Over the past week, a simple spark—trying Claude Code for the first time—turned into a full creative surge as Chris built, broke, fixed, and shipped two live applications while discovering what’s now possible when AI becomes your dev team.
This episode explores how fast capability compounds when you stop outsourcing your potential and start building alongside AI.
How AI tools unlock rapid learning through trial, error, and iteration
Why multi-agent workflows (Claude Code, GitHub Copilot, Codex, GPT, Gemini) are becoming the new norm
How to build functional, beautiful web apps without knowing how to code
What “capability building” really means in an AI-first world
How AI transforms documentation, diagrams, and past work into new products
Why self-service experiences (search, mega menus, assessments) are now trivial to build
The risks and opportunities for professionals whose value depends on others not having capability
00:00 — Opening + Nico’s recovery
01:05 — Chris’s coding breakthrough week
04:10 — Using Claude Code credits to build fast
06:25 — Mega menus, UI fixes, and debugging with agents
10:40 — Automating YouTube ingestion and leveraging GitHub workflows
14:15 — Building contributor pages and adding search/breadcrumbs
17:00 — Learning to trust the tools and push through breakage
21:20 — The great repo collision: breaking (and recovering) an app
26:45 — How capabilities compound once the foundation is built
31:30 — The future risk for people who profit from others lacking skills
35:40 — Why AI toolkits differ for every person (GPT, Claude, Gemini, Perplexity)
41:10 — Preview: AI turning calls → proposals → kickoff decks → slide builds
Claude Code for Web
GitHub Copilot & GitHub Agents
Vercel deployments
Astro, React, and Vite frameworks
Codex (OpenAI)
Gemini
Perplexity
Automated workflows (YouTube ingestion → site → pages)
Value Path systems mapping
AI-driven diagram generation
Multi-agent development pipelines
AI is now a capability engine, not just a tool. You learn fastest by breaking things and letting AI help you fix them.
Trust builds speed. Once you trust the tools—Claude Code, Copilot, Codex—your velocity skyrockets.
Your existing docs and diagrams contain huge untapped value. AI can turn them into apps, visuals, and systems.
Developers and non-developers now work on equal footing. The limiting factor becomes imagination, not skill.
If your business model relies on others lacking capability, you’re at risk. AI removes that gap instantly.
AI is best used as a team, not a single tool. Everyone’s stack is unique because each model does different things well.
Chris: Good morning, LinkedIn friends, Value First Nation. Welcome to another episode of Value First AI Daily, your collaborative AI intelligence report. It is Monday, November seventeenth, twenty twenty five. Happy Monday, Nico. Good to see that you're feeling better this week.
Nico: yes thank you so much um sorry to everybody for last week but uh came down with this with some sniffles with a little uh sinus annoyance and uh i'm probably here that it's just on the tail end i'm feeling much better so back at it and uh yeah not soon enough i gotta tell you a lot of lot of stuff between last week and this um well i need to i need to say how are how are you doing i know you got some news over there i know i know you've been working chris has been putting in some time and uh why don't you why don't you show us what you got cooking
Chris: Man, I need to thank you, first of all, for unlocking the vibe coding part of my brain last week and over the course of the week. It's just timing in all things, right? You need more time, don't you?
Nico: Well, no, just like whether it happened at the right time or not. I am. And... Uh, like the timing this time was, you know, um, odd code for where I would come out recently and there's a credit to use, but also, uh, you know, paying for my own like individual account that I've never wanted to like get rid of, even though I probably should have a long time ago. Cause I have a team account, but trying to get the cloud code seat on the team account would have been a hundred and fifty bucks like a month. Uh, so luckily I had the, uh, cloud code or I had the cloud like personal account ready to spin up and, and because I could, I was able to start cranking on that, um, on that. a two hundred and fifty dollar credit, which I'm halfway through, by the way. I think I spent like maybe sixty bucks since we talked yesterday.
Chris: Yeah, I've been waiting for one of those nights to come where I just get comfy on the couch with the laptop and just, you know, have some TV on or doing something and just cranking.
Nico: So this is one of the projects. And by the way, for me, this is just stuff I've probably wanted to learn in the past, at least to understand how it all works instead of being so reliant on others. And let me tell you, when you're breaking stuff, While doing these kinds of things, there's no better way to learn how it all works. So we've got this site in the works, which is up because I just want to like, let's just publish and ship. It's not waste time for the value first. stuff that we've been doing. And, you know, one of the last things I was working on last night, or actually this morning was a mega menu, right. And so just to show it's not all like, this is like, it's not all magic. It's not a genie in a bottle. Like, so this is what the mega menu looks like right now.
Nico: All right, like, not good. So a couple more prompts. This is prompt number three after the first two times he created it and it wasn't showing up at all. Nothing was showing up, right? So that's been the thing over the last twenty four hours is just trusting and learning how to triage with various agents and tools. But yeah, we're just getting started here. Really excited about this assessment.
