Time to Talk about the world of AI!

Why is this.

Good morning, LinkedIn, friends, Value First Nation. Welcome to another episode of Value First AI Daily, your collaborative AI intelligence report. It is Friday, October 24, 2025, and on Fridays, Nico and I get to hang out with Mr. George B. Thomas. How you doing, George?

Howdy. Howdy. How's everybody doing?

Doing good, sir. How you been?

I have been pretty good. Had vacation last week. Been trying to catch up this week. Been looking to this conversation because I don't know exactly what I missed with AI, but I'm sure I missed some things while I was on vacation.

Just a few things, but I wanted to find out, did you get a chance to see Tron Aries?

Oh, yes, sir. Yes, sir. We made it a point to see that before I left on vacation. And I. I personally loved it. So I don't know what some of the people are talking about. Maybe they bumped their head on the bathroom sink or something. But I. I quite enjoyed it. Did you see it?

Yeah, yeah, we saw it twice.

There you go.

I always like to. Especially with movies like that, I always like to see it twice in the same weekend just to, like, make sure that I didn't miss anything, let's say.

Yeah.

And yeah, to me, I. I agree. Like, Leto played.

He.

I mean, he's kind of like that as a person anyway, like, in real. In real life. So I thought he was the perfect person to play that character. And I was like, oh, cool. He's going to bring the exact same. I. It was great that he did. He's going to bring like that exact same character that he had when he played Wallace in 2049 to this role because it's also sort of. It's not post apocalyptic, but it's like future sci fi role.

Yeah.

And yeah, I thought it was great just to see him. Like, he just, he did a great job of like, sort of breaking into being human and then it doesn't really happen until the end. Like the last, like 10 minutes. You're just like, oh, okay. Like, now he's starting to, like, really let go of, like, the rigid structure and whatnot.

Yeah. Which, by the way, like, as I sat there, I was envisioning there are many humans that actually live life like he did in the movie, where they're not fully embracing the humanity that they have at their disposal, and they're just kind of living in this system, if you will. And so for me, it was an interesting kind of mind journey into the falsality of the Tron movie and the reality of the real world. I. I definitely will say too I did enjoy some callbacks here and there to the original and. And some things that they got got very right. So a great, great movie. In my opinion, some life lessons could be learned along the way if you truly pay attention for more than what it is just a movie.

Yeah. And while you, while you were gone, just couple. Couple things dropped. Except for you. So I mean, I know that we talked about Comet at one point. I know that we went over and we looked at it, but now we also have chat GPT Atlas. Yeah, or just Atlas rather. It's not even Chad GPT Atlas, but we've got Atlas. And then with GPT we now have GPT for business. So we have that same sort of bring the business into your chat context with GPT as Claude has. Not sure if the same functionality is there, but I read something this morning, I posted about it yesterday and read something this morning that I thought was honestly fairly profound because when it comes to this stuff, the number one, I don't want to say blocker, but like the number one thing that people talk about in terms of like, what could go wrong, the first thing that comes up is generally, well, if there was, and we kind of experienced it a little bit earlier this year, if there was a powerful enough solar flare.

And.

We saw what that looks like earlier this year when Spain got hit by that, and it was kind of out of nowhere that like, wow, an entire country has been out of power for like a couple days now and it's basically crippled the entirety of their systems. Right. So it's not like this little small joke anymore. And it was from, man, I hope, I hope I'm saying your last name right. I'm. I'm so sorry. And just ahead of time. He is not Russian, but his name is Boris Cunia. And he said, maybe it's just the cynic in me, but with all the browsers launching and more and more ways to personalize the experience, it seems like a race to who can capture the most amount of data. And that was only followed up by Chad Sedaris who says people saying if AI gets out of control, we can just shut it off. All forgetting exactly this. We are all embedding it into all parts of our business. So it's this race to get to let me consume as much data as you can. I just heard another thing this morning about Uber now tracking a completely different aspect of what their drivers do. So now they're no longer necessarily tracking driver data, but they're tracking, like actions, their Drivers take, which to me could be, I mean Uber drivers, I'm not sure about Lyft, but they have to have cameras. So they're collecting this sort of data on natural speaking, right through the driver and their passenger. They're collecting this data on sort of like human movement and human motion and reaction and that kind of stuff. It's, and it's gotten to the point where, you know, I've had a couple friends I've been talking to in the last few weeks and they're also, they're all talking about moats and they're all talking about like programming based moats. And I don't know, I've been saying the same thing for a year now. If you guys been ignoring it, not sure what to tell you. Data, data is your only moat. Whatever data you have, that, whatever data you have, no one else has innately, right? Because it's either specific to your business, specific to your line of business, specific to your type of business, your configuration of business, whatever it might be. But you know, that's the moat because if nobody can get to it, then you have to make an API key and then people have to pay you via tokens in order to get access to your data.

