E63: The AI War Machine with Rabah Rahil

Speaker A: Get me wrong, there was things to be stressed out about. Um, and honestly, there was kind of this weird once. Cause there was a point in time where I was transitioning into product, and so I wasn't in marketing. And that's when I could really see his magic of, like, being able to. Here's what happens from my perspective. You have as a CEO, like, you're everybody's parent, essentially, and, like, marketing product and I don't have kids. You guys have kids. But it feels like every going to sell this rich. The product isn't perfect. And he would just have this kind of like, parental smile of like, yes, I'll. I'll let you b. Tch. And like, okay, we'll go find your toy. Hold on. But like, in his head, he realized, like, it's just not that deep. So anyways, I think there's something to that, Jeremiah, about having, uh. In specific, if you're going to lead a company, I think that is a. I think you do want your org leaders to candidly be a little neurotic. But as the CEO, uh, you can't take on that, that, that. That neuroticism because it'll eat you up, dude. Because you running the marketing org, I had so much drama. Richard had every single org rolling out. Like, he has everybody else's drama. You had everybody else's drama. And I could barely live with marketing drama. And I had a small team, too.

Speaker B: So anyways, it is true, like, obviously, I don't see all of your life, but you are very even from what I see. Just like, constant even. Like.

Speaker C: Can I tell you guys something funny? It's the characteristic my wife hates about me the most.

Speaker B: That actually makes sense.

Speaker C: She's like. She's like, I can't tell what's going on. Like, uh, you know what I mean? She's like, you don't talk to me. I'm like, I don't know how to say this, but actually, I tell you everything. It's just that I'm not saying it multiple times and I'm not crying about it, you know, like. But actually, you know, everything. All the information has been transferred.

Speaker B: It's like, very factual delivery. Not, uh, emotional.

Speaker C: But you've. You've met Isha, right?

Speaker A: Yeah. Yeah.

Speaker C: Oh, yeah. You can see her. You can see her saying that, right? You can see that.

Speaker A: Of course. Yeah. She's a very alpha woman.

Speaker B: She was.

Speaker A: She was great. But you threw me under the bus. This very strong woman. And then Rish was like, I think you forgot to send her Like, I was always very adamant of, like, here's merch links for, for leadership. Like, buy whatever you want. Give them to your partners. Whatever do they. And then the first time, he's like, this is the guy that didn't give you the. There was something that we made, I think fmrt or something. And this is the guy that didn't give you.

Speaker D: To you.

Speaker A: Oh, by. This is our cmo, Rabba.

Speaker C: Oh, my gosh.

Speaker A: Yeah. Wonderful woman. Very, very, very strong. Very beautiful woman. But, uh, yeah, it was funny. It was, I think, the San Francisco event or something, but yes, exactly. You look great, by the way. Your hair, the beard. You look healthy. You look really healthy.

Speaker C: Dude, I just got. Did I tell you guys about getting a DEXA scan? Have I shared this story I love? No, I have never gotten a DEXA scan before.

Speaker B: No, I need to.

Speaker C: Jack, don't get one. Um, I don't want to hear your results. I can't. This is, I'm not, I don't have the mental capacity.

Speaker D: Dexa scan. Like, 45% ish.

Speaker B: Right.

Speaker C: Have you ever seen what Jack looks like? Rubba.

Speaker A: Jacked and tan. Well, not the tan, but the Jack part.

Speaker C: Yeah, dude, Jack is jacked.

Speaker D: Yeah. I'm a little less tan than you guys. Yeah.

Speaker C: Yeah, a little less tan. But, um, anyway, I, I, I've been working out for, like, three years now, and so I was like, oh, I should go get a DEXA scan. Like, I've been lifting three times a week. It'll be useful to know and to, like, see where I am on the curve. I have 28% body fat. And it turns out that the reason m for this is my lean muscle mass is, uh, Sorry, my lean mass is very low, even though I'm getting stronger. And it's because as a vegetarian, my protein intake is very bad. And so, literally, I'm, like, interacting with ChatGPT about my DEXA scan. And it's like, actually, man, like, everything is fine, but your lean mass is simply not getting added and everything. And I communicate with ChatGPT a lot about my health, so it has lots of context. Um, and so it's like, you've gotta increase that protein intake. Like, you've really got to, but, like,

Speaker B: a lot more legumes and stuff. Is that the planner or what? What do you do in that context?

Speaker C: Every morning and night, I have a cup of Greek yogurt with a. With added supplements of protein. So every morning and night, I'm consuming 45 grams, and then, so that's 90 total. And then I'm trying to use real food for the other 80.

Speaker A: Yeah, that's tough for you. That's tough. There's a lot, a lot of menstruation.

Speaker D: Cockroach cheese could be a good thing. Yeah, there's a, there's a, there's a bunch of different stuff. Did I get to tell you guys? Uh, I probably haven't told you this story, but back when I was doing Muay Thai, there was a guy at my gym. Uh, I call him Emilio. I don't know what his name actually was, but he'd always come in in the most inappropriate outf.

Speaker B: For.

Speaker D: For training. Um, and then one time he came over to me and he was like, ah, nice. Ah, body fat percentage. And I was like, oh, thank you. And he was like, what is your body fat percentage? And I was like, I don't know. You just told me it's. It's good. So I just said to him, I don't know. And he was like, oh, okay. Well, yeah, nice body fat percentage. And they just walked away. And I was like, oh, that was so weird. And he's like shirtless and jeans with work boots on. I'm like, what's going on here?

Speaker A: You know, it's amazing. That's amazing.

Speaker D: I think he was hitting on me. I'm not sure.

Speaker A: Yeah, definitely hit it on you. Definitely hit on you.

Speaker B: He was your deck. Our DECA scan.

