So um AI slop is a very real thing, and and uh across my team and the people I work with, I I very often I'll ask a few things when I think this is like when I really want to get a deeper understanding. One of the silly things that I'll do is just say, what was the prompt? Like, tell me what did you give it? Like, what did you feel like was really important for it to know for you to produce this thing? Right. And like that's what I really want to know, right? I really want to know what where did you source this stuff from and why did you think that was the important context for it? What I think that does is it allows people, like that's enabled people to start to think way more critically about their prompting. Um, and the output of AI is a lot better with the documentation when you do it. Um and I've even considered asking people to put the prompt at the top of the document. I haven't done that yet, but I have thought, isn't that kind of a funny idea of like just link to the cloud thing? You know, just link to the the chat.

SPEAKER_00 9:14

Um yeah, I bet the first time you did that, that was kind of uncomfortable, being like, well, what do you mean? I of course I wrote this.

SPEAKER_01 9:23

It's just a little vulnerable, don't you think? Like um it gets you to that kind of like um it cuts through the mud pretty quickly to get you to like, well, what was important for for the construction of this document? Um yeah, and then I also think, you know, like some of like people's deeper understanding of things and like the ability to demonstrate these critical thinking skills and stuff play out over the course of weeks and months. Like in the context of a product manager making a PRD. Well, three weeks later, when this new product roadmap conversation happens and there's some question around prioritization and why we did what we did, and you know, hopefully your product manager is like, well, hold on, we produced this whole PRD, like we moved forward on this thing. Is this not solving the same problem or or what? Like those types of things really reveal themselves where like they did truly understand like the concept on a deeper level. And obviously, you want to reduce finding out about that three weeks later.

SPEAKER_00 10:20

You kind of mentioned before, like, show me the human decisions that are invisible in this document that that made this output. And oftentimes when we're working with AI, there's a series of iteration that happens. There's the original prompt and then further prompts to to then produce the refined document. Yeah. And I think that's okay because nested in that chain is a lot of that what you're talking about, that human taste, that opinion. And then AI really just becomes a co-author with you in it. And what I find is the best pieces of work that I do that I leverage AI for take a really long time, you know, because you are kind of uh you have a lot of opinions, you're refining, you're refining, you're refining, and finally you get an output where you're like, wow, you know what? That took me probably just as long if I would have like written it properly. Um, but it's definitely now grammatically correct. Yeah.

SPEAKER_01 11:15

So yeah, exactly. Like it's funny, like what you did before is when you truly authored the thing yourself, is you'd go through and you'd write it all and then you'd like reread it a couple times. Now you don't write any of it and you edit it a hundred times, right? Because you're kind of like having this back and forth of like um those, those taste, those judgment calls, the critical thinking, all those types of things that the LM didn't consider when it authored it, right? Um, or maybe not author isn't the right word, but you know what I mean.

SPEAKER_00 11:45

Yes. Yeah, for sure. Um I'm curious, like, you know, being uh in this space uh and also with the go-to-market experience and everything that you're doing, yeah. How do you view um like human judgment and working with AI in a world where you know everyone else is doing it and how you uniquely stand out? You know, I think that is something that I certainly ask our teams about because everything is starting to sound the same. Everything is starting to kind of look like the same shape, that a conversation exactly about what we're talking about, this taste and this judgment and what makes us truly unique, is more important than ever. So I'm curious as to your take there and how you kind of direct your teams accordingly.

SPEAKER_01 12:29

What matters is my ability to look at what else is out there and perceive what else is out there and find an angle that makes us different and articulate what that is. The LLM will help me with 10 different ways of how I can say that thing, or 20 or 50, or however many you want. Um, that might be different grammar or they might be whatever, but it's really important, I think, to um, and especially in the context of marketing and positioning and everyone's messaging is starting to look like the same thing. I'm totally seeing that. It gets back to making a human call. If if messaging is the thing that you own, and as a product marketer, I own messaging. Um, it's gotta be something that I own, not Claude. And I think that that comes down to me doing what's necessary to writing, you know, proper messaging and still using, you know, LLMs as a tool. Here's a funny thing that happened very recently that your question made me think of. We had this debate around, you know, we generate these um graphs, these graphs with these apps. And the graphs, we wanted to kind of upgrade the baseline. And one of the versions that we did was like this 3D histogram. And so someone, someone looked up, like, hey, there's someone went and asked Claude, like, are 3D histograms like an acceptable, like for data, you know, experts, is like a 3D histogram like a reasonable thing. And Claude was like, no, this is like no one really does that anymore, blah, blah, blah. And I actually responded, if anything that makes me want to do 3D histograms for this chart more, you know, like it sort of is starting to have this like opposite effect for anywhere where you want this like creative ingestion to whatever you're doing, you know? Um, and anyway, your question made me think of that because it's starting, it's starting to have that like you know, reverse effect on on some decisions that I make.