Right? And it's, and it's like a new layer of data too. And I said this a few months ago at profoundly as we were talking about our internal system set up and like how it was going to be usable for us. And I said, look folks, data capture at all costs, we've got to find a way to bring it all in. And what you're describing there is changing the way people are doing things in HubSpot too. Because that new layer is that behavioral unstructured data that is now accessible and consumable. AI is there to provide insight on what it means. People have been capturing a lot of that stuff for a long time. But now the sentiment scoring in HubSpot, the behavioral scores, it's pretty easy for me to see. Like, I mean you got to give AI guardrails to tell it like what you care about for a business. But you certainly should not be in my opinion, saying this web Click should be 1 point and this form fill should be 5 points. Because AI is just better at doing that in the context of when they clicked and how they clicked and what was happening before and after that clicked. Like it's, it's a new again, this is infrastructure. And it's laughable to think it's like saying you could turn off electricity.

Like, so I, I think There's a couple of things where my brain went here. One, I think that we have to add a level of specificity because when I heard the words come out of your mouth, data at all costs. I can hear a marketer in the back room screaming. But we just ungated everything. And that's not what we're talking about. We're not talking necessarily just about form fills, although that is a type of data. We're talking about the unstructured data. We're talking about the transcripts of calls. We're talking any, any way that you can get information that you once thought was unusable by mere mortal humans. You should be figuring out how to get that layer of data into your CRM or system so that AI can actually make sense of what we are not historically good at because we can't look at big chunks of data at the same time. So please open your mind to not just historical data conversion, but like conversational data and stuff like that. The other thing that I'll just throw out here real quick, Nico, when you're talking about like the solar flare, I'm a big fan. My wife and I both are a big fan of the 911 series. Like have. There's 91 1, then there's 911 Lone Star and now there's 911 Nashville. And the new 911 series came out recently. And what's funny is they have this, this episode where there's this solar like space storm and it knocks out these. And not to give a lot away for if you haven't watched this, but it knocks out a lot of these satellites. And like one of the funniest parts for my wife and I is when it knocks out these satellites. By the way, there are some humans in space. I'll leave that there. You can go watch the episode when this happens too. But one of the shots that they come to is the, the fire department comes and there's like Waymo esque style car driving in a circle because it had. And there's a person in there like with a sign saying help call 911 and the things driving in circles. And I start to giggle because I think back to San Francisco and I think about how much we're relying on AI and information and technology to do the things that we call everyday life. And when it breaks down, down how badly it could break down. Just the other day when I got back from vacation, like the Internet had a flu, like it was sick. There was all sorts of things that were going awry because of some DNS issues here, there and everywhere. And like, and so like this genie is out of the bottle, ladies and gentlemen. Like it. And I mean fully out of the bottle. Got its tennis shoes on, walking around doing what it wants to do. Like anyway, that, that's where my brain goes when I hear what you guys are talking about is those two different kind of parallels of we gotta pay attention because it is, it's just here.

It is and it's, it's definitely not slowing down. You know, like that's, that's what I was so not necessarily like blown away, but again like that it made me stop and think twice for a second about like, oh yeah, the, the fact that this stuff is like interwoven into everything that we do. Right. But at the same time it's like we've already always had like the same sort of natural fear when I think about it. Right. There are people who live off grid per se, and you know, those people are technically better off in the event of like, whatever, power outage, mass power outage, things like that. And so it's crazy to me a bit to think that like that's the only thing that we are really seriously dependent upon. Right. Is just so long as we have electricity. Right. Because then we can keep digging further into the earth for water or keep digging further into the earth for, you know, thermal dynamic, you know, thermal, thermal heat and, and power. But yeah, without that electricity though, one.