Speaker D: Yeah, I haven't had one. No, no, I have no idea. But I imagine he bought them.

Speaker B: He gave you a scan.

Speaker A: Here's the cool thing though. When you start putting on muscle rish and I gotta get you. I built a whole app where you can drop in your Dexa scans, your blood work, everything, and it does everything for you, like and it tracks all your programming stuff. This is cuz when I, uh, left Vermont, got really into trying to get my health. I'd been healthy my whole life and then the last 5 years had. Anyways. But the cool thing is when start to put on muscle. Um, because the other thing. Are you allowed to do Peptides as a vegetarian? You are, right?

Speaker C: Let's say yes.

Speaker A: Okay. I would highly recommend, um, looking at. I got a plug too. The guy's great. Real stuff. Um, but you need to get your protein. Your protein needs to be there. Peptides aren't going to fix that, but it's more of like a turbocharger. But I was able to even though I gained weight, so I had a huge weight loss. I did awesome. I got that. But then I've Been because I don't want to be like small. I like actually feeling a little bit, uh, I don't mind being a little chunky.

Speaker C: Oh.

Speaker A: I'm trying to cut anyways. But the too long didn't read is you're amazing at the maths. I put on five pounds of muscle in my last bulk and only half a pound of fat. So even though my weight went up, my body, uh, fat percentage went down. Because it's a ratio.

Speaker D: Hell yeah, dude. And how long did it take you to build this app?

Speaker A: Um, maybe a month or something like that. It's actually really cool. It's uh, it's what I use. It gives you a little, um, uh, like motivational story from like world leaders or like Napoleon's and stuff like that. You track your, um, all your. It pulls in all your Apple health. So it has like my. My prep score. So it has my hrv, my sleep, my resting heart rate, my wrist temperature, um, my sleep quality. Um, so pretty much it's all automated around things. The things you have to drop in are like the DEXA scan, my blood work, and then all my nutrition is synced from my fitness pal, so it knows if I'm hitting my numbers or not. And then it'll have like a little weekly review and then you can talk to Claude through it and stuff. Yeah, uh, it's been really awesome. It was one of those kind of side quest things. Cause I just wanted to get better at cloud code and then I just started going deeper and deeper and I was like, oh my gosh, you can deploy iOS apps.

Speaker D: Oh yeah, dude. Did you build that app with Zach? Because I feel like he was building something around Michelangelo. And by the way, did you say Napoleon? Like, ah, are we thinking of the same Napoleon?

Speaker A: Yeah, he was, uh, he was a G. I mean, he's also a great example of uncapped ambition. We'll eat you up. I think him and probably Alexander the Great. But I think there's a lot to some of the things that us, uh, three wise men can impart wisdom on you Jack, on your young gun. But as you get older, one of the best skills you can develop is being able to separate great signal without taking the whole person. Because some people are really great at doing things that are just pure assholes that are like, I never want to be that person. But man, they are very good at X, Y or Z. Let me just internalize that. But not, you know, the way they show up in personal relationships or intimate relationships or um, things like that. And so there's very few great, uh, markers.

Speaker D: A really smart Napoleon, but I imagine there's something to admire there. On the topic of building apps and this app that took you a month to build, what do you think about this? I mean, I can't think of a better person to talk about selling SaaS with thanks than you. Rather, in the age of AI, I'd rather open a restaurant than sell sass. What do you think, dude?

Speaker A: I think if you got a bunch, uh, I think one, when you have things that you're starting, you should always start with a goal in mind. Like, what's the goal if you got a bunch of money and you want to have a cool place and you want the best vegetarian restaurant in San Francisco that you can host all of your fancy VC friends or something? I think a restaurant's always going to be cooler. Uh, a restaurant and a nightclub are probably the two worst things you can start. The other thing is, most restaurants, their financial profiles are, like, 80% alcohol. So, like, if you're not, like, my dad wanted to start a hookah bar. He's Muslim, like, a long time ago. He just retired. He has some money. I want to do something. And then he didn't want to sell alcohol there because Muslims aren't, uh, they don't like, like alcohol that much. I was like, dad, don't. Unless you just want to light a bunch of money on fire, don't start it, because you're going to make all your money from alcohol. And if you don't do that, then don't do it. So I think kind of side change, I think the whole SaaS is over kind of thing. I think it's so overblown. I think you're seeing it now as well, with kind of the AI stuff cooling, where we have, like, one of the best math skies in the world. There's a thing in math called regression to the mean. I think you're starting to see the regression to the mean in AI, and I think you're also starting to see, um, I know we have two CEOs here, but I think the people that have the most AI psychosis are the people farthest from the work, and that's usually the CEOs. They're like, oh, my gosh, look, you can do this. Why aren't we running our whole marketing department from that? And, you know, just like, okay, great for the demo, terrible for production. Not to say there's not a. Like, I'm working on a GTM in a box that's AI powered, not to say that there's not legitimate services out there. But I think the AI psychosis there is definitely a correlation with how close you are to the work. Um, to actually have the understanding and the domain expertise of like hey, this is a great demo, but this is not going to actually um, run the business. But I'd love to get your guys takes.

Speaker C: I mean I think Sean is on the record saying it's important to be popular and this is a great way to be popular.

Speaker A: It's true, it's true.

Speaker C: So I mean like I, you know, if someone tells you what they, who they are and what they do, you should kind of just listen to them. That's my. Yeah, I like people, people say who they are. I and I. And Sean is like a master, you know what I mean? Like he has correctly recognized that uh, being popular online gives massive leverage and massive alpha, especially in the age of AI. And so this is like a great. First of all it's an excellent tweet. So like a plus Sean. Second is like I don't even know if anyone can define what SaaS means in the context of that tweet.

