Latest / The Goats of Growth / AI Fluency Is Building Tools Other People Use — Scott Hanson, Founder of Tenfold Labs
Transcript
- Jay Webb: What's good, everybody? Welcome to another Scott Hanson: Mm-hmm. Jay Webb: episode of the Goats of Growth, where I interview go-to-market leaders and founders at tech companies to learn how they've been adopting AI and the effect positively or negatively it's having on go-to-market teams. Today's episode is brought to brought to you by Common Room, the AI Go-to-Market intelligence platform, trusted by top-performing teams like those at Okta, Anthropic, and Fireworks AI, to name a few. Common Room's agentic flows help your team know who's in market. Act on those signals automatically and run go-to-market motions at scale. Click on the link in the show notes to learn more. today's guest is Scott Hansen, founder of Tenfold Labs, who builds AI systems. Hi, Scott. AI systems for Fortune 500 Teams, who is currently building out the internal AI team at a $50 billion company. I'm not sure if he'll let us name who it is or the subsidiary of that company, but I'll I'll leave that up to him. Scott Hanson: Mm-hmm. Jay Webb: He previously built the world the largest event staffing marketplace in the world, which I'm quite envious of and and and wanna wanna probably learn a little bit about that. nevertheless, Scott, welcome to the Goats of Growth. Scott Hanson: Thank you so much. It's an honor to be here. I'm grateful. Jay Webb: Absolutely. So yeah, I'm excited to to have you on because I know you're really into the weeds. I want to get into Tenfold in a moment, but before I do, tell me about your AI journey. I know Scott Hanson: Yeah, you bet. Jay Webb: that you had started this, you know, certainly digging into it well before most people. Tell me where the inception was, where that began and Scott Hanson: Mm-hmm. Mm-hmm. Well, a little bit backing up, my history is in tech, as you mentioned. And in 2014, I started Pop Bookings, which is the Uber of event staffing. So if you go to any concert, trade show, sporting event, anywhere that you see temporary staff working, odds are pretty good they get booked through my app presently. And we have something like 400,000 active staff on the app. I did it the old-fashioned way where I raised money from Silicon Valley and investors and grew it to a team of about 30 and and had a lot of fun doing it. But that's where my my journey in tech began. I was a COO and co-founder. And I was a non-technical founder. And I imagine that's where a number of people who are really big in the AI scene now are starting is like you are as a visionary and as like an entrepreneur, there was nothing more frustrating than having this like product flow that I wanted to work exactly the way I designed it. And then would need to translate that through developers, then well, through front end design and then through to developers. And then ultimately get something that was like what I thought, but we'd iterate for months until it actually finally got there. And I'm one of those people that I'm like super nitpicky and want pixel perfect designs and code and things to just function with little delight me features like animations on buttons and things like that. And you have to pay extra for that. And you have to try to find developers that you can communicate these visions to. And then they think creatively in order to do that. And we've been hiring people overseas. And I don't know if you've done any overseas development, Jay, but it's it's a it's a complex thing. You you you trade off oftentimes the creative thinking aspect of a really high quality coder in the States versus price. Really, at least that's how the game used to be. So along comes Chat GPT, and I have stepped away from pop bookings and we can dig into why or or all of that. Yep. Jay Webb: What year was that? So tell me the time. So you so along comes Chat Scott Hanson: So Jay Webb: G P T around what year? Scott Hanson: 22 is when I started playing with it. So Jay Webb: Okay. Yep. Scott Hanson: I wasn't like the first of the first, but I was pretty dang early. And Jay Webb: Yeah, really juice. Scott Hanson: immediately I saw the ability for it to like create cool things, whether it was, you know, everyone started with pictures and just chatting with it and mysteries of the universe and trying to understand what AI actually was at that time. Cause you know, you're talking to something and it's responding like it's very smart and conscious, but then it's like, okay, so are you alive? Are you not? And So, that's where it all started, but that got into coding. Very quickly, I started building my own apps. I was doing executive men's coaching and needed an app to help men track their progress with their goals and commitments and various things like that. And I built that with its own little kind of chat flow within it. And it really just launched my my app building on AI. And so I was super, super early on