SPEAKER_00 14:17

Yeah, that's so true. Yeah. Um, that is so true. And like some, but and sometimes even you'll say, like, well, the did this company do it and aren't they successful? And it'll be like, you're right. I didn't think about that. So exactly. Um I'm yeah, in a product marketing context where, like you're saying, your angle, your unique approach has to be very uniquely yours, and and you have that ownership. How do you specifically think about your relationship with AI? Like, how do you use it? Um where will you kind of what work will you delegate to it? And where do you exact your human judgment and taste?

SPEAKER_01 14:59

Yeah, I think it depends on what the specific like task is, but um I try to use it as much as I can, like a way to, I guess, like here's a real example of something that we're doing right now. We're researching competitors. And Claude is um, you know, I very quickly wrote an agent that allows me to go through all my competitors' G2 reviews and derive sentiment and like what the common threads are and things like that. Very easy to do, allows me to get quickly to like what public information is out there that I could, you know, quickly get to. Some of these competitors don't list pricing. So the LLM is not gonna help me really figure out what pricing is. So how do I go find pricing? Right? How do I, the LLM can't sign up into the platform in a free trial and start using, well, I guess it can in some instances, but I'm not gonna set it up to go, you know, use a platform and tell me about the features. So I while I start my research with some of the sentiment on competitors and you know what their messaging is and all that public information, there's still a whole slew of things that are much better if I go and and get my own uh foundation, my own opinion, my own taste uh on. And that's things like, well, let me sign up for the self-registration, go through the process and see if I know what to do, I know how to build an app, you know, something like that. Let me um let me go talk to people in this space. Maybe I go to a consulting firm, whatever. How do I find out what how they're pricing this? Maybe I maybe we need to have our salespeople start asking these qualifying questions so we can get better ideas. Like those are different ways where it's like that's one problem of just research of how I'm using both LLMs, but also kind of leveraging the human side to where I really need to make sure we have like a strong foundational opinion on it. Um, and it's almost like the LLM is like, it's like it's like data points to support your thesis, but it's not your thesis, right?

SPEAKER_00 16:53

Yeah, or or kind of like directionally where you might look and explore a little further. I mean, one of the things we talked about is like it might give you like a very like dry breakdown of a competitor or a company you should be looking at. But sometimes looking at YouTube and at a video that has like 25 views, right? That kind of goes through something might give you actually the most insight that you need. And that's just not something that the LLM is going to pick up because it's like this probably isn't worth uh worth me servicing. Exactly right. From a team perspective, like in terms of how you work with your team, you talked a little bit about kind of questioning the prompt and questioning their approach. Do you find now that you have your own kind of human layer harness around like, okay, we are going to use AI, but this is exactly how we're going to be using it and what we're going to be using it for?

SPEAKER_01 17:43

I would say it's evolving. There's there's both the element of like you want to drive as much AI use case and adoptions as possible, but I think you also need the checks in place to like validate everything we talked about, right? Um I think that what I tell people is that it's great on my team, like it's great if you can produce this document. This document serves a purpose for many cross-functional reasons. I typically am asking them to edit it down, more concise, more concise, more concise. I do find that as a consistent thing. Um, but really what I want to know is can you articulate what's in the document without the document? Like if someone, if a salesperson asks you, I'll use the research example. If a salesperson comes up to you and asks you about something about a competitor, do you feel like you can answer it on the spot? That's the real test, you know? Um, and so that's what I ask of people when they produce, if it's, for example, a competitive battle card or something, that's what I would ask them. Um and so um that's that's typically like a framework that I have around. I don't I I'm not doing the like I mentioned the example earlier of like show me your prompt. I'm not quite doing that yet. I don't, I think that's a little bit much, or like maybe a little uh gimmicky in a in a way, but uh I think that either way, I want to I want to know that the people on my team that are using AI are feeling like they can pass that sort of like water cooler test, you know, of being asked about it.