Of the things a point before you move on.

Yeah. Yeah.

So like when we say it's everywhere, I, I was like, what's the, what's the level to which I could say to somebody it's everywhere. And my brain works weird, but I literally just googled toothbrush with AI. Why would you need an effing toothbrush with AI?

However, to tell me when I'm done.

This website that's Feno and it's an AI powered 20 second brush. Diabetes defense, cancer defense, oral health, overall. Like, and it's this thing that doesn't even look like a toothbrush, but it's, it's an AI teeth cleaning brush device. So when we say it's everywhere, when we say we're embedding it into our lives, the thing that you're supposed to do two to three times a day now is AI powered, AI assisted.

Like so this manual thing. Right.

Well, also business, but we're just talking about life.

Yeah. And not even like that's not just because they're having fun or like they, they had a good idea and what and like the mouth and the cleanliness of our teeth and mouth, like is one of the main entry points for, you know, sickness. So. And like. So it doesn't surprise me that like the diabetes defender part. Right. Just being able to. It's like another way to analyze what you put in your gullet. And again, so much of this is just mindset because I just got a notification as around the show about Q 2.0. Like the AI OS. I haven't watched it yet, but we've been talking about this as the next layer beyond browser. When can it just be around us in our computer? One of these comments. I love the ui. It seems intuitive, but to call it an OS is a stretch as an OS is for running EXE and DMG files and including file managers. That's a very strict definition of OS or operating systems in my opinion. And as an example, I have a document that I worked on with Claude that outlines my natural operating systems as a human that does not involve computers like at all. So to suggest that to call it an oet, it has to be this super specific. Those are the kinds of things, if you think about it that way, you'll never get to the next step where we don't want to be running any files at all. We just want stuff to be happening as we see cloud code for web and just these things converging on each other to make the same functionality get better and be available in new places in new ways. It's like if you're going to use these old definitions and old. There's the keyword, you know, old constraints that just aren't there anymore.

Yeah, listen, it's. It's very easy for us humans to get stuck in history. Some of your identities are based on the history that you've had to live through because you haven't operated or, or optimized or installed the new OS to free you from your historical baggage. Oh, I'm sorry, Wrong podcast. Wrong podcast. So that definition is of history. It is of not where we are, and it is not of where we are going. And see, I think that's going to be the hardest part for people is to understand that life once in my mind was a nice lazy river. Ladies and gentlemen, you are traveling down Niagara Falls at the current moment until we get to a peaceful pond at the other end, hopefully.

Yeah, so show us, Nico. We got to see some stuff because I think it's always good to show some of the outcomes that we mean by this and even being able to compare it to what you were trying to do three months ago. There's just so much to learn. And like, the takeaway for me from this week, it's like we saw Haiku 4.5 come out and then code, Claude code for web. And it just should be a reminder, like, Claude code was just for desktop and IDEs, and. But because the model got better, more efficient, you know, faster, cheaper. Now it can be available in a browser because we don't have to worry about the load or whatever you worry about. And that now it's accessible to so many more people. I still haven't, haven't dug in yet, but. But it's getting closer and closer to that point where I feel like there's almost no barrier for me. And, like, it's just going to take one time where I pull it up and try an idea of my own, and then I'll be hooked. So, luckily, Nico's been hooked for a while and he's got some time off right now, so he's building some cool stuff. You're muted.

You're muted, brother.

I know. Oh, well, it's too much time off, to be honest.

Apologies.

I don't. I don't know that that's. That's a thing. It's funny because my, My wife was like, yeah, well, you know, I'll let you get back to work or whatever. I'm like, yeah, like, you really gotta stop calling it that because that's, that's for other people. Like, that's not for me. Like, that hasn't been for me for, like, the last nine months. Like, I don't. I don't know what, what that is anymore. And I'm not.

That is meaning work.

Yeah.

Yeah. Like, I mean, I'm still, like, orchestrating, so it's not, it's just very different. It's not really. It's just not really working.