Speaker A: That's super fair point because we're, we're

Speaker C: consuming more software as like in general than we've ever consumed before. Like last I checked, at least $44 billion more, maybe $80 billion more. Right. Like we just added two whole uh, salesforces in the last two years worth of consumption uh, of new software. So sure seems like we're consuming more software.

Speaker B: SaaS is a business model. I think you could make a case there. This idea totally.

Speaker C: Everything has become usage based. Yeah, everything's becoming usage based. But did you know that they sell contracts for a discount at Anthropic and OpenAI? And that's like a big part of the business now too. Um, so I don't know man. I, you know, like, um, this is like a, it's like, it's like any other hyperscaler. Right. Like you sell credits, like all that sort of stuff. Right. It's like any other hyperscaler. Yeah. I just think we're consuming way more software. I think that's amazing. I think the software is providing a ridiculous amount of leverage. So like something I think that's very underappreciated is how many people do you think work at anthropic?

Speaker A: 500?

Speaker C: Maybe like 6,000. Yeah, yeah. How many people do you think work at Salesforce?

Speaker A: 50,000.

Speaker C: Yeah, it must be. Yeah. So it's like ask that question but

Speaker A: you don't have the numbers. Yeah, that's about right.

Speaker B: Well, they got 83, 83,000.

Speaker D: Twice as many. Sheesh.

Speaker C: No, I'm just saying, guys, 80,000 was a big number.

Speaker D: 83,000 is, you know, it's almost twice as much.

Speaker A: That's, that's crazy.

Speaker C: Wait, wait. Yeah, yeah, but I'm just saying there's an order of magnitude and then maybe a doubling. Right, for less revenue. Okay, so, so the leverage gap is not to be underestimated. Right. It's like, um, it used to take, you used to have to create 80,000 jobs in order to produce $45 billion of revenue. Now you only have to create 6,000 jobs in order to create $45 billion of revenue. And so that part I think should not be underestimated. But the consumption of software has sure as hell gone up. So, so I, I, I think that that's actually the like under appreciated part of like what's happening right now is, yeah, there has never been more leverage in software than, than there is right now.

Speaker B: I would just say agency. Like I've been, I've been thinking about the agency business a little bit and all of the agencies I know are leaning into software ultimately. Like they're, they're creating their own software or they're leveraging more of software to give themselves better margin and to, to create more leverage with the employees that they have. And so like that's the, you're not going the opposite direction there. You're actually leveraging software more in those roles than has been done previously.

Speaker A: This was the corollary, I was going to say, where, um, I think there's going to be a really big boom in SaaS, businesses M or services based businesses with outcome pricing masquerading a SaaS, where like you're an agency and you're like, hey, I'm running, I'll do all of your gtm, et cetera, et cetera. I just think a lot of people would rather have the outcome than the software. And I think that previously that just wasn't possible. Where there was, you had to choose, you had to say, hey, do I have to have all this headcount? It's kind of dovetailing with Rishi's point of like, do I have to have all this headcount to actually service these people versus like, hey, I have a bunch of hitters. I have way more domain expertise, you know, your business. I can make this GTM brain and then actually service you. You don't really care about the software. And the other thing is when you do that, you can actually accelerate your software way faster because it's a paradox. But it's actually easier to design an F1 car than it is to design a mass market production car because you have all these people. And the analogy I'm making there is I have to have an interface that a bunch of people are going to understand versus I have one interface that I can train my team on to get the output that it needs. And then you can have HTML artifacts, et cetera, et cetera. Like the client doesn't need to see the messiness, but I think that's the, the, the software leverage paired with. I don't need to build a ui, I just need to train a certain set of Tier 1 operators to do the thing to create outcomes. I think is a really interesting, I mean, obviously the very top of mind because I'm building in that space. But I think like outside of gtm, I think there's a lot of applications verticalized where domain expertise is actually very expensive. I see the value chain in three ways. For AI, you have data, you have models, and then you have prompting inference. And with the data, the models thing is essentially commoditized. All the frontier model, that game is done. You have the players that are going to play there, but then you have data and then you have inference. And prompting data is a really interesting place. I think there might be a fourth linkage in terms of like how do I route to the right model for cost. But I think that could probably land in the inference data or in the inference, um, value or link. But then you get to the data and the inference and the prompting or the prompting and the inference. The only way to do the best prompting is actually be an expert in a domain expert. And so I think that's, that's kind of my little correlative hot take. So, um, that would, I think, be more aligned to what Sean was saying. Like I would rather start a. Because that's a services based business and restaurants are services based. I would actually rather start a services based agency that's AI powered than a restaurant.

Speaker B: Agencies have so much more pricing leverage than software does too. So if you just talk about like this software services component. If I were to build software again today, I probably would try to distribute it through an agency and really focus on that avenue because there's just so much more pricing power when you go to a. Uh, I remember Sean Frank had this tweet probably a couple years ago and it was basically like if Your software's under 500 bucks a month I'm signing up and I'm trying it um, if it's over that it becomes a whole conversation. Well 500 is like nothing when you start talking about service fees. Uh, right. Like I don't know how much they pay agencies but uh, a company that size to be spending mid five figures to six figures on agency services is nothing. Right, but you're gonna, you're going to then care about a 500amonth software subscription. If you layer the software into the service, you can charge way more for the software. Ultimately. Um, I think that's a, it's just a, ah, that's, that's the way I would, I would try to sell software at this point.

Speaker C: This is very vertical dependent. I don't, I don't agree with this.

Speaker B: I mean that's true but that was

Speaker A: my, that was my thesis though. It needs to be verticalized. Like this is not horizontal play. This is, I can run GTM or I can run a certain vertical that you are trying to get done. I don't think you can say hey I'm going to come in and run your company. Like I think, think that is. But I think verticalize the thesis holds.