SPEAKER_00 19:06

Yeah, absolutely. I think that goes to this whole notion around understanding and that carpathy uh quote that's floating around. You you can outsource thinking, but you can't outsource understanding. That's such a big part of it because it's not only the taste that you bring to something and bring to a prompt and bring to a project that you're leveraging and building with AI, but it's also how much did you understand from an output of someone else? Like, are you taking the time to digest it? Because I I even feel like today there's one part is the output, but another piece is now the the way we consume things because of AI. We'll get sent something and then we'll like put it into the machine and just say, like, hey, give me like the bullet points. Okay, sounds good, right? And that on the consumption side, it's Jen like all the nuances that I think really make something unique and make an article stand out, are being lost by the filtration of like give me a one-page brief on this, right? Do you find this as well?

SPEAKER_01 20:08

Yes, totally. Um, and it's funny you bring up the point of what is anyone who's using the LLMs now, like this document was produced by an LLM and now it's also being summarized by an LLM, so the person doesn't have to read the whole thing. Like, there's just so much irony in that. It's hysterical. Um, and yes, I find that all the time. I do it my, you know, I'm sitting here guilty as anyone. Um, I do it myself. Um and I think um to your point, it's it comes down to like the real understanding of things. And it I feel like we're gonna get to a point where less is more again, less has always been more, but um, LLMs really love to add this like more is more type component to it. And did 10 page purities ever really need to be 10 pages, like even before LLMs? Probably not. Like you uh you could probably make a good argument before that it didn't. And so let's focus on that, right? Let's make a one-page purity um that gets that cuts through the mud and really gets to the point. Um, maybe one page is ambitious, but you you understand my sentiment.

SPEAKER_00 21:10

Yeah, for sure. Where's that famous design quote, like great design is when you have nothing left to take away. I think that's very true of like uh any sort of uh content or how you present something on you. I think I think you can only get to something, uh something's um or a thing's essence when you truly understand all of the edges around it and can remove it. And AI honestly can only do so much, right? Like you that that human understanding is needed to really refine it down.

SPEAKER_01 21:40

Yep, I agree. But the consuming it part is interesting because you still have to read it, right? I mean, like at the end of the day, you can ask an album to summarize it, but are you gonna miss it? And how is it is it worth it um to miss it?

SPEAKER_00 21:53

Yeah. Here what I found, I'm curious as to what you do. Um for the longer form. Stuff. I still like I kind of have ADD when I'm reading like a really long thing. I get lost and distracted by something. But what I will do is I will copy the text and I'll have it dictated to me. Um and that, like I'm like very much like a podcast person.

SPEAKER_01 22:16

Yeah.

SPEAKER_00 22:17

I can listen to it and like understand all the nuances. And the dictation is getting so good these days that I'm like, okay, it's just like they're talking to me now. And I will understand the nuances that way.

SPEAKER_01 22:29

That's really interesting. I want to try, I'm going to try that a little bit more. Um, I've done a few, I've played around a little bit with like notebook LM before, and I'll like I've had to like turn documents into like podcast interviews. Like I've I've simulated like interviews, I've simulated like conversations through a doc, like a single document that's not even a dialogue-based thing. And I found that to be like kind of interesting. Um, but I haven't done that habitually enough to really like stand by that method. But um dictating a document makes a ton of sense to me, especially if you're like driving and commuting or something like that. Um what I like to do for longer documents, I'm very commonly asking the LM to go through very step by step. Like a part of my initial prompt is like, I want you to first go through this section with me and just summarize this in detail and in whatever detail I want, you know, but just go step by step. And I find that I'm able to more quickly get to the brevity that I want and need for whatever, however much I value whatever this thing is. Um, and that just helps me, it takes a little more time. I typically find like, you know, I mean, I have like probably like four clawed uh chats right now that are still ongoing chats of me consuming something bit by bit, inch by inch, you know. Um and I think, but I think those little like breadcrumbs or nuggets allow me to kind of get what I need in the instance and like move through it without, I don't know, asking it to summarize such a big document. It I find it often like is very inconsistent or doesn't quite know, doesn't quite understand that point that you made of like what you really want to know, like what the real human parts are, what the real meat is that you do want to know.