that. And just had a lot of fun because I finally could see this vision I had in my head that I always had to translate to other people finally come alive in my own hands. Jay Webb: Let me ask you this. So you mentioned ChatGPT app, by the way. I at that time I was using this app called, I don't know if you even if anybody even knows about this, called Inflection, which was something that Reed Hoffman from the founder, obviously, of LinkedIn Scott Hanson: Yeah. Jay Webb: built. And I was using it, and it was I was describing it to my friends as basically the smartest person, knowing the smartest person in the world and having it in the palm of your hands. And I would just sit there and chat with it. And then what happened is I remember. He I got we got a message in the platform or something like that that it was going away because of ChatGPT and Gemini all started to come in and I don't I think Reed was like, okay, that they they've got this. So it's it's interesting. But let me ask you this. because ch most people when they hear Chat GPT, they think exactly how you were just alluding to in terms of chatting with it, getting an output. At what point did you like what were you using to evolve to those apps that you just mentioned? Because I just stumbled upon Apps recently, vibe coding apps recently. Were you were you using ChatGPT interface with something else, an early app that none of us were paying attention to? Tell me about that. Scott Hanson: Yep, I was one of the first, I I would say first hundred or even earlier of users of this app called Rork R-O-R-K, which a lot of people know of now. It's kind of a lovable competitor, Jay Webb: Mm-hmm. Scott Hanson: Rork.app. But they had just launched and I saw them on Twitter and was like, Holy shit, this is exactly what I've been looking for. It's like an interface for Chat GPT that will allow me to see the code that I'm creating. And I I was immediately one of their first power users and was chatting with the developers and and the founders there were only two people in the company at that time which is super impressive and seeing them vibe code that just inspired me to go even deeper. But I was consulting them like, hey, okay guys, you need to add this, this, and this. If someone wants to build a company off of this, here are the things that they need. And we would have chats back and forth as they were developing that company. Jay Webb: And just wanna rewind for a moment. You first discovered them on on Twitter, is that right? Scott Hanson: Yep. Yep. Yeah. They're like Ukrainian guys or Russian or something like that. They're they're from the Eastern Bloc. Jay Webb: That is truly amazing. I love that. And and so and what was the first maybe you said was the first app the tool that you were using for your coaching, men's coaching? is that Scott Hanson: Yeah, I was building a coaching app, but very quickly all of that evolved, the the coaching piece, I still do that. I just still do men's executive coaching, which is fun. I I coach CEOs, I've coached CEOs of of Fortune 500 companies, actually, one. And Jay Webb: Mm-hmm. Scott Hanson: and just learning that I could build an app that was super robust, that was storing user data, that was doing all these things that I knew from all of the build at pop bookings, how to set up the back end architecture, And how front ends work with back ends, which a lot of people, when they start their vibe coding journey, don't quite understand. It's like you go on Lovable and you're like, make me an app. And Lovable will, but all of the extra stuff that happens at scale, none of that is there in some of these like vibecode build yourself apps. And Jay Webb: Yeah. Scott Hanson: so that forced me to actually build my own AI agentic app system. It was like right at the time that OpenClaw came out. I saw everybody run into OpenClaw and I went a different path. And I think that path is still really relevant in the time. Like I this would be a super interesting topic to dig into the difference between bot development, like a gentic swarms developing versus staying in the CLI, you know, the harness, and having human doing the coding superpowered by AI. And we're at a tipping point on that, but early on. A lot of people thought open claw was gonna be the path that they could, run their whole business off of open claw. And we all found out pretty quickly that that doesn't work very well. And so I went the full harness route right from day one and it served me extremely well in business and in in the contracts and things that I got early on. Jay Webb: I'm gonna talk about tenfold in a minute, but CLI Scott Hanson: Yeah. Jay Webb: because I know that a lot of people may not know what that stands for. What does it stand for? And tell you know what what's give me an application for. what what is it? Hold on, let me see if I can help you. That's right. Yeah, okay. Scott Hanson: man, you're gonna you're gonna put me on the spot on exactly what CLI