Speaker C: I guess what I'm m trying to say is as a software company I spend more money on software than I do on agencies. Um and I would not spend more money on an agency than I would on software. And in fact I would be like this is, I would, I would almost categorize it as crazy if we spent more money on um. Like how is it that we are not able to use software to figure out how to do this even if it's expensive software. So that's why like that's, that's sort of more. What I'm saying is like I think depending on the vertical every ver. Like so commerce as a vertical or whatever, commerce as a vertical I think has been coached to pay uh, for agency services in a way in which other verticals have not. And I think that that matters a lot um, in as much as they've been coached not to pay for software by Shopify. Right. And so like the sort of like, like the playing field matters a lot for the industry that you're selling into. Um, and not all industries are the same. Like if you were selling dev tools for example, like imagine being an agency owner trying to sell agency services to a developer. Like developers notoriously will not work with outside developer agencies or things like that. Like it's, it's like um, an you get like allergic reactions like no matter how good the outside agency is, but they'll pay a ton for dev tools. So, so it, it just, it, it shifts a little bit. Like, by the way, that being said, in the places where you've been trained to consume services, I think every fast growing AI app company is selling services just for clarity. So like Sierra sells services, Harvey, uh, sells services. Like all of these companies are doing it, and they're doing it in like, very nuanced ways for, in order to drive adoption of their platform. Because if they don't, then they actually get relegated to, um, like Jeremiah, what you were saying, like an outcome that's like not worth it at all to even do in the first place. Right? So there's this like, crazy map that you have to navigate depending on which vertical you're selling into for what go to Market Motion. You should choose depending on the industry, how they buy, how they prefer to buy. And so like, you have to repackage based on who you're selling to essentially every time. Um, and I think that that's like, that you didn't have to do in the SaaS 1.0 world. In the SAS 1.0 world, everybody bought the same way. However, well it worked. And this nuance is like a new nuance, um, that I think has stretched. Most go to markets in ways that they were not expecting to be stretched. Right. That's why you have all of these crazy things happening. Harvey hires a lawyer to go be part of the deployment team in every sale. Right. Like, that's unusual.

Speaker B: It's, uh. My dad owns a small, uh, well, health food store, grocery store. He's got, I don't know, 60 people

Speaker C: that work for him.

Speaker B: And he was telling me about his time tracking software and how much he pays for that. And I don't remember exactly what the number was, but I just remember being shocked. I was like, how are you paying so much for something that's like so simple? But it's because it's like, designed for what he, for his business, uh, uh, and his use case. And so he was gonna, he was gonna switch to a different Software and save $6,000 a year or something. And I was like, the fact that you can switch time tracking software, which is literally just like punch in, punch out, there's an app to make sure that like, your employees are actually like, there's a notification system if your employee isn't clocking in when they're supposed to what their manager gets. That's it. It's like, you could probably build it in Claude. I mean, there's probably more nuance to this. Right? But you can build that in cloud in an afternoon. Um, but this is something where it's like he's, he's saving $6,000 a year by switching to this other thing. I'm like, how are you paying so much for this? But in that context, it's an important piece of infrastructure for him. And, uh, and they don't have any, like, it's, It's a very. They don't have much software. Right. It's like Excel and QuickBooks and this thing. And this thing is like, it has value that, that you perceive, uh, value for him that's worth paying for. Even though I look at that as like, this is a very simple thing that shouldn't cost you anything.

Speaker A: Yeah. So I think this is a really important point. And going to the building off what Rish was saying on the paradigm shifts previously, you could build a really great product and figure out distribution around that. Now, because the price of development has dropped so much, building the great product doesn't really matter that much anymore. And you actually should concentrate. You should have some sort of mvp. Yes, the product needs to actually have value. Uh, but this is actually what the Cluley, uh, guys are doing now, where the Cluley guys, they actually have a crappy product, but they crack distribution, which is actually in my opinion, more important now. And so they're basically have. They buy up these Constellation apps that they know they can take viral and that's how they're making money. And so the, the, the broader point here is that what's fascinating is spinning up a sales team is really hard and it's really, really long, like time to actually spin up a great sales team. The perfect example every. I don't know if I used to work at Whole uh Foods back in the day and we use a software called Workday. I don't know if you guys know it's arguably the worst software you've ever used in your life, but here's the thing. Their moat is not their software. Their moat, their sales team is cracked. They have the best distribution and their sales team is absolutely cracked. So going back to your point, Jeremiah, if you cloud coded a better Workday app and it. So there's that 10x, it needs to be 10x better than the incumbent. Say it's a hundredx better, dude. You're not selling into them. You don't have the distribution, you don't have the sales force. So I think this is something that again, the AI Pill. Uh, it's just not true. Like, just because you build a cool app, dude, if you don't have distribution and if you're a sales ed, it is really, really hard to spin up a really great sales team. It's really, I would argue behind finding a cracked cmo. A sales leader is, is probably the other hardest hire you will ever have. It is really hard to find a really great sales leader and then they have to find people under them. And then on top of that, you're constantly fighting economics because these are like, love salespeople, but they're all mercenaries. So it's like, okay, cool, well, Stripe's going to pay me all this money because they see me killing it now. And you're like, dude, I want you to stay. But, but like, the economics just don't work. So I think that's one of the biggest fallacies that I've seen crop up in the AI development. We can build in Claude code kind of or Codex or whatever. Um, and it's something that has actually been really interesting because I'm also seeing this in B2B SaaS proper of the B2B SaaS, like go to market playbook has completely changed in what. I've been out of the game for like a year and it's, It's. The old playbook is dead.