Chris: That's really cool.
Nico: Things like this, you talk about self-service tool experiences on websites. It's never been easier. One of the next big steps I got to do is connecting tools like this to the CRM in HubSpot. And then see if I can click into it. This is something I'm really excited about. Like, auto, it's got some placeholders on here right now. But one of the shows, let's see here. I was able to automate the suggestion from YouTube. So now it's really just getting the dots connected from all the things that already exist. And this is Workflows and GitHub doing that. I've never done that before, and I didn't that sounded real scary to me, honestly, when the first thing I started working when I started suggesting workflows in the GitHub repo area.
Nico: Let's not, let's not do that. That's kind of thing like, you know, you tell people in HubSpot, like, like, unless you know what you're doing with the workflows that are going to impact like, don't do that. Right? Get a professional and it's like, well, Luckily, we're surrounded by professionals now with our AI friends. And then we've got some nice contributor pages. I've got to figure out how to deal with media, like getting pictures and stuff going. But we've got a nice page here for Nico. We'll be loading this up with our episodes. Added the search bar this morning. uh, I finally unlocked the part where it's like, Hey man, go ahead and take a look at everything and tell me what can be better. Right. Like, well, how about a search bar? How about, uh, how about, uh, breadcrumbs? I was like, yeah, let's do it.
Uh, So that's project one. And I'm probably just as excited. This is where it's like, I need to get this up for businessy reasons, obviously. Give people a place to go. And oh, I haven't even seen this since he last updated it. This looks pretty damn good.
Chris: Isn't that awesome?
Nico: We get in there like, oh, shoot. Oh, yes. OK. OK. It's like, that's what happens when you just start. When you trust the system, honestly, you're just like, okay, it's ready. Push, pull request, whatever. Just start hitting buttons. Now, I did experience how that doesn't always work well. And so the other thing I've been excited about Is this, I showed off, showed this off a lot last week. Uh, this has been so helpful in terms of just creating visuals to share, like during all the content, we're trying to help people do things a little bit better when it comes to scoring and, you know, customer life cycle management, things like that.
Nico: Well, uh, over the course of yesterday, well, here's what I did. Right. where I broke everything. So these apps, these have been like two separate projects, right? Two different repos in GitHub. And I'd come a lot farther with this one to the point where it's like, all right, we can make a design guide. Uh, and here's all the settings and like, and so then I go with the, with Claude and I say like, all right, let's document all of it. Um, fill in any gaps that you see like value first language, stuff like that.
Nico: I bring it over to this, uh, repo, uh, using, uh, GitHub co-pilot and agents that I was still learning how to use. And instead of, you know, um, trying to pare down the part. So this is a learning app, literally a simulator, not a content-heavy business website, so completely different purposes. And there's different frameworks built underneath. This is using Astro primarily, and this was a React, Vite, or something like that app. Mm-hmm.
Nico: So I take the design guide from the website, and I just give it to GitHub's co-pilot and say, apply to this site. Apply to it. Can we just use this? Well, I couldn't see anything after doing that. Like, zero. Like, broke it. Just black page. And it's been a journey to get back. But we got back without having to just go back to the last publish.
Nico: I actually asked Ryan at some point last night. I was like, how do I go back to a previous Vercel build? Because I wasn't sure I was going to be able to get back here. Cause it was like the first step back was just like one page worth of blinks, like just down. Right. So no styling and structures and stuff. Yeah.
Nico: And I had to trust since we started from a place of all of this content and all of these instructions of how to operate and things like that were in the repo. I was like, okay, I just need to like step by step with the co-pilot, with the agents. At some point I had to realize that those are two different things. Um, uh, cause co-pilot kept telling me he couldn't do anything. And then he was just spinning up agents to do all the stuff. I was like, oh, okay, I get it.
Nico: Uh, and, um, And I woke up this morning to getting to this part where it was like, okay, the styling's back. Like, the app is back. Right? Because before that, it was like, oh, man. Just, like, it broke. Broke everything. Nothing was working. And it didn't necessarily, like, make sense as to, like... Like you would have never thought that the things behind the scenes would have looked like the way that they were looking just because the first shot that Claude came out with last week was like close to this. Right?
Nico: So it's like you just skip ahead. And then when you try to get in there, like the way that you do that, like there's a process, right? And luckily, like again, half the battle is just believing that you can get to the end with the tools that are around you, but also understanding how all the tools work. And like I said, I was losing steam in terms of whether I was going to get back to this. That's when I asked Ryan, like, all right, maybe I just need to go back.
