The Q1 Recap: Platform Wars, The Future of Commerce & Custom AI | GTMN

The first quarter of 2026 is officially on the books, and AI isn't just a tool anymore—it's the foundation of how we reach customers.

Welcome to a special recap edition of GTMN! We've curated the most essential segments from all of our recent chats to give you a high-level view of the massive shifts that took place in Q1. From the AI platform wars to the future of commerce, this is the ultimate catch-up on the strategic pivots every leader needs to understand right now.

Transcript

SPEAKER_03 0:00

All right, everyone. The first quarter of 2026 is officially in the books. And welcome to the special recap edition of GTMN.fm. Looking back through our episodes, it's clear that something big is happening in the go-to-market landscape. We've moved into a new era where AI isn't just a tool, it's the foundation for how we reach customers and measure success. And in this episode, I've curated the most essential segments from all of our recent chats with Austin to give you a high-level view of the major shifts that we are witnessing. We'll walk through the platform wars, the future of commerce, the strategic pivots that every leader needs to understand right now. It's a deep dive into the technical and creative forces defining this year. Grab a drink, settle in, and let's dive in. OpenAI launched their, like officially announced that they're launching ads. We covered that a couple of weeks ago. But then OpenAI and Stripe launched a Gentec Commerce Protocol. And then Google and Shopify launched a universal commerce protocol. Like so there's ACP, there's UCP, and then OpenAI also has a deal with Shopify where merchants are going to pay 4% of every transaction that happens through Chat GPT if it's a Shopify merchant. It's just bananas. Like all the everyone's trying to angle and figure out their preferred way of doing all this. But in the middle of all this, one big thing that happened is OpenAI actually released some details about their advertising platform. So the pricing is going to be a $60 CPM, which is basically about 3x what Meta charges. And so it's very premium and it's all based in CPMs. There's no conversion metrics, there's no conversion optimization. It's basically like you're paying us 60 bucks, you're getting 1,000 impressions on Chat GPT, and take it or leave it. And they're starting with e-com companies or, you know, large sort of consumer advertisers. Some people hate it. Some people are like, hey, they'll figure it out. It's just their first thing. Like, don't give them hate. Um, and some people are like, hey, like, what are you doing? Like, why aren't why didn't you build out an actual ad manager and have conversion tracking and do it the right way from the very beginning? So I'm divided. I'm like, hey, it's OpenAI, they'll figure it out. Like they've hired all the best people from Meta to do this. So I'd be surprised if they don't figure it out. Um, but I'm curious what what your take is on that whole like how OpenAI has launched ads and the details that are emerging.

SPEAKER_02 2:39

Well, I mean, I think you touched on a lot of the important elements, right? Like it's a brand new platform, they don't know what they're doing. Cost per mile is super easy. It's just like, you just pay us and hopefully it'll work. Like we don't know, they can't guarantee anything, right? So I would argue they might actually have better results and get more payment in the future if they can do something other than CPM, but they just think they can't yet. They're just starting somewhere. So I don't I don't think it's like some grand nefarious scheme. I think like these things are complicated, hard, and take time. And even if you're the world's biggest AI company, these things are hard, complicated, and take time. That said, hey, you you know the game, man. Like new channels are the best sources of growth. So I think if you're one of those select merchants who can advertise on there, you're pretty fucking excited.

SPEAKER_03 3:32

I couldn't agree more. I mean, it is gonna be a gold rush. Okay, here's my here's my prediction, and I don't typically do predictions, but I think they're easily going to make 10 or 20 billion in revenue off of ads just this year alone. I think they're gonna completely eliminate their burn rate with ads coming out of 2026. Like that's my bold prediction. All right, we'll see. Jan of 2027, did did this prediction turn out right?

SPEAKER_02 3:58

Because if even if they can't do anything more sophisticated, just assume that they can't do anything more sophisticated than CPM, just based on the number of users they have and the amount of impression volume that they have on their core product today, I totally agree.

SPEAKER_03 4:11

Okay, should we talk about our favorite topic? Uh are the Claude ads. Oh, yes, the one I sent you. Yes, let me pull this up. Should we should we play it for everybody?

SPEAKER_02 4:24

You gotta let people you gotta let people watch it because this is our continuing theme. Uh I feel like this is created out of an episode of Mad Men, though. It was so good.