Speaker C: I'll say something somewhat provocative. I think traditional SaaS is not. I mean, this part I agree with Sean. Traditional SaaS, like, doesn't really have a place. So you have to actually be like AI software, right? Um, again, that's why the definition of SaaS matters. But you have to be AI software to be used for some sort of department or industry or vertical, whatever. If I, if you do that right, I actually think one of two things is very likely to be true. Thing one is if you don't have some way to get PLG adoption, I think you're effed. Um, and the simple way to see that is most AI software that's growing really fast has a mechanism to be adopted in a PLG way. Okay, so that's one sort of thing. The other thing, on the other side, on the enterprise side, if the CEO is not selling to the CEO of the other company, I think you're dead. And this is Decagon, Sierra, Harvey, Lagora, uh, Serval, whatever. Uh, based on like, like all of these sort of like enterprise companies, the learning has been actually, you need to build a new sales motion that is centered around the CEO's ability to compellingly deliver the vision for what AI can enable for that organization using their approach for enabling it. And then there's a delivery team to map against that vision to how do I enable it for this particular organization where all of the services come in? Because like you gotta map the vision to the implementation for what feels magical as a first step, second step, third step, fourth step, so on and so forth. And so if it's in between those two, I think you're effed.

Speaker B: And um, the, the challenge then is that like there's only so many monetization strategies that can actually support that. Like you, you have to gotta be big numbers, bro.

Speaker A: If you got a forward deployed engineer, you got a CEO to a CEO talking like those are big numbers. Like it can't be a thousand dollar, like Sean Frank contract that he's gonna get. Like these need to be big numbers. The only thing I would add Rishabh to that is that I think not only do you need a PLG motion, you need a fast time to value like and so like you need that like really fun, quick, awesome moment. Um, in the plg, like it just can't just be plg. There has to be like the lovables of the world, et cetera, et cetera. Where it's like, oh my God, it launched my first website. Like it doesn't need to be productive but it's like the value of being able to feel like you can actually publish to the web is such a, uh, I think that's another thing where we're all in the.01% of the 0.1% of AI pilled.

Speaker D: Dude.

Speaker A: AI penetration is so like. And I think that's one of the pros and cons of San Francisco. The con is, or the pro is like you are so bleeding edge. Like you are absolutely. But it's also, dude, it's such a bubble. It is such a bubble to what the rest of the country is like and what the rest of the market is like. And I think that there's so much um, room for AI adoption, um, where we have not seen it. And I think it's hard for. I don't know if it's just me, maybe I'm just projecting, but I'm on Twitter and all these things like, oh my God, look at all these things. Oh my God. New models out. It was just like, dude, like I think there's, there's a, there's a balance that you're going to have to find of like having inspiration, exploration but at the same time like keep the main thing the main thing, and I think it's really. This has been the best time to be exceptional, the worst time to be average. But I think the problem with that is this is the easiest time to have really bad shiny object syndrome. And you never actually commit to an idea and see it through to see if it's actually really, um, a good bet or not. Because you're just constantly like, look at this, look at this, look at this. So that's my little rant.

Speaker C: Rabbi, just quick switch of topic. I'm curious what you. When you were leaving Vermont, uh, you were going to build your own company, which you're starting to build with War Machine. Then you also announced this, like, CMO job at tie. So help us tie that together. See what I did there?

Speaker A: He hasn't lost the step. Oh, my God. I don't care what that Dexa says. You are. You are full of muscle. Um, yeah.

Speaker B: Yeah.

Speaker A: So it's actually been.

Speaker C: More importantly, I. I think the thing that I'm more curious about is what have you learned in trying to build your own software company? Like. Cause that's like the sort of, like, again, for other people who want to build a software company and. Or who may be, like, running a department inside of a software company. Like, what are some of the things you're like, wow, now that I see this, I can better understand X or whatever.

Speaker A: Yeah. So I had a really awesome. It's. It's been a really great experience of ours. So for people that know I'm building something called doctrine internally, it's called the War Machine. It's. It's basically a go to market in a box for B2B SaaS, um, with a bunch of AI awesomeness. And, uh, Michael, CEO of Ty, he approached me. He's like, hey, I know you want to do this for the next three to five years, but I'll tell you what, Come build out a marketing department here. We'll deploy the software, you be the cmo, um, we'll pay a little bit of money. Everybody wins. So it's actually been great. It's almost in this weird way, like a incubator slash CMO gig. So it's been great. So the first thing I would say the learning was the War Machine is light years ahead of where it would have been if I didn't take this role. Because being able to deploy in a production environment with real people, real things, this is a real business, making real money. You see the. You don't build. One of the biggest issues, I think with AI, it was Always an issue. But with AI it's even worse. It's called supply side innovation. Basically you're just like, this is a cool idea, let's build it.

Speaker B: Yay.

Speaker A: If nobody uses it or if it doesn't generate value or like. And so being able to stay away from that because there's real people I can talk to the AES, I'm running the marketing part myself. So I know this. So that would be the one thing is get the software in somebody that matters hands. Even better, get somebody in the software's hands that is paying you. Like that is a, it's a big step because you'll get the homies helping. But once you have somebody where it's like, hey, I'm um, paying for this, there's this actual transaction that needs to be covered with value. So I would say that's the biggest thing there. Like obviously you want to build. Um, the second biggest learning was when you are building a software company, there is a certain point where the product is well and good and you will start to find yourself working on the product more because you can hide in the product. The product is very tangible. I can make impact on this versus actually building your company. And that has been my biggest learning where um, candidly this product was probably ready six months ago and I, I was just playing with it, doing it instead of. Actually I'm looking for my founding AI engineer. I'm trying to build an actual team for bringing on uh, an operator to run an agency motion like these things. So those would be my kind of big takes of like get it in hands and get feedback um, as soon as possible. Because that, that will really help to hone. Um, it'll, it'll bring you back to earth. Um, the second thing is like your ideas are great, but great doesn't. You can't eat great. What you can eat is actual money and value. And so get it into somebody that will give you money or some sort of exchange that puts you on the hook for generating value. Um, and then the third thing I would say is, um, if you're very, very kind of product domain expertise, heavy kind of bend towards ic, um, I would really, really watch yourself in trying to build a product versus build a company.