means. command line interface. But what I'm speaking of when I say CLI is clawed code. That's an example of a CLI. Codex. So it's the app or the platform that you're using one of these models in. Some people use Hermes, some people use open code, some people use conductor. That's actually what I use right now. VS Code is a CLI, and that's what most developers were using prior to. All of these AI tools. Jay Webb: And tell me what was it? Because that's important, right? Like I think a lot of people are frankly, I'm I'll I'll raise my hand and say, yeah, like this is this is what I do, right? Like I'm self-taught and I've been in the terminal for a few months now and I'm building on Claude Claude and Claude Chat and I'm using Replit and trying to just piece it all together the best I can, but I'm not so n knowledgeable, I guess if you will, to be like, you what, I don't need this over here. I'm gonna go over here and do it this way, right? with open c and everybody was running to Open Claw and you went the other way, as you mentioned, what was it that you weren't getting from OpenClaw that you felt like you could get yourself? And then how did you how did you learn it? Was it just trial and error or Scott Hanson: Yeah. Yeah. I think these are really important questions that if I could drop a lot of value on your audience right now, it'd be to set them up for success on how to get really moving quickly on developing as an agentic developer. So I think the benefit of going the route of using CLI, which I would call human in the loop. It's like what, and this is my own terminology. I don't, I'm not like industry lingo all the way here. I kind of bridge the gap, but With OpenClaw, it was really get it set up in a way that you then could just pass work off to this black box that you don't know what's happening, but something's happening and it gives you an output. And sometimes it gives you a great output, and sometimes it gives you total shit, and you don't know why. And so, the big issue for me with it was both security because you don't know what the agent's truly doing if you're running a professional open claw setup. You don't want it to have access to your personal data, you don't want it to have access to your machine. your business, all of these kinds of things. You want to treat it. I I heard early on, you wanted to treat open claw like a a a homeless guy that you hired off the street and gave him like a a machine gun and hope that he hits the target, but he could shoot you too. And so Jay Webb: Right. Scott Hanson: when you use just clawed code or or in my instance the the pro the setup I'm running right now, which is very much about to evolve here probably this week, is I use a thing called conductor, conductor dot build. Jay Webb: That with a C or a K? Okay. Scott Hanson: C. Conductor.build, it's a a VC backed, I think they're Sequoia backed startup, don't quote me Sequoia, it might might be in Dreessen Horwitz, but they are a it's like an evolution of VS Code. And what it allows you to do is it allows you to run work trees, and I I don't know how technical we want to get here. Like a work tree, so like how how who Jay Webb: This. Scott Hanson: am I speaking to here? Like, should I speak? To total newbies or should I speak to people who have an understanding a bit of this stuff? Total newbies. Okay, well then we're gonna zoom out and not get in the weeds at all here. Jay Webb: No, total newbies. Yeah. Dumb it down. Well no, here's what here's well no, here's what I want. I because they're total newbies, which I wouldn't necessarily say I'm a total newbie, but total newbies who are highly curious and engaged and actually want to get into the weeds, right? So Scott Hanson: Yeah. Yeah. Jay Webb: I would put myself in someone who wants to learn more about VS Code or at least at a high level, whatever. Scott Hanson: For sure. I'll speak to I train CEOs of very big companies how to implement AI into their business. So I'll speak to that person who understands very smart, very technical, but just hasn't dug into this stuff. So we're not going to get into work trees, all of that kind of stuff, or what the the benefits of conductor are here at the moment. That's something that I'd be happy if anybody wants to reach out and ask me. I'll explain it, especially because I'm moving off of conductor here very soon. But the point that I'm trying to get at. is I wanted to see what the agent was doing all the time, which is inefficient. It's not like if you set up agent swarms really well, let's say you're Carpathy and you just you're an absolute genius at AI, okay, amazing. Go for it. But most of us aren't. And what I found was that if I start with something like, we'll just stick with clawed code for now. If you start with clawed code, you can build trust with these agents over time. You can see what it's doing. You can start to take actions. And to me, that was the most important thing is to the speed at which you can