SPEAKER_00 24:08

Regardless of what works on the consumption side, I think that where where AI is really fantastic is for literacy and understanding. And if you're using it in a way to where you can increase your understanding in ways that you really couldn't before, um I think that's the best use of AI and something that I have found. And I can embarrassing embarrassingly say, like, I don't understand this concept. Explain it to me like I'm five years old, please. Yeah. And then I can show up and sound confident when I'm when I'm talking in public.

SPEAKER_01 24:39

Sure.

SPEAKER_00 24:40

Based on everything that we're talking about in terms of like how you think about team building, it obviously has implications around like the type of people that you'll hire and their approach and how they view working with AI. Maybe talk a little bit about that.

SPEAKER_01 24:58

I think people's viewpoints of AI, different perspectives. Like, I don't I don't mind people having different perspectives. I do think that like if you're not at least interested in it as a tool and how it can change and augment your job in a different way. I think that I would be concerned. Um, I guess I I don't know if there's a better, there's probably a better way of putting that, but I think like most people are finding like the faster you can get to finding its limits, the quicker you can get to what part of your work isn't visible that really matters to the point of my article, you know. Um, and the best way to do that is to use it every day. I think with with, I mean, particularly with interviewing, I think it's been interesting because like, you know, you ask someone to produce a PowerPoint now, you know, before or whatever, you know, the homework assignment, the infamous interview homework assignment, and then they come through and it's like, well, it's a lot different now, right? Like it's pretty easy to output that. The floor has been raised, but how do I tell that this person is going to raise the ceiling, like and knows what the ceiling is? I think is like an interesting question. Um, because it's easy to make a deck, it's easy to make a PRD, like we talked about now. It's easy to do some like top-line public information research now. So, how do you really like understand if you have the right person for the job? Um, and I think, you know, I've seen some companies come and invite. I'd be curious if you guys do this. We don't, but I've seen people invite people in now for like a week-long session. Like the last stage of the interview is you come in and work there for a week. And while it's a huge ask, I get it, because I think what you're trying to get to is can they do the human parts? Do they show up to a meeting and apply critical thinking? Do they connect the dots between things? You know, and so I think whether or not you have, you know, you fly someone in for a week or not is a different decision. But I think like it's getting back to the human parts of the job. And we have to sometimes interviewing, I think historically has been kind of lazy, where like, yeah, can they produce this document or this PowerPoint? You know, and a part of it is to help you because you only have so much time to spend with a candidate. Um, but I think another part of it is it is it's gonna ask us to get back to that a little bit more. Um, you know, how how well people can articulate themselves and and the extra mile is no longer that they made an extra slider to, like it's something else, right? Um so yeah, uh we think about that a lot. I think about that a lot. Um funny enough, I still have people make PowerPoints. I assume they make it with Claude. Um, that's like a baseline. The floor is there, I expect it, you know.

SPEAKER_00 27:29

But yeah, it's interesting. It's like a rediscovery of what makes us uh like human again and what our contribution looks like. And I think when people talk about hiring, they kind of talk about, oh, I definitely want someone that's AI native. Well, that's that's true. And I think that's to your point, someone who understands how to use the tool and leverage it and they understand that they exist, obviously. But there's all of this human stuff that is also required because you can have a very AI native person that may not be able to connect those dots, like you said, and may not be able to present on something or to show that they can understand and then apply that to business strategy. Um, so I think all of those things are very true. I'm curious, like working cross-functionally. You know, you're in you're obviously product marketing, the marketing organization. Um, as you have to communicate with product, uh, with engineering, how are you finding that collaboration? Do they share the same worldview? Are there ways that you found like uh leveraging AI has either helped or hurt that cross-collaboration?