Speaker D: I couldn't agree more, dude. I really couldn't agree more. You said a lot there. We're talking about the services versus software. We're talking about, I guess people versus software, uh, for deployed engineers, agencies, um, building their own software and supply side innovation. I think you could be really careful. My SaaS uh, if you can call it SaaS. Yeah, it's a SaaS. It's a Shopify app, is very small. We have like 40 something paying customers. But some of those paying customers have all of the supply side innovation. You could say it's demand, but it's not. It is the agencies asking to build what they just came up with in their head. They're like, oh, can it do this? And then you can ah, be persuaded into thinking, okay, yeah, I've got to build this thing. And that can open up, don't get me wrong, for my business. It's opened up a few service opportunities which are interesting, building bespoke solutions. But that is the same supply side innovation. It's just coming from the agencies that are, ah, slightly less technical or um, you know, are coming up with ideas as they're seeing the demo of your product. And I couldn't agree more with you. Ideas are great, but you can't eat ideas. You actually have to sell the product that exists because it exists to meet a lot of their use cases, not the product that they're imagining when they're watching the demo. Right. Uh, because that customer, um, sure, they might be a customer of a 10, $20,000 contract, but they're not the customer of the few hundred dollars a month, um, SaaS. So yeah, I couldn't agree more. It's ah, something I've learned very recently in the last few months is like, let's get back to what the product actually does right now. Unless we're going to go there, right? Unless we're going to go to building something custom for you.

Speaker B: Isn't this one of the challenges with AI though? Like the, in the past that was okay to. And I'm not saying it's. You shouldn't. If you're charging 20 bucks a month, you have to do that. Right? Um, you can't just build everything for everybody, um, or whatever. There's a number at which you can't do that. But I think like what you guys are doing is kind of the opposite of that. Right? Because you've looked at it and said, well, if you can go, just do this with AI because somebody could say, I had this idea, I love this thing, I'm just going to go make it happen and then they're not going to pay for your product because they can go do it on their own. Um, there's that trade out there, that cost and benefit analysis that everybody's doing now in a way that they didn't before in the past. They Just evaluated between two or three options, and they would pick one of those options. Now you can do whatever you want in a way that you couldn't before. And I think that's. Where's the line? I guess, because there is a line there. I just don't. I don't know where that is either. Um, or I shouldn't say either. I don't know where that line is. But I'm curious what you guys think about that.

Speaker D: I think that's why this emphasis on selling is so important, though, because you have to sell at the end of the day, and if you have something to sell, you got to sell that you can't be in the sales call. And they're going, oh, can it do this? And they're like, you know, they've drifted off and they're imagining a different product and you stop selling to then start just like, re building the thing that they've half baked in the moment. You know, it's a. Yeah, I think you gotta be careful, right?

Speaker B: I mean, this may be why I was always so bad at sales. I always, like, the product is never good enough for me. And like, my own product that I build that I think most people would say was pretty good, was never good enough for me, um, in my mind. And it made it really hard for me to sell.

Speaker C: Wow. I'm the opposite. I'm the opposite. I love fucking selling. And I didn't realize it until the last three months.

Speaker A: And I don't think I'm always a great salesperson. I always wanted you at events because you were, uh. You had this weird or not weird, but this unique skill of, like, incepting people because you have this very kind of chill demeanor, but your. Your intellect is so loud that you don't need to be this, like, loud, vivacious, crazy guy. And then by the end of the time, uh, you leave, parodying what Risha was saying. Because you're. You were the best salesperson. You're the best salesperson we had on the team, hands down. Well, maybe you're just really. But I'm.

Speaker B: I'll tell.

Speaker C: I'll tell you something later that I cannot share on the pod about actual, like, pipeline, uh, and, like, closes and things like that. But anyway, over the last three months, I've been doing more direct selling in the market. And what I have learned about direct selling, Jeremiah, is exactly what you said. Which is if, first of all, you have to have extraordinary confidence in what you believe about the world and how your product articulates what you believe about the world. But I don't think that, I think you can hold true in your head. Like the product has to get better at the same time as evangelizing the vision for what you see the world should be and why you believe the world should be there. And this is what I mean by like this founder led sales thing, like in this bucket over here and then PLG in the other bucket. And I think in between the two is the zone of death. Um, I really believe that, uh, and I struggle to actually find companies growing really fast in the zone of death right now. I mean, I want to see counterexamples. But Jack, my curiosity for you is if you're not gonna sort of like go in the direction of like, I have this vision. And Jeremiah, you're right. Like I paint the vision. I hear what the customer, how the customer attaches, how they want to shift their business based on that vision, and then we sort of deliver a thing. But Jack, if you want them to buy a tool, shouldn't the forcing function on yourself be like, what drives this to get adopted? PLG, like 100%, 100%. Shouldn't you at some point be like, I can't get, I can't even get on the phone with people. Like if I get, if I pick up the phone, something has gone wrong. And I, uh, mean, maybe you pick up the phone in order to learn what went wrong. Right. But is the disposition, I'm on the phone, this is wrong. Or is it, I'm on the phone, this is right. Do you see what I'm trying to ask? Like, I'm very curious about that because I know Raba used to push me a lot on this. Um, and I think now it's 100% true. In the world of AI, like the selling needs to happen on the site and in the first interaction with the product, that's your salesperson is your marketing your website and the first experience with the product. And if that's not your salesperson, the other salesperson is all the way over here. It's like Brett Taylor doing sales this year.