take actions and then give it work that genuinely makes your life easier. So, when I have people start out, I have them come up with a task that annoys the crap out of them and they just wish somebody or something could do. Well, if you pass that off to open claw, you're passing it off to a black box. But if you keep it in clawed code, you get to watch what's happening. Again, inefficient, but Allows you to start to understand the inner workings of AI. And not only that, if you keep it on your computer, you are much safer than if you're running cloud code. I don't want to tell anybody that you're super safe by doing this stuff on your computer, but you don't have an agent living on a cloud computer that is accessing the internet and doing God knows what and giving people God knows what information or that can be hacked or any of these kinds of things. Does that make sense? Jay Webb: Yeah, it does. Let me just ask a follow up question. When you say so you can see what it's doing, when I s hear that, I I I I'm hearing that I'm looking at my terminal after giving it instruction and I'm watching it write essentially the code, I guess, if you will. Is that where you is that what you mean? Yep. Scott Hanson: Yep. That's exactly what I mean. And you're you're seeing and starting to familiarize yourself with the language of coding. Jay Webb: Yeah. Scott Hanson: And I think that's super important. It's it's if I could use a metaphor, you don't necessarily need to be the mechanic to work on your car, but it benefits you highly if you want to understand how systems work together to watch the mechanic doing his work. You might not understand all of the terminology or why this part goes where and all of that, but you start to get a feel, the engine's here. The transmissions over here, the oil pans down here, okay. And I see how those flow together. And that's what I think people need to start training themselves on. And it's so simple if you just watch Claude do it. It's boring as hell, but if you just sit and try to let it absorb, it really Jay Webb: Yeah. Scott Hanson: helps. Jay Webb: Yeah, I thought of, just in my own world, what you the what you just gave there as far as the the mechanic is me as a go to market recruiter and working with go to market teams over the last twenty years or so, CROs, VPs of sales and so forth. And I don't I'm I've never held any one of those positions. I can't Scott Hanson: Mm-hmm. Jay Webb: do it nearly as well as them. I don't know all of the things that everybody knows, even when I'm interviewing people and so forth, but I can have enough I can know enough to be dangerous, right? I have a I I understand what you're saying in terms of the the jargon and all all that stuff, for sure. So that makes a lot of sense. Scott Hanson: Absolutely. Absolutely. Jay Webb: before I get into the actually know what I'll ask this I'll ask this this question actually and then I'll get into the other one. let's talk about talk about Tenfold, Tenfold Labs in terms of Scott Hanson: Yeah. Jay Webb: the the large company that you're working with, how you came across them and and what were the challenges that they were having That you that you're helping them with. Scott Hanson: You bet. So this journey of building apps on my own helped me build a lot of confidence. I started by building them for myself and started like having cool tools. And I, like I said, I built my own what I would call harness, which Jay Webb: Mm-hmm. Scott Hanson: is like a bunch of different essentially software that agents have built that just run things for me to make my life simpler. So for example, something that's running my email, something that's running my finances, something that's running my calendar, something that like all these things that do. What everybody aspires to have, which is like a Jarvis, like a Tony Stark Iron Man type AI that comes and just runs his life for him. That's what the the aspiration of OpenClaw is. That's what the aspiration of Grokbot is. Again, not there yet, in my perspective. That said, when I had everything set up to allow myself to just be faster. So we're talking like setting up things that allowed my own personal productivity to skyrocket. And When I have all those connectors up, even though they're not happening autonomously, meaning I'm not walking into my office and some voice is saying, hey, you have 15 new emails today, blah, blah, blah. I've come and in clawed code just hit a button and say, What's going on with my emails today? it's a it it's still me having to go type. That process allowed me to maximize my productivity probably a thousand fold, honestly. And it built a significant amount of confidence when it comes to Doing a gentic development. And that was the most important thing that led to an amazing conversation where I was at Gold's Gym in Austin, Texas, which is where I work out in the morning at 6:30 in the morning every day. Jay Webb: Well it. Scott Hanson: And I was in the