SPEAKER_03 4:33

I mean, the Bruhaha, and we'll talk about it like right after we watch this together. And you know, we're giving some free marketing here to Alex McFarland. I don't know who that is, but how do I communicate better with my mom?

SPEAKER_00 4:53

Great question. Improved communication with your mom can bring you closer.

SPEAKER_03 4:58

Wait, I love how it took like less than a second for uh the lady to wake up and respond. That was hilarious.

SPEAKER_00 5:09

Start by listening. Really hear what she's trying to say underneath her words. Build conversation from points of agreement. Find a connection through shared activity. Perhaps a nature walk. Or if the relationship can't be fixed, find emotional connection with other older women on Golden Encounters. The mature dating site that connects. Are we gonna get flags for this? What would you like me to create your profile? So good.

SPEAKER_03 5:45

Oh my god.

SPEAKER_04 5:46

What's the difference between me and you? Me and you, me and you.

SPEAKER_03 5:51

So good. So good piece of an ad, right? Just so good. And you know, it's interesting. Okay, well, what do you think about the ad? You sent it to me, so let's let's start with you.

SPEAKER_02 6:02

First, it just it fully captures one, the the depth of emotion that people are going to to talk with AI. Two, I think it captures what I'm hearing and sensing from a lot of people, which is like GPT is pretty lazy, like is providing lazy responses these days. I don't know if you caught it, but it was like the answers were so generic and fluffy. And yes, he was happy to have the answer, but they were pretty bad answers, objectively. And then, you know, the ad at the end was just the cherry on top. The fact that it was like not at all related to the topic, was only loosely related to the topic, which now, you know, look, will ads be that bad? Probably not, but it captures the point, which is like you're sharing private information, which could be utilized to advertise for something that's completely unrelated. And if we've seen how other advertising has gone, that's certainly what happens in the world of programmatic. I can't tell you how many times, like, I've bought a product and then I receive an ad for it still on YouTube because they never did re-engagement ads properly and they didn't have, you know, an opt-out list. So these are the types of things that can happen um pretty easily. So I I thought it was a brilliant, brilliant advertisement, especially given the fact that I think Claude is like winning a little bit of the war right now, at least in consumer perception for the Silicon Valleyites. But, you know, what it doesn't address, which is kind of side to all this, is that there's still a vast majority of people who are not using AI in corporate America. So like this is a great audience, great ad for our audience, us, but won't really make any difference for somebody in Wisconsin who hasn't figured out how to use AI yet.

SPEAKER_03 7:40

I mean, what a great point. And that's what I wanted to bring up here. So if you look at Google Trends, and hey, don't don't get me wrong, like the fact that we're looking at Google Trends to understand the growth of ChatGPT versus Claude is like the whole thing is just so freaking meta.

SPEAKER_02 7:56

It's so hilarious.

SPEAKER_03 7:58

Um but when you look at sort of the you know the the market share and the interest share, um, you know, other than this like holiday period sort of lull, uh Chat GPT is still crushing. And you know, the the and and that is the same point that another one of my um feed people were making. Feed people, uh somebody I followed.

SPEAKER_02 8:25

That's a great new term. Feed people.

SPEAKER_03 8:28

I don't know what to call them, right? Like the people I follow, this is too long.

SPEAKER_02 8:32

Random random N NPCs on LinkedIn.

SPEAKER_03 8:35

Um and they were saying the same thing that you know the awareness for Claude in America is probably less than 3%. And I don't know how they came up with that number. There must be some study of that or what have you. And so the thing that I think the Claude ad did really well, and I think this is going to play at the Super Bowl, is they did not mention Chat GPT by name. Uh-huh. Right? Unlike the other sort of competitive ads that you see, where even the Pepsi and Coca-Cola ad, like Pepsi has literally got Coca-Cola in the ad. So you're actually giving airtime to Coke just as much as you're getting attention to yourself. So, from a just a pure advertising perspective, one thing that they did well. The thing that is interesting is like the absence of ads may not be enough of a needle mover in terms of conversions for Claude. You see what I'm saying? Like this might be a nice little story to tell, but it actually may not inspire more people to go check out Claude because you didn't say anything about Claude. You just said it doesn't have ads. But what does it have? Right? For an average uh person.