And again, historical mindset, this is what work was. New mindset, this doesn't feel like work, but this is work.

And, you know, again, like, I'm not really sure how to best explain, like, that transition. There really isn't, like, an easy way for me to explain it. Oh, sweet. Made the update already. What you're looking at now is Claude code in the browser, as opposed to having to do it through the command line interface, which there's nothing. Oh, a huge difference. In other words, when you're typing in here, it is like, claude. So you can cut. You can use, like, markdown to help drive things, but it's not like you're going to Build in here. It's not like Claude, where you're going to build in here and then you're going to see it and you're going to be able to play with it from here. You do need to like know a little bit about, you know, hey, like how to work with. How to work with like GitHub, how to work with an IDE, just so that you can actually figure out like, oh, okay, well now I need to come in here after I've finished working with this quad commit and you know, check to see if the last commit was actually properly made or if I actually have to do anything differently with it. Let's see. Let me go to the actual latest. There we go. Yeah. And then from there like I download it to a vs, upload it to my server and then all of a sudden we have a HubSpot D and D game. And this was something that I started like yesterday. You can kind of see. I mean it's pretty if you want to see like how much work Claude code can do. Right. So it was just as a challenge. Do you think you could create a 4 player, didn't even spell right. 8 bit style dungeons and Dragons game that's entirely themed around HubSpot and loop marketing. So it went through. Absolutely. Here's my vision, four player game, blah blah blah, update to dos. So it creates its own to do list and then really does scrub through it. It's actually really amazing to watch this guy go to work. Like it. Like, I'm not sure what to tell you, but like I have abandoned replit and not necessarily like code helpers. So I still use like copilot within Visual Studio. But this does everything now. Like, I don't, I don't know another way to put it to you guys, I really don't. I. And it's not even that I wish I do. It's just like I tell it to do stuff and it just goes and does it.

Can I ask you a question? Fire away, because here's where my brain goes. Because you once needed to know almost everything to do what you're trying to do.

Yeah.

Your humble estimation. How much does somebody now need to know to be able to do everything? Like, right, it used to be like, let's just give fake numbers. You needed to know 75% of the process of coding to be able to actually like do something powerful. Are we at 20%? Are we at 40%? Are we at 10%? Like, if you know these seven things, you could probably create your next business. Like where.

10% or less. Wow.

Yeah.

Like, Chris, like let that hang for a second.

And that's starting from like earlier this year. I had to know raw if I had to put a percentage on what I had to know. Now what I was doing earlier when I first started, it was just, hey, GPT, pump out code for me and I'll take the code, put it into an HTML file, take that into replit, and then I'll host that. And here's my app. That was like, I needed to know 90%.

Yeah.

And then when the auto coding started, that brought it down to around like 60 or 50 because I still needed to know a lot of terminology, still needed to know, like learn GitHub, learn all that stuff. Like it was, it was quite the pain in the tuchus. So yeah, like now it's by comparison it's 10% hands down or less. Probably like probably closer to 5%.

So watching this and then I'll shut up and y' all can do whatever for the rest of the show. The point I'm trying to make is the barrier to entry for your latest and greatest idea to become reality and become a business and put you in the race of potentially being one of the one founder, one employee, billion dollar businesses in the future continues to get easier.

All right, so George, like if you only need, if you don't need to know this and that's lower than 10%, what do you need to know?

When you ask me that, I feel like I need to know the humans that I'm serving better so that the tool that I can make with ease is triple 10x20x as powerful as the original idea. Idea that I would historically have to have spent a massive amount of time to launch. Now I can spend more time understanding the humans that I'm serving.

Yes. And SAS in general has spent more time building and trying to like learn and get to the 90% so that then you can build the tool so that then you can launch your idea and help the people. And because that's like number three or four, a lot of people do it without knowing or spending time with customers and understanding the problem inside and out, understanding the solution inside and out, and understanding the humans inside and out. So that's what I mean when I say things like AI is forcibly removing unnecessary complexity to where. When I see this on the screen, like this could look like, oh, Nico's just having fun, which he certainly is. But when you think about, I know so many of us have dreamt about things like this where it's like when, you know, how humans learn and how some don't and we have this huge HubSpot knowledge gap to fill. I remember 2019, I was thinking like, I had a friend that, you know, could illustrate comic books. And I was at a new business and I was trying to think about how like we could create like a comic book around the business with characters and like, because people learn from stories. Stories and like seeing themselves in the story. And it's just back then, you know, the jump from my idea, like, well, it never happened, like to tell you, like, so that tells you how far, you know, whether it's time or money or, you know, expertise, one of those things are not around, they're not available. Right. And that's where this democratization of the knowledge of how to digitally represent, you know, these things, like the gap is gone. Right?