Speaker D: Yeah, Yeah, I think it, you know, I haven't, I haven't thought about it that way. I thought about it when Jacob Postle recently brought up that he was in back to back sales calls. And I was like, isn't this what you get sales reps for? I haven't had, uh, enough sales calls in a day for it to feel too draining to me. I think it's good to discover what it is that people want. People want XYZ product and then that's actually a good idea for um, what we build next. And then also I prefer it because the actual contracts, the um, services contracts are so much more significant than the actual software itself. The software is almost just like a gateway into uh, the services. So I think that that's why the sales course makes sense. But at the same time I can't build everything.

Speaker C: Right.

Speaker D: Um, and there is a lot of things that the software does that people just don't even know because they're dreaming of what the software could do.

Speaker B: Right?

Speaker D: And those dreamings of what the software could do are half baked. They only just came up with them. So after a while of hearing those half baked ideas, you realize I can't build all of these just because they came up with it. I have to find a way to bring it back to selling the thing that actually exists because it exists to solve a real problem that they have. Um, you know, like I said, like you said actually Rabba, ideas are great, but you can't eat ideas, right? This problem that is actually solved by the software and does this thing that right now your team do and it takes them, I don't know, 10 hours out, uh, of their 40 hour week or whatever it is.

Speaker B: Right?

Speaker A: See this is why I'm so bullish on AI powered services. Like uh, an agency. An agency can solve all of this. I can sell you into the vision, I can nuance the software. There's, I mean I could also be, I'm a hammer, here's a nail kind of thing. Uh, but it's like X CMO selling into marketing departments saying hey, all you guys need to do is give me one or two tastemakers. I will handle all the toil. Because I think the biggest thing about uh, and I can nuance it where you have the, so again it's hammer nail kind of thing. But there's the War Machine code base and then there's the GTM brain. So it's almost like uh, a Tony Stark Iron man suit. So I can take Fermont, I could take your app Jack, and the War Machine still works. It just has different contexts. And so that's where I think you get leverage. To your point, Jeremiah, because to do that previously an agency would have to have all these bespoke workflows and like you just couldn't do it with AI. You can now because you have, okay, cool. If you're a domain expert, you know what the marketing ecosystem is and you can bring it online and then you have the context of the company that stays up to date. But that's why I'm very bullish on that aspect of it. I think we're. The challenge comes in is knowing what you want to build, like kind of what you're doing. Jack, I would argue that what's your goal? And if your goal is to make a bunch of money, I would use the app as a loss leader for a services powered or like an AI powered services business. And the uh, uh, best way. So people go through three phases. You go through problem aware. Do I know I have a problem? Solution aware. Is there a solution out there? Product aware. Is there a product that provides this solution? What you're doing there is just showing. Hey, Jeremiah. Yes. You have a problem for your time tracking software. You don't know. Okay, cool. Hey, here's a solution. You know, we could track all of your people, um, with this. Oh, by the way, we offer all these services that we can do for you, or we could sell you a software. But do you really want a software product or do you just want us to handle everything? You pay us money, you yell at us. And this is, this is the example of like yes, all this like AI VA stuff is very cool, but it doesn't beat just being able to text your VA and say do the thing. And I think that that is in this AI world, people having somebody to yell at, that's one of the biggest things in enterprise is like, hey, if something blows up, my head's not on the block. I can call and say, Rishib, why is this doing so well? Oh my God, I'm going to get a promotion from you or Rishib, hey, this thing broke like uh, support contracts and having that kind of direct line contacts again, it could be hammer nail bias because I've been thinking about this a lot. But AI powered services gives your business leverage. So what did Rashid bring up at the earlier revenue per headcount so you don't run into the supply side kind of issues of agency of having to balance bringing on more supply to meet the new demand. The demand leaves. Now what do I do with all this excess supply? And then the second thing is you still have the power of nuance. I think it's. And you get the CEO to CEO selling. I can sell you the vision. I can say, hey, we can do this, man, this is awesome. And then you stay out of that like trough of sorrow or that kind of dead zone.

Speaker B: This is why I think software, even like inexpensive software that you theoretically could buy code will always have Value what you just were saying, Rob, about the, uh, you need somebody to yell at if something goes wrong. I mean, do you really want to, like, let's say you're running a Shopify business, selling on Shopify. Do you really want to be responsible for the 20 different apps? And, uh, what happens when something goes wrong? Um, because it will go wrong at some point. And I think, obviously, I think Rishabh, you talked about this in the past. Like, the technology's going to continue to get better, the ability to maintain those things will continue to get better. But at a certain point when the Shopify API changes and you didn't catch it, it takes you a month to figure it out. Uh, and now this, uh, I'll just use surveys. Your survey tool, uh, hasn't been collecting data for the last month and you didn't notice. Well, when you're using something like a. No, that would be fixed in hours, uh, not in a month and that. So like, it's worth some amount of money and that the amount of money that it's worth is going to differ from person to person. But there's a, there's a, there's some value there, even if it's just like the insurance layer, for lack of a better term, uh, that exists in that software.

Speaker C: Um, so yeah, I think you only end up buying it once you have a couple of problems.

Speaker B: Yeah. So you're like a 6 to 12 month timeframe there where it's like vibe code, everything, all this is great. And then after there's uh, some amount of time where those things start to go wrong. You're like, shit, do I really want to like, manage?

Speaker C: No, dude, even, even worse than that, the person internally at your team leaves.

Speaker B: Yeah.