SPEAKER_02 9:42

How is it better?

SPEAKER_03 9:44

Like the absence of it doesn't make it better by you know, if you if you don't talk about for an average person, not the Silicon Valley Elite. So I think it was a hilarious ad for us, for the insiders, but I don't know that it's gonna connect that deeply with the broader audience.

SPEAKER_02 9:57

Especially since the broader audience might not understand yet how to use GPT or Claude in order to do work and help their lives, right? I do think it lands in the sense of, hey, ads are bad if you're talking about private sensitive health information, private, you know, intimate information about your family life. You don't really want to be advertised for that. So it lands that that point quite well, but it's a making an assumption that the audience already knows how to use these LLMs.

SPEAKER_03 10:21

Google also launched a bunch of new stuff. So their new Nata Banana stuff has launched. I don't know if you have played around with it, but it is darn good. They launched it with this new feature called a window seat. So, what you can do now is say I want a view of the Golden Gate Bridge in SF. And I can say Golden Gate Bridge, and then I can give it a perspective from Bay Bridge. I don't actually know if you can see the Golden Gate Bridge from the Bay Bridge. I just like I'm thinking. But and so what it will do is it'll create the perspective from the Bay Bridge looking at Golden Gate Bridge. And so you can do this for like all types of um images where you're not just giving it the object, but like the framing of it, and it does like some pretty cool things. And then they launched yeah, and they launched something similar for their SMB product called Pomelli. I don't know if you've seen Pomelli that's doing something similar, but like allowing you to create a photo shoot for your product. So you give it a product and you give it some prompts for creating a photo shoot, and essentially it allows you to create a professional grid, you know, photo shoot using Gemini with that product in different settings and what have you. And for the first time at a dinner two weeks ago, somebody said that they've moved all of their static to Pomelli. This is like a multi-hundred million dollar brand. And I was like, really? Like Google Pomelli? And I'm like, yeah, they're just using Pomelli all day long. So Nano Banana is is epic. And I had this experience awesome. I was creating an image, I used uh nano banana in Gemini, and then I went to Figma and I tried to create the same image in Figma with Nano Banana, it didn't work in Figma, but it worked straight like single shot in in in in Gemini. In Nano. Yeah. It's kind of wild. And so I don't know what's happening in Figma, but there's there's something concerning here about you know their business and maybe even Adobe's business.

SPEAKER_02 12:21

Well, I think like, I mean, I can only speak from the perspective of how I'm using these tools. I think when I'm trying to come up with interesting images and iconography and creative style artwork, I actually find that that starting with a human is actually the best spot. So, like my my workflow is if I'm trying to design something, I will hire a very, very inexpensive offshore creative to come up with 20 ideas for something. And most of the time they're like really interesting in it because I I'm not always the most creative person in the world. But say you say you need to design like a new logo or a new brand for something, put all your thoughts on paper, have somebody come up with 20 ideas, then you can take that. And I think like with nano, you can get some really incredible outputs because I think nano is really great at taking information and even existing sketchwork and making things that are beautiful. I think it it can one-shot a lot of stuff, especially if it's simple. But if you're looking for complex artwork and outcomes, and then actually where I feel like Figma is thriving right now is in using all of its tools with MCP to take like the interactive brainstorming and kind of troubleshooting work of creative design and making it single shot. Like two examples that come to mind is one, you know, I need to like build a screen for this app I'm building. So I took a bunch of screenshots of similar apps that I care about and I had agents go and catalog their CSS for me. And then I basically was like designed me a couple of screens for this app in Figma that I can walk and talk through with a team. And it was able to one-shot that really quickly. And actually now the Figma MCP supports layers. So, like, we'll create, I don't know, HTML and CSS style outcomes with layers. And so that's one use case. The other use case is I always use Figma for brainstorming. I've predominantly use Fig Jam, but I'm actually switching over to Figma now because of the Figma MCP. And the example I'd like to share is when I previously run OKR sessions with teams, I will sit down, I'll have a brainstorming session, we'll have, we'll do crazy eights or something. You know, it'll be 10 of us, we'll put our ideas on a board. Then I have to manually collate everything. I'll go walk through that board with every individual and gather feedback and then incorporate it into the OKRs. Like OKR is a really expensive, time-consuming process, but it's important for teams to get aligned. And that could live in Figma or a Google Doc. But this year, I actually just did it exclusively with Clarify and Figma. So Clarify is my call recorder. I went and interviewed six or seven people on the team, asked them questions about our company OKRs. Then I had a brainstorming session afterwards where all I did was like take the themes from the calls and put them on a board. We all talked, we all put our feedback down. But then I used the Figma MCP to like pull that into themes and groups. I spent, you know, an hour working through those groups one by one with Claud Code, distilling information. I also went when there was something missing, I could be like, hey, could you go to that call with Pranav? He said something important about this OKR. I'm willing to bring it in. So it was a little time consuming sitting there with an agent, kind of working through it. But the outcome was now we have these OKRs that are literally the words of the humans I'm working with. So it's not, you know, there can't be any speculation that it was, you know, made up or that it was hallucinated because it literally was just consolidating six conversations worth of text and a giant board into things that made sense. And it was audited by me as a human. And like that to me, I actually believe is like one of the most powerful aspects of AI now is not necessarily as a creative tool, but as a consolidation and synthesis tool. And it's actually very, very good at synthesizing lots of text, especially if you have it in the form of transcripts.