SPEAKER_01 28:37

Um, great question. Um, I think I've found, I mean, like, you know, working with tech, the tech team is always so interesting because there's probably the the, I mean, developers are the fastest adopters of code, but there's still even skeptics amongst them, right? That feel um, and even within our org, we've had some people that have now come around to it, but at first were very skeptical of it all. Um, and funny enough, they've come to the same realization that we've talked about, which is like, whoa, writing code was never really about me writing the code. It was about the decisions I make to write the code that way, right? And that's like a really fascinating thing. So I think internally at our org, like cross-functionally, we for the most part think the same way about that, and people see that the same way. There's probably less meetings now where we review a document together. Like, you know, that's again, put it in the claude five minutes before the meeting if you need to. Worst case scenario, summarize it for yourself and show up prepared to talk about your opinions on this thing. And I think that's a little bit more of some of the changes that we've made. Like a lot more meetings are just about alignment on these things, our interpretation of it, and clarifying ambiguity amongst them. And so I would say, I would say that's a huge part of it. And it's still sort of subtle and nuanced, but that's probably been like the biggest change that I've seen cross-functionally. I will say, I'd be curious how you feel about this as well. I will say, like, I do find the skill level of AI tools really interesting cross-functionally. Like, developers are amazing. They are in cloud code, they're doing all this, you know, they have open claws set up at home and they're doing some wacky thing, you know, out of office with with you know open claw. Um, and and that's one tranche. And then like salespeople and BDRs and more like forward-facing people, I think are just still like I think they're still realizing in some instances how they can actually use Claude to research a company and show up really knowledgeable about, you know, like that learning curve, I think is still hitting some people that is fascinating to me. And I spend a, I mean, you mentioned I'm cross-functional. I spend a decent amount of time actually like educating, making sure that like other teams are educated on on how they can use AI better in their job, even though it's not technically my my day-to-day. I just think there's a lot that people can still leverage and utilize.

SPEAKER_00 31:02

I totally agree. I think usually there needs to be an AI enablement person internally whose job it is, is to make sure people are gaining the most that they possibly can leveraging AI. Usually someone in operations who's like very AI focused, um, that can either start with, you know, sales or marketing, company research, et cetera. You're right. I've definitely noticed that as well. Like developers are, you know, they're kind of this is the way that uh everything has gone. If if you're not using AI to do all of your development, you're already behind. And you also see exactly what you're saying, that domain expertise and that taste kind of made visible in the quality and the output. A very experienced engineer leveraging AI and agents, their quality and output, I mean, is uh exponentially more than a junior engineer, right? Like, and that totally makes sense. So you see that expressed in the development uh or engineering organization. And I think that um it just hasn't made its way, or at least more publicly talked about within the other organizations. And sure, there's tooling and workflows and things of that nature. Um, but I think we haven't found like the pattern of like, here's how a marketing organization should leverage AI, here's how sales should, et cetera.

SPEAKER_01 32:24

Do you think that a junior engineer who now comes up with AI natively, like they just graduated college and they're just coding straight up, you know, with LLMs? Do you think that in aggregate, those developers, when we look back 20 years from now, will be better at the human invisible decisions or worse?

SPEAKER_00 32:47

So this is where I stand today. And it's gonna depend on the environment that they're in. Yeah. But let's just talk about their discipline being a junior developer. Yeah. What they will lack is the muscle memory from all of the banging your head on the keyboard experience that all of these other more senior developers have gone through and understanding the nuances of why certain systems are built in a certain way and where the highest leverage use of an engineering's time is. I just fundamentally don't think that they will understand that out of the box. Now, if they have a great mentor and they're leveraging AI to upskill themselves and exponentially upskill themselves, then I think that they can get there. But it takes a certain type of individual to want to do that, right? I often question, okay, for a junior developer, those invisible skills, there's an agency aspect to it, there's a taste aspect to it, but there's kind of like a motivation and drive and a unquenchable thirst to learn that has to be there for a junior developer. It's hard to imagine what their value is outside of that. Yeah. Does that make that's where I stand today? I'm not sure.

SPEAKER_01 33:56

It does. It does. I stand in a similar place. I like the way you articulated that. It totally makes sense.