Speaker C: Like the pragmatic way that this will end up happening, by the way, is not like you're at a company and you're like, man, I don't need all this SaaS. Like, I'm gonna build this thing. Like, and now I'm Jeremiah. I'm like, I built version one. And then I'm like, yo, Rishabh, get over here. I built V1. Can you make V2 of this? And like, you maintain it? And I'm like, yeah, sick man. Like, I'm, um, gonna, like. And then every now and again something will come up. And then I'm like, three weeks later, I'm like, Jeremiah just got pushed by Anthropic. And he's like, well, you kind of have to take that job because I'm not gonna pay you as much as Anthropic is gonna pay you, so good luck at Anthropic, you know? And then that piece of software is now an orphan, and then it breaks, and literally no one has any idea how to fix it. And they're like, I know what to do. I know what to do. I'm gonna ask Fable to go look at it.

Speaker A: No mistakes.

Speaker B: Yeah. And the problem, though, is like. Like, literally yesterday, I. Cause I'm like, I'm out of. No, now. Out of stamps, all this kind of stuff. I spent two hours yesterday working with one of the engineers at Node trying to get access to my AWS account, uh, so that we could make sure that other people had access. Because this is just something that was, like, set up years ago. We really only use. We don't use AWS for our main infrastructure, but there's a very important piece, one very important piece that lives in aws. And so it's just like, it was, uh. We needed to get everything set up. And, like, the. I haven't logged in in so long. The authenticator was, like, on m. An old device, and it didn't work. And it's just like, this whole thing, right? So it's like, that's fine and good, because we did that in a moment when it didn't matter. But what happens when something that matters is broken and you have that experience? Somebody's left. You need to deal with it. You don't know they're on vacation. I mean, I've had this happen in the past where it's like, you have something break. The person that set it up three years before is on vacation. Uh, they don't have their phone. And now you got to figure out, how do I get access to this thing, uh, in a timely enough manner to actually get this back online? So, yeah, those are just the kinds of problems that you don't want to be dealing with when it matters.

Speaker D: Raba, before we break here, because people will get angry if they listen to this conversation for 45 minutes. And we spoke to Raba for this long, and we didn't ask how the playbook changed. You mentioned that the playbook from last year has completely changed this year. Um, I'm personally interested, uh, in software.

Speaker A: I haven't validated the new one, but I can give you my theses if that's fair enough. So the, um. I think AEHR is probably doing the best. Ramp was my power ranking. And I think Air, uh, Ari over at air, Shane, everybody. I think AEHR is probably doing the best marketing. And what I'VE realized is kind of going back to that usually you could have had great content. You put a few influencers around it and then that content would surface. Then you get your distribution. It's just not the case anymore. And so what you really need is. I boiled it down to essentially three pieces. So one, you need a villain. So Air had Dropbox. Air has like having a villain, it doesn't necessarily need to be an actual company, but like, what is your villain? The second thing is you really need almost like a talent scout. And this was Ari. And Ari has a couple of people under him, but this was Ari where, dude, I'm old. So I remember when memes used to be like days, weeks long and like you could write out a meme do. The Zeitgeist is so quick now. So you really need like, uh, I'm going to sound ages here, but like some sort of Gen Z or that's just perpetually online, like just always online, knows the memes. This was one of the things that we had at Triple well, where Tommy was literally Twitter, like he could tell me everything, who's who, what the memes were, what the. Like. So you need that kind of talent scout to then be like, oh, Jack is on the up and coming. He's not there yet, but we need to get on him, pay him ten grand instead of a hundred, because in three months he's going to be really expensive and all the legacy companies are going to pay him and we're going to be head of the curve. Ari did this with the Rizzler. He did a bunch like, so you need that really awesome talent scout. And then the third thing is, I think all the alpha is gone in TikTok and Twitter. Is it still useful to be there? Absolutely. But like, if you're trying to penetrate or if you're trying to work, just the alpha is just not there anymore. So really where the alpha real, um, lies, especially in B2B SaaS. I think enterprise Richard would agree. I think if you're up market, you really just need to be events and relationships. Like that's, that's the show that's. It's not a hard playbook, it's an expensive playbook. Not a hard playbook. But this is more of like that 50 to 150 to 250 ARR SaaS of like, hey, you're growing, you want to get, uh, you know, mid market upmarket. Um, so the third thing is building a essentially like a shadow army of propagandists. So basically, I don't know if you guys are familiar with comfort, but like, uh, essentially what comfort did but for B2B SaaS. And so you can take over TikTok, YouTube, Instagram, you give them weekly briefs, maybe you incentivize some of the micro to mid influencers and then you just have. Have you guys ever heard of the, the Tom Cruise experiment, where basically, uh, you come over to my house and I go, hey, Jack, did you know Tom Cruise is in the room next door? You know Tom Cruise in the room next door? Then Rishib comes over and goes, jack, you know Rabba has Tom Cruise in the room next door. Jeremiah comes over and goes, dude, can you believe Rabba has Tom Cruise? Tom Cruise is in the fucking room next door. But once you hear the same messaging over and over and over and again, I'm not saying you should do shady shit, but this is how you can start to propagate your distribution because you're giving these content briefs. So I think that's where you're gonna see a lot of cracked, uh, B2B SaaS. Um, playbooks going is really indexing heavily on distribution, leaning out of LinkedIn and Tikt, finding somebody, some sort of talent scout to keep me inundated of what the, the, the people are talking about and having some sort of villain. So I think that's going to be the new three three Play playbook.

Speaker D: Dude, you got to stop. You got to stop. We're going to see a bunch of B2B sass with Tom Cruise and their logo reels on their website,

Speaker A: and that's not bad. Uh, we got Tom Cruise here.

Speaker D: All right, man, uh, you're off to Econ Cowboy now, right? Fun. Well, listen, thanks, boys.

Speaker A: This is incredible. Appreciate you guys. Appreciate you, man.