SPEAKER_03 15:59

Lots going on in 2026. Um, Austin, I want to kick off with this whole my feed has been full of, and and it's I think it's also because it's CES week. Everyone is at CES literally right now, but it's all about agentic commerce. Agentic commerce. And one of the headlines from the last couple of weeks was Do you know what IAB is? Have you heard of this acronym IAB? I think it's the Internet Advertising Board, something, something it's like a nonprofit group. Yeah, yeah. So it's a it's a it's a digital advertising technical standard setting body. Talk about jargon. Uh, but they did good work. They did good work. And they essentially announced a an agentic roadmap, which is basically here are sort of reference architecture recommendations for everybody in the ad tech and the martech industry on how to go towards agentic commerce. And I was looking at it, it was like super technical. I went through the doc like for 20, 30 minutes. It's on their website, IAB Tech, I think. And um, there's so much momentum for AI and um agenc stuff. And at the same time, over the holidays, I also listened to a podcast by Rory Sutherland. Have you heard of that name? Have you come across uh okay? So Rory is, I think, um, sort of a uh a leader at Ogilvi, if I'm not wrong. And he's also the author behind this book, Alchemy. You should read that book, by the way. It's fantastic. And he was on the Knowledge Project podcast. And I'm gonna play this for you because I think it's pretty, it's pretty amazing what he said about agents and AI, and then we can just have a discussion about this. So this is uh that that podcast, if you haven't, if you're watching on video, you should go check out the Knowledge Project podcast. And this is specifically Rory's episode. And I'm just gonna hit play here and let's let's talk about it after.

SPEAKER_01 17:59

The whole question of the decision science is uh is I think at the moment completely critical. And actually, by the way, it'll be very, very interesting with AI. Because I keep hearing people saying you will say to your uh AI, find me a staying holiday, and it will provide you with a perfect paying holiday. But I keep saying people don't design like that. At the very least, you'll have to show them three or four or five paying holidays from which they choose. Because we can't really choose in the absence of comparison. You see what I mean? It's a freaking element of free market capitalism, which is that we can only like something if we choose it in preference to something else. That's interesting.

SPEAKER_03 18:42

All right. So I thought that was actually just a very short snippet, but that point about single-shot AI commerce or agent e-commerce, where you say, like he said, I want to sking hauled and it just books the whole thing for you, versus it gives you a menu of options. Hey, guess what? That's what the current paradigm does for you, and it's just a better way of coming up with those uh with those choices. What's your take on all of this? And and do you think that agent e-commerce is really going to take off the way people describe like you'll just have agents buying everything for you?

SPEAKER_02 19:14

I don't know. I'm not convinced because I think a big part of the buying experience is the actual discovery of items and products.