Yeah, yeah.

And this is the kind of stuff, like if you can deliver this, there's a whole, you know, section of the market guaranteed that if they enjoy playing something like this, they will learn HubSpot at the same time in ways they never would have before.

Yeah, coach, put me in the game. CMS Rogue, please. Let's just go ahead and add me a spot. Number two. Like, ladies and gentlemen, get in the damn game. In the game.

I don't know what, what people are still waiting for, I gotta tell you. Like, I mean like, because of, because of this, it was like, okay, well, like the other day, I'm like, you know, there again, there's been stuff that I've wanted to develop for a long time now that I just haven't had the, like, haven't had the time primarily because even though it seems like super easy and like you can crank like now, I can. So even though it has in the past seemed like it was super easy and stuff like that, to be able to do this and be able to run like multiple things at the same time. Oh, you're not going to want to cut away yet. There's, there's been an, or not an issue, but there's just been a whole thing of like, okay, but where do I go from here? Like, yeah, you can make like silly little games and stuff like that. Like yesterday I've wanted to for the longest time be able to make something that can actually translate what I'm saying and do transcription and be able to actually put that into like wherever I'm at, application wise. And so yesterday I basically just asked Claude code like, hey, I've got a OpenAI key, if you don't mind. Throw together a whisper application for me that I can use and on like Windows desktop or whatever. So now I have this Windows, I have this whisper application that just works if my mic isn't, you know, engaged already. That just does the dictation for me. And because it's using the Whisper API key from OpenAI, it's as accurate as the OpenAI dictation which if you have used is amazingly accurate and will even catch slang and will even catch, you know, like all different types of, you know, whatever type, you know, words, sayings, you know, all of that kind of stuff. Slurring your speech, whatever it is. It's like the most perfect dictation app I've used. So it's like outside of that, I was like, yeah, I really hate HubSpot's, you know, like, card building stuff. So I was like, yeah, I should probably just create a card builder for myself so I don't have to do that stuff anymore. So I, I just did that and pretty soon it's going to have, I mean it's already got some plugins, but it'll have some more plugins as it goes along. Has like an AI wizard for you so you can go through and like determine like, hey, what do you want to build out as a. So you have like 317 properties available card to be able to show. I don't know, what do you want to see on companies? Usually you want to see like contact notes or something like notes from primary contact record. I don't. This is literally the first time I'm doing this, so I don't know if it's actually gonna work well. But yeah, I mean it's like even still. So. Okay, there we go. So like that's my card, like whatever. I will eventually have that be the case where it's like it's just on here and then you can just manipulate it on here. But for now it's very simple. This was like the first push out of Quad code. So this was the first ask was like, hey, build me a card builder for HubSpot that takes API key and my Open API key to be able to do like custom cards and you know, involve plugins so that I can. So that we can also plug in like additional applications and stuff like that. Shout out Ryan Ginsburg and you know, be able.

Yeah, I need to talk to you. Not on the podcast. It has nothing to do with HubSpot and I'm going only going to allude to the idea. I, I will split it 5050 with you whatever we make. But I have had an idea for an app for podcasters, and I want to talk to you about it to see if it is actually feasible to do the thing that has been in my mind for at least two years now. That's. That's I'm gonna say in public, but I'm just throwing it out that you and I need to talk. Oh, man.

It's gonna be a fun weekend, looks like. So we'll wrap there. We hope everybody has a fun weekend. I mean, how could you not at this point, access to all this power and intelligence and technology. Just start trying to have some fun, Try and get in there, start trying it out, and we will see everybody on Monday. Thanks so much, guys. Thank you.

See you guys.

Bye.