SPEAKER_00 34:02

I also I also think that junior developers sometimes may have a unique approach, less so from like there will be a world where all code is like written by AI and even this whole notion around human in the loop or revision probably ends up going away. So then in that world, what do developers become? It's really all about the taste at that point, right? And uh it's gonna be interesting to see what that evolution looks like. Yeah. Where people become more or developers become more product oriented because they don't need to develop anymore. Um, so it'll yeah, it's just gonna be very, very interesting.

SPEAKER_01 34:44

Yeah, I agree. I I think you brought up an interesting thing in in your previous point too, which is like environment matters a lot. Like who are they learning from? You know, who who are the senior developers and the staff, you know, engineers that are the ones teaching them? And are they are they distilling taste and critical thinking and those things in them as well? Um in addition to probably the biggest point that you brought up, which is like everything comes down to internal motivation and desire to learn at the end of the day, doesn't it? Um, and taste.

SPEAKER_00 35:15

Yeah. And I think that that internal motivation, like the men, even the mentorship aspect, it's like even I can speak transparently in our company, like there's a desire, but we're trying to move so fast and fix so many things that it's just like you don't you almost don't even have time to do it. You just have to like try to keep up. Like we're just trying to share as much information as possible, but like it's kind of like a sink or swim type of thing. Yeah. Um, I mean, I I constantly feel like I'm drowning and I'm trying to keep up with everything that's happening. Yeah. Uh, and one of the benefits of doing this podcast is I get to uh talk to thought leaders like yourself and learn like, okay, this whole notion, like, you know, my previous conversation was on like data and harness and AI governance. But I think this missing layer, this human judgment layer, is a very important piece around where do we fit in a world where AI is doing most of these things. And uh where we fit is actually what makes not only us unique, but what our output uh is unique. Because in a world where everything starts to look the same, um, you have to stand out, you've got to be different. And to your point about that 3D histogram, like maybe you do the opposite of what the LLM is telling you.

SPEAKER_01 36:31

Yeah. Yep. I think especially so creative functions, I think taste, like marketing and taste is is your number one asset now. Um, it really matters. Um and I think, you know, for developers as the other comp that we gave before, I think it's critical thinking and strategic decisions that you make to develop something one way or another. Um, and to some degree, that's also kind of taste. Um, so yeah, uh it's fascinating. I the way the world is changing is so interesting. And I'm not, I'm not the I'm not one to jump on the AI is gonna take our jobs camp. I'm one that thinks that it's changing, and I think it's changing in a way that allows everyone to be a little bit more human again, and we should all like relax. Like, that's like a relief for me. If I don't have to produce a 10-page document anymore, that's great. I just need to understand the parts of it that always were the parts that mattered, you know. And that to me is a better, better outcome than what existed before.

SPEAKER_00 37:35

Wow. That's a very powerful and and great place to start closing on. So um I guess one more question for you.

SPEAKER_01 37:43

Yeah.

SPEAKER_00 37:44

You know, we have a mix of enterprise architects, application development leaders, and technical builders that listen to this. If you had a one piece of advice in terms of where to start in their thinking of making that invisible visible or just an approach, what would that be?

SPEAKER_01 38:03

I would ask, I would ask people to, you know, evaluate the work that they do themselves and reflect on what decisions they made as a person and and get better at defining those things for themselves. Um to make sure you're valuing yourself first for what isn't visible. And then I would ask others, I would then I would ask them to start seeing that in in others um and start driving a culture that um makes the invisible work more visible uh pretty cleanly. Um yeah.

SPEAKER_00 38:44

Fantastic. And then finally, if people want to learn more about you, read more about your your musings and your thoughts, where can they go?

SPEAKER_01 38:52

Yeah, um my sub stack, um it's called somebody manage me. Um I think there'll probably be a link um somewhere in this. Um, and um uh that would be a great place. And then you can also find me on LinkedIn. Um, Paul Trafficanti is exactly how it is on LinkedIn. Um, I typically share, you know, interesting musings and thoughts and and link to my to my sub stack there.

SPEAKER_00 39:15

Okay. Well, Paul, this has been a fun and fascinating conversation. Thank you so much for your time.