SPEAKER_03 19:22

Okay. The the major news of this week was um the first ever ad placement on an AI surface appearing through Google direct offers. Have you read much about this? Do you know what this is? No. Okay. So what Google is saying is they've got this AI mode, right? So I think there's AI mode and AI overviews. So essentially what they're saying is whenever Yeah, yeah, yeah, exactly. So whenever you are I in in any of the AI surfaces that Google has, and there is a very high intent query from a consumer. Um, and if that's happening in those two modes, AI overviews or AI mode, advertisers can now place what is called a direct offer. A direct offer, it's very constrained for the time being. It is only, I think, for e-com uh type of merchants. And it is literally a price offer. So you're doing a 5% off or 10% off or, you know, buy two for one or whatever else. So these are price offers, they're starting with price offers, and it's only going to be targeted towards high intent queries. Now, how does Google know that it's a high intent query? We don't know. They're doing their own determination of saying, is it a high intent query? Is there intent purchase? But I have to assume that Google's pretty good at figuring that shit out. And so they're going to match up. Uh, so you have to opt in as an advertiser, you have to create these offers, and then they will show up programmatically in these surfaces. And the rationale was really interesting. What Google is saying is like, hey, if you're in research mode, right, when you're researching a lot of different things and whether it's search or AI, that's not where these things are gonna show up. These things are gonna show up when you're very high intent, you're ready to make a purchase, and here's an offer right then and there to capture that intent and get you that customer. So it has begun. Ads are now in AI. Um, and you know, we're gonna start seeing some crazy stuff happen. I am excited, but also kind of intrigued that they chose pricing promotions as the the first version of this. It can be scary because basically you're just discounting. Maybe those people would have come to you anyway. So, you know, it's an incrementality nightmare, but I will tell you everybody in the marketing world is going to be trying these out over the next like three months. So fun times.

SPEAKER_02 21:50

And you should, right? Because I mean, new channels often are where you find the most growth.

SPEAKER_03 21:54

Okay, so I want to share one more thing about Amazon, which is really interesting. So, do you know Neil Patel? Is that name familiar to you? So Neil is interesting. He was one of the early like bloggers and online influencers for all things marketing. So he's had a you know blog for like I think 20 years talking about all things conversion rate optimization and online marketing and what have you. And I follow him on X and he brought this up that Amazon just pulled off the biggest PowerShift in advertising. And I don't think most people have realized that now, if you go to Amazon's DSP, you can buy Netflix, Spotify, Roku. All of that advertising inventory is available through Amazon's DSP. So not only are they, you know, doing commerce and doing, as we were saying, right? Like the actual sort of infrastructure, but now advertising not just on Amazon, but everywhere else. And using Amazon's um, using Amazon's first party like conversion data to inform the audience definition on those third-party inventory. So this is a pretty massive shift. Like you you you you go to YouTube for advertising, now you go to Amazon for advertising, not just for Amazon, but like all the third part, like all the open web, you know, inventory.

SPEAKER_02 23:10

I'm not I mean, again, I I'll have to I'll find the article from like 2019 that basically foresaw this coming. But you know, there's a lot of power in having a platform.

SPEAKER_03 23:20

Yep, and they're just killing it. This morning, or I I don't know when I noticed it, but on my X feed and on my LinkedIn feed, all I see is Manus. Have you heard about Manus and you know they got bought by Meta, like and and what's happening now? Okay, so I know very little.

SPEAKER_02 23:36

I just know that it happened, but I haven't done the research yet. So tell me.

SPEAKER_03 23:40

So Meta bought Manus AI for something like $2 billion or something like that. I don't know what the exact number is. And and Mana's just launched, or I think it was yesterday or the day before, but they came out and said that we have built a Meta Ads Manager integration into Manus. And if you know anything about LinkedIn and X, there's a whole bunch of like e-commerce media buyers who are just, you know, obsessed with automating everything. And I get why, and they're just going nuts about how amazing this thing is. And, you know, so so that was an interesting sort of evolution. And what what I'm seeing here is Meta going all in. And and Mark Zuckerberg even said that by the end of 2026, I think he said 2026, he envisions automating all media buying on Meta. Now, that might be an extreme point of view, and you know, maybe what he means by automate is okay, you have a MANIS agent that's doing it for you, but just a you know, pretty big piece of news if you're a media buyer, if you're a meta media buyer, which is obviously sort of a significant portion of all media buying that happens on the internet. So I haven't played around with it, but big moves, and I expect I expect like you know lots of people to have points of views in the next week or so. So I don't know what you think about it. Yeah.