Nico: But the thing is, is when you build it back up, I'm in a place where I probably use Like I'm super comfortable using GitHub Copilot now to manage and understand this app, which needs to be managed differently than, you know, this website. Right. Right. So it's been a journey for sure. Um, but man, one week and I've got things out here that have been in my head, you know, forever. And like, once this foundation, man, this is how bad or good it is, folks. It's like, maybe we can go without another, without another episode today. Maybe we can start tomorrow.
Nico: It's hard to tear myself away from this stuff right now. I've never felt so powerful, but also so hopeful for what I'll be able to accomplish. Like, honestly, like, you know, The thing that the automatic YouTube ingestion and like the GitHub workflows opened up for me or put back on the table is like, okay, once this settles, like, are we getting back into NNN? Like, are we getting back into that next level, right? It's just this foundation. It's this, you know, capability.
Nico: And it's a lot, you'll find a lot of this language on the Value First site. It's like, If you're paying lots of money or giving lots of time to somebody else who will do the thing for you or accomplish the task while not building your own capability, you're just missing a huge opportunity right now. And I think from the other side, if you're a person who profits off of people not having capability, You've never been in a greater spot of risk because it's literally like AI allows us to go from I can't do it to I can do it in a matter of seconds.
Nico: And in a way that it's like, I didn't, it's not just luck, right? It's like, or maybe it is lucky, but it takes no effort for me to try again. and then again, and then again. And as I get more comfortable, more trusting, It's just that it serves this human will right now, which is based on response time and relevance and meaning that we've just never gotten from this space that is full of SEO gurus, to be honest.
Nico: Right. So resist the urge, folks, to go and find that AEO guru and figure this out your damn self.
Chris: yeah i think um i think it's one hundred percent correct once you it's it's the same as when we all started getting into like all the different platforms right like migrating away from gpt for a bit and trying gemini and and using claude um and i think that's really the it's kind of funny i think um It's unique in the same way that any other product is unique, right? So people are wondering about, well, in a world of so many different tools, how is it gonna be possible to determine which is gonna be the best, which one to use? I don't know what to tell you.
Chris: I'm still extremely agnostic, but that's only because there are features in each one that I use for very specific reasons. So it could be that I use Gemini maybe five times a month because of whatever I'm doing that I need Gemini to do that thing for. I might use, in order, it's probably, yeah, it is actually GPT-Clawed Gemini, just in that order.
Um, and perplexity kind of just weaves its way in there because it's like the search backbone for all of them.
Chris: Um, so it's like, that's, that's what my AI team, uh, just for like normal toolkit looks like, but it's been interesting to see that like George is a huge GPT fan. You're a huge Claude fan. Um, I don't think I have any huge Gemini fans yet. I think we will soon.
Chris: And I'm really sure. I can't say for entirely sure. But given that the trial's over on the eighteenth, I'm guessing that cloud code for web drops tomorrow. Sounds appropriate.
Chris: The only reason I'm pretty also sure about that It's because every time I try to get into it now, two things happen. One, I get greeted with this, which I have not seen before, but it's certainly some indication of this all transitioning here. Whoops.
Get the bus out of the way.
Chris: um so i was trying to work on a project hopefully it'll it'll do it again it popped back there's there's like a new there's a login now not a login there's like a start screen so it's and it's branded and everything i wish of course when i try to do it it won't pop up but I haven't seen it yet this morning.
Chris: Let's see if I can... Of course. Yeah. I actually used... I touched Gemini, Perplexity, Claude, GPT, Codex for the first time, right? Yeah. And that was me, like starting with Claude and Claude code and then wanting to do more concurrently. So it was like, all right, uh, Claude, give me a prompt for GPT. Give me a prompt for, for, uh, Gemini to do this app, uh, to do this piece of the app. Right. Right.
Chris: And you use them as like your dev assistance, right? Right.
Nico: And if I could show that it's just like, again, we talked about this a lot, like proof of concept, like getting to that just really fast. Right. As I was thinking about the media hub and the assessment actually like this thing, like I got this done and then I just gave it to cloud code. to put it in the website. And it didn't start like this, of course. This is where you work with Cloud Code to get it into. And I started making sure the design guide was applied.
Nico: But the way this stuff can work together is just so freaking powerful. Uh, the other thing, so what, so of course this could only happen if I had some content done, like I'll keep reminding that. So all of the stuff that I've been doing club projects and documentation of everything is now turning into other forms of value.
Nico: And, um, I want to highlight that just because like, Oh man. you've probably got a ton of value just sitting around you and documentation or diagrams um and let me show you how like how easy this could be right so this is a value path systems map what i'm trying to do is say okay this is a long damn list of touch points and it's awkward and Right, yeah, I got it to collapse and expand. But what if we could just see it on a visual map, right? This is like a step in that direction.
Nico: But where this came from was, and I'll pull that up and we'll close it. Oh, no, you're good. I just knew that when it wasn't going to be shown, that it would actually show up. Right, so let's get yours up. So I'm really sure this is going to be a scratch page. That sounds fancy too, doesn't it? Create your first cloud environment. Talk about democracy.