SPEAKER_02 24:59

I yeah, I I want to figure out how it works. A conversation that came up very often in my trip out at West was the idea of all marketing agency actions being automated. You know, we're seeing this happen. It's just happening in different parts of the ecosystem. But the one that people most know about is AEO, SEO, and content, right? Like there are a lot of agencies now that will help you write content, you know, print off. And as a founder CEO, you can have a content strategy for LinkedIn and X and Reddit. And that used to be somebody's job. It used to be sometimes even an agency's job. You'd hire a content writer or content agency to write on your behalf, and that would kind of like fill the hole in the content parts of your of your marketing team. And then we saw it happen with SEO and AEO. I'm saying SEO and AEO. People are going to debate that and say it's separate. I think AEO is becoming SEO and vice versa. We'll see that happen over the next three years. We talked about that in a recent episode. So that's another piece of the marketing team that's being outsourced. So this kind of is now entering performance marketing, right? What if uh, you know, in the beginning, maybe you can do programmatic and you can do media buying with the help of an agent, but then what's to stop you from spinning up an actual agent to do all of it for you? Um and so I think this is just gonna continue. We're gonna see more parts of the marketing team split into agents in the same way they're seeing marts for parts of the software engineering team split into agents.

SPEAKER_03 26:18

Yeah. And, you know, a counter opinion on this, which was I follow this guy, David Herman. He is awesome. Um he shares a lot of interesting stuff. He's kind of like an unknown quantity, but a known quantity. So very under the radar. Like he's not big about shilling his stuff, but he's been doing media buying for some of the best sort of consumer brands for the last like 10, 10 years. Wow. And um, what he said was really interesting. So I'm just gonna put this up on the screen and you can just kind of read this on your own. Um, I think he, you know, he tweeted about it yesterday or or today, or do we call it X'ed about it? Oh my God. Um, but to to for everybody who's listening, the difference in in my opinion on having Manus or any AI run your media buying always comes down to the gut reactions versus the data reactions. And so what he's saying is, hey, you know, lots of big agencies do need to fear automation and AI, and those who deeply get the thing don't. And so what he's essentially saying is, you know, all of these AI tools and manus and whatever are going to keep getting better, keep sort of you know improving uh the automation and the efficiency. But the the ones who know when to pivot uh from a particular growth tactic to you know a different tactic, those folks are not going to stress because if AI and agents are doing all of the automation, what is left? Where's the arbitrage? And the arbitrage is coming from that human sort of loop, human intelligence and intuition. And then he does go on to say that, hey, maybe the the day that AI learns all of this and everybody loses their jobs, and we all end up looking like, you know, the fact blobs and WALL-E, you know, maybe maybe that day comes, but that's a different sort of problem altogether and different sort of you know uh reality that we would live in. So I thought this was a very balanced take. And this is coming from somebody who's done media buying for 10 years and a lot of fit on meta. Um, so it's it's interesting to see how this is going to evolve in the next like six to nine months, but it's just a crazy time to be alive.

SPEAKER_02 28:23

Yeah, I totally agree. And and actually his point is valid and hilarious. And I think like that that's what's kind of creeping into the um the intellectual conversation behind all this. I mean, Elon talks about this, Sam talks about the same altman. Um, like if AI becomes so advanced that it eliminates the need for jobs, then can you create a world of prosperity where people can actually live? But I to be honest, like in a democracy like the United States that's driven by capitalism, I don't see that ever really happening. I use this term all the time called deterministic attribution, which is to say, hey, I know exactly who clicked the ad. I know their phone size, their screen size, their their operating ID, I know their IDFA. Question for you, Pranov. Is that type of deterministic attribution dead, alive, or on life support?

SPEAKER_03 29:10

Uh good question. Well, I I don't think it is dead. I do think it is misunderstood and maybe misused, but it's not dead. And the the first thing to sort of clarify is, you know, this is this I I hate to have a semantics conversation, but I think it's important just to kind of set the stage very quickly. When you use the word attribution, what do you really mean? Yeah. If you mean cause and effect, then I don't think that the thing that we call determinat deterministic attribution does cause and effect. So that's the the irony of uh the the label. Now, what it does do that is extremely valuable, and I think you need to have it, is telemetry or tracking or whatever you want to call it, right? But like understanding the knowable parts of the journey that a certain user is going on with your mobile app, it does it. And, you know, pick your solution, whether it's App Slide or Branch or whatever else. Um, some will be slightly better at that or not. But it's not debt. It's it's it's got a very specific use case. It is overtrusted by people to even the best of marketers might think that it implies causation or it implies that you can use it to do budget allocation. But those assumptions are extremely flawed. Let me pause and see if that helped.

SPEAKER_02 30:38

Yes, it helps because I think that what people are trying to do is they're trying to understand how them spending money is making them money. They want to know where to spend money. They want to optimize, you know, their campaigns against existing acquired customers. And that's, I think, what people fundamentally are trying to achieve when they say attribution. And that's a big part of the problem is there's like there's the story, there's, hey, what am I trying to do? Then there's the kind of statistical analysis of what actually is the answer to that question. And then there is data. And data can be informative in answering the question, can be informative in establishing causation, but capturing data is a separate and discrete practice from attributing campaigns from answering the question. The problem, though, is that executives ask the question and then operators have to go figure all this shit out on their own.

SPEAKER_03 31:35

Another thing that's happening, there was a massive amount of panic on X, and also just like talking to a bunch of my uh peers about the SaaS sell-off, right? Figma's down some like 70% from its highs. Um, right, people were just talking about like all of the major SaaS companies are down. Like there's a major sell-off in SaaS, and everyone's freaking out because of things like Claude and Claude Covert and you know the plugins thing, Claude bot, mold bot, all this stuff. And I'm like, yeah, but like, are you really gonna rip out you know your your SaaS tools with with AI build tools? I I it just feels so misguided to me that there would be such a big sell-off. And I I think it'll come back, but who knows, right? It it's uh what what's your take? Like, have you have you what what's the discussion with your customer base around this stuff?

SPEAKER_02 32:27

Man, I mean, I think it's real. Like, you know, I talked to you about this over the weekend. I'm building a um kind of like a centralized brain. I called it like a personal data platform, and it's written in Cloud Code, and the purpose of it is to collect information and context from a variety of sources and provide a context layer on which I'll build applications that do stuff. Because I my personal limitation so far is like I have lots of directories that I've built in either Cloud Code or Cursor that do things, but there's no thing to tie it all together. But every time I take an action now, I spin up a directory and I apply Claud code and I have it go figure out a solution and do it. There's almost no SaaS application that you can't build with enough time. So the question is like, is the value that you're paying worth the value received from not having to go build it with an agent? And I think that's the case for a lot of um there's focusing especially on like legal and human resources tools, right? Like the example I came to in our space often though is um, you know, those privacy consent banners? Like, man, I would not want to be in that business. Because what I looked at these maybe six months ago, and they're basically like a JavaScript snippet with a back-end system that collects data in a warehouse and they're charging $10,000 or $15,000 for it. So if I ever were consulting for somebody who had to do GDPR, I'd be like, just build it. There's no point in paying one of these providers to do it. Oh yeah, maybe it's $15,000. Well, you're gonna set up, you know, a snippet on your site, which Claude can do for you. You're gonna set up a warehouse to collect opt-out data and federate GDPR and sure there are compliance rationals. But most of the time, an agent will be able to figure those out and you can set up some type of, you know, um, you know, cron job to just alert you every single day or week or whenever you need to. So I think that's what people are scared of is like there are lots of these niche SaaS applications that are great businesses because it's like a silly, stupid thing to have built, but now it's silly stupid because you can build it. Another example is have you heard of like uh dams, like data asset managers? Yep. So like a dam is just a giant ass database with assets. Yeah. Do you really need to pay $100,000 for that? Or should you just go spin up a snowflake warehouse and build some type of UI where people can log in and store their credentials or their um I've heard people doing it with Google Drive.

SPEAKER_03 34:48

They use uh you know, Claude to sit on top of Google Drive and that's their dam. Mm-hmm.