Nico: I never in my life thought I would ever say I'm going to create a cloud environment. Right? Right. And it's like, I mean, if you've tried Codex, Um, it's not the same thing at all. Uh, it's so much less intimidating.
Nico: Like the first time I tried codex. First of all, that I'm aware of, I don't know if they've changed it since then. I use it in Visual Studio, but that's only because Copilot allows for a conversational sort of aspect to it. But in its native environment, Codex is not a conversational device. model so when you ask it to do something it just goes and does stuff even if you ask a regular question it will still go through your repo before it comes back and gives you an answer which could take it anywhere between whatever five ten minutes
Nico: this is entirely different in that as soon as you type something in and and you go like you're just naming this environment whatever it's going to be um So I know it's already there, and I don't want to create a second one. And then your options here are basically how much internet access you want to give to Cloud Code in order to work on your project. Developers know this a hell of a lot better than I do.
Nico: So you can either lock it down entirely, which would be the equivalent of developing locally. You can use trusted networks, or you could give it full-on access. Again, I think it's a matter of maybe you don't want to do full access because security risks. It could go pull bunk code from some social post or something like that. Who knows?
Nico: Basically, it could do something weird in order to actually help you or solve your issue and then ends up actually screwing up your project as opposed to helping you. Yeah. See, that's the other thing. It's it's slammed right now. It is getting slammed so hard. Like I can't log in. I can't do anything. I launched.
Nico: And the only, the reason I know that is because at seven AM it was beautiful. There was no problem at all. And now that it's eight o'clock hard stop, like good luck using this today if you don't have your API. So like if you're outside of cloud code, yeah. Okay. You stand a chance, but if you're in cloud code, I don't know that you're going to get to it today.
Nico: I just got a retry connection, damn it. Yeah, real sorry, peeps, but that's the way it goes.
Nico: And if you want to actually use it outside of quad code, Visual Studio and Cursor have really nice plugins where it's essentially the same thing. It's just right in your IDE.
Nico: Yeah, so this definitely... It was a launch point for me, right? So I probably, there was enough friction where if this wasn't free, I probably would have done more to figure out the other tools that I would want to use. Like started with the IDE Visual Studio. I liked how easy it was in Cloud Code for Web, but also Cloud Code for Web was free. So it's like, let's just go nuts there.
Nico: You know, you run into a bunch of, you have to believe in a lot of things that aren't apparent on the screen. Like, are we refreshing right now? Nothing's happening. Oh, I can just type a couple of characters in the reply. Oh, there it is. Oh, everything got done. Cool.
Nico: Uh, yeah, I had no idea. I thought we were frozen. Um, and just, but just to unlock the possibility is really what this, this credit has done. Um,
Nico: And, you know, again, like, it's the worst it's ever gonna be. I know now that I can get past a lot of these, you know, barriers.
Nico: And, you know, to close, like I came here, right, this map. Yep. Like, this is like, five years in the making. And what I did to get here, this is where I think it was Codex. I was like, Cause one thing I, again, there's limitations all over, uh, Claude code for web. I can't get pictures too. Yep. That's super frustrating because I love communicating via pictures and guess what AI does too.
Nico: um, so, uh, so when I discovered codex, I was like, all right, let's just, let's, let's go all out.
Uh, can we take this thing? And let's build it into the app. So this is literally a map of if you're tracking all of the things, and you go to market, and all of the stages, and all the ways internally and externally, the signals that are on the other page should show up somewhere on here to give you almost like a heat map is what you should end up with, whether things are working well or not.
Nico: And that's where to start. this is what it turned into, right? And that's where all the dummy data is in here to actually like, like all of these different things would have, like, are so many different capabilities, like coming together, like coming up with a good data set to even make this stuff make sense. Yep. Is it is a challenge even for AI early on, it was hard to like, you know, get meaningful dummy data together. Right.
Nico: And now it's like, I didn't, I didn't come up with any other dummy data. Um, cloud code did right. I just said, here's the brief of the app and let's add this kind of thing. Let's add this. Oh, I want to add manufacturing. I need to, I want to capture these seven kinds of revenue types. And then all of a sudden, we've got, you know, twenty touch points that are focused on the back end.
Nico: Notice how this changed to, like, internal and, you know, back end stuff, back office stuff. Right. It's a new world, man.
Chris: I know. Like, now, like, it's so... That's so different.
Chris: And tomorrow, I want to show you guys something I put together that basically takes your sales call, turns it into your proposal, takes that, turns it into your kickoff deck, and then takes that and actually builds the deck.
Chris: Sounds pretty spicy. Join us tomorrow if you want a piece of that. Everybody have a great day. We'll see you next time.
Nico: See you guys.