SPEAKER_02 34:54

Yeah. So this is where I think people are freaking out a bit is like there's a lot of these easy applications that have been built, even like in the last year and a half, on top of AI. And you're like, well, why would I pay $50 to $100 a month for something when I can just build it once and maintain it forever?

SPEAKER_03 35:10

If you don't invest in your own education, you're gonna get left behind because you won't know what questions to ask AI. You're gonna be stuck at simple questions like summarize this thing for me, right? And so I'm like getting extremely bullish on ed tech. And the reason is, yeah, like there's a there's a part of me which is like, okay, you can just go learn through these tools as well. You can, but again, it's the problem of like I don't know what to learn. Like somebody has to structure the heck out of it. You you see what I'm saying? So there is a reckoning coming for and and hopefully a boom coming for ed tech, which every parent should be paying 50, 100, 500 bucks a month for all the emerging ed tech platforms to get ahead of this and and get your kids like that type of education. I don't know. I just feel like that's gonna be a massive opportunity in the next like five years.

SPEAKER_02 36:01

Dude, you hit it on the head so much. This is the reason why I'm also not worried about at least my own career. Like I know a lot of people are freaking out about white collar work, but if you are a specialist with domain expertise in an area, it doesn't matter if an LLM can do all of what you do. What matters is the ability to interpret all of what you do and to explain it to humans in a way and validate that what has been done is correct. You know, for example, with this um, you know, this this data product uh genie, if you don't know what questions to ask and you have no way to validate it, you're effectively giving up complete control of your company's strategy and the ability to have OKRs and goals to effectively AI. Which, you know, either means the founder doesn't need you or you know, you're gonna be replaced. So yeah, I I just I don't I don't really buy the idea that that these things are gonna displace people. I think it's gonna just displace generalists who don't have specialist skills. But this is why, you know, if you're an existing specialist and now you have a tool set that can help do your job 10x faster and you have the ability to interpret it, well now you're able to just maximize your time, right? Yep. And you know, it's so funny on the education piece, a similar debate has been raging around like, is upper education even useful anymore? Or like why get an MBA? Why get special degrees? And I'm like, at the end of the day, what a lot of these programs do is teach you how to learn. That is the ultimate goal of school, in my opinion, is to give you a great network of people and community that you have for the rest of your life and to teach you that the process of learning is never over. A lot of people go and get degrees whether they're from great schools or terrible schools, and they think that the work is done. That is just like that's like a poor behavior to learn. And I think in in the age of AI, whether you have the tools structuring stuff for you or not, you still gotta learn. Um one thing though on ed tech, real quick, like did we talk about this last week? I think Mark Andreesen talked about this in his in a podcast he did recently, where what's cool about this upcoming generation of AI is that you can imagine every student in the world could have their own private tutor to teach them. And I actually think like what's crazy to think about is you go to school and maybe you have teachers that are helping regulate the classroom and regulate what's being done and learned, but everybody has their own private tutor teaching them not only the required course material, but more. So you say, Well, look, hey, AI agent one, two, three for Austin, like you gotta teach Austin basic arithmetic and grammar and all this other stuff. And there's a set schedule we teach him. But hey, if he's curious about these other things, here are the bounds in which you can teach him. And you should start to like learn his personality and allow him to explore these other topics in a way that he feels comfortable with.

SPEAKER_03 38:42

I I'm gonna give you a $10 billion idea. All right, like I genuinely believe this. And if somebody's building this, I want to invest, even though I have no money to invest because I'm a poor founder. I think that you can build a $500 device, this is a custom Android tablet and a $500 or maybe even a thousand bucks a year subscription for a kid of any age. Okay. And this is a fully AI enabled tutor on your iPad. It has no entertainment, right? And it's configured by the parent and the teacher. So their school teacher. And it's essentially a progressively, you know, the same thing you do in lifting, right? You progressively load, uh, increase the load. It's just progressively overloading based on the interests of that particular child. And that device and that subscription, I I can't imagine any parent not paying for it. Like it's a slam dunk idea in my head. And the reason it you can't do this on top of an iPad, it can't be an app, it can't be on an you know, uh just the Play Store. And the reason is the moment you go onto a regular device, all the other distractions of social media and apps kill the value of this thing. So I think you can design an amazing product, and I know this has nothing to do with GTMN, but hey, here's a free idea.