sauna, and there were two other guys there. And I I'm gonna share the details of the story because it's a good story and it's kind of uncomfortable too. So there was a kid in there, and he this was like, I don't know, this was back in February. And There was a lot, and I'm I'm gonna be careful in how I frame this. There was a lot of conversation about Israel and its effect on the government and various things. And we're not making any political statements here, but in this sauna with just three people, this kid reaches out to this guy who's an older man and is just like, So what do you think about the Jews taking over the world? And that is how he opened the conversation. And immediately I could feel the tension in the sauna just be like, whoa. Jay Webb: Crazy. Scott Hanson: And I'm just watching, and this guy's sweating. And I'm like, crap, this is like a landmine of a conversation. what an opener. Holy crap. And no kidding. Jay Webb: Yeah, no kidding. Talking about an attention grabber, yeah. Scott Hanson: And so, like, he starts trying to pry on that. And and this gentleman isn't responding to those questions, obviously. And the kid goes over to me and starts asking me. And I I'm like, I'm not, I'm not going there. I'm not touching this. Jay Webb: Yeah. Scott Hanson: But I redirected him and was just like, Yeah, that's cool and all, but I'm actually more interested in AI right now. What's your take on AI? And just pivoted the conversation. And that Jay Webb: Okay, yeah. Scott Hanson: brought this other gentleman back in. Jay Webb: Yeah, okay. Scott Hanson: And we got into a really deep conversation on AI. And it turns out he is the VP of a very, very big Berkshire Hathaway subsidiary. And he and I connected and started having a very deep dive on AI. And that I think was, I mean, that was the birth of Tenfold Labs. That was the birth of a lot of different things, but that Conversation ultimately led to a massive contract as a solo developer with a Berkshire Hathaway subsidy. And when I say massive contract, this is something that Deloitte would most likely charge around a million dollars to do. I charged a fraction of that, but Deloitte would have delivered it in eight to nine months, maybe a year. I delivered it in about two to three months. And I delivered it at a fraction of the cost. But thanks to AI development, I have almost no overhead in my business. So all of that overhead that a massive company, one of the big four, consulting companies would have to eat, I don't have to. And so like I can take those haircuts on my on my actual like what I charge. So what that ultimately led to is way faster development, higher quality development and way cheaper for for the company. yeah, go ahead. Jay Webb: I was just gonna say there's so many different lessons there that we could get into. I Scott Hanson: tons, yeah. Jay Webb: mean, my my my my goodness. okay. So tell me what you learned from him in that conversation in the sauna, no less, that Scott Hanson: Yeah. Jay Webb: you know, ultimately you ended up ended up turning into that contract and what did you what did you help them with? What what was the what was the problem and then what was the outcome that you achieved? Scott Hanson: You bet. I think by being able to like if I don't say the name of the company, I can give more details on what I did. so I'm gonna go that route and Jay Webb: Sure, yeah. Scott Hanson: people can probably do their own research to try to figure out what this company is. It's not that hard. they are in distributions and they run trucks for some of the largest companies in the world and they do 50 billion in annual revenue and they're a Berkshire Hathaway subsidiary. They're freaking huge. And what's happening with a lot of companies that are in that position, these Gigantic behemoth companies is they're still running on very old systems because they made those investments, millions and millions of dollars, to the big four, companies and development companies, all that over the years. And it's a big, it's it's a lot of work to shift off of old technology. So internally, they are they have been deep in that shift. And that has led them to adopt a Atlassian product that. tracks everything within the company across all of the different portfolios and divisions. And what that looks like is all of the all of the initiatives, all of the ideas, all of the the things that executives in every department, what all all of the stuff they want to work on and the stuff that's in process. So think of it like a giant project management system across the entire company. All of that data is in very unstructured format right now. And so there's a team of executives whose job is to pull that from each portfolio and consolidate that information into a digestible report that then goes to the C-suite of the company. And this is across pretty much every major company on the planet, even Amazon and Apple, and you name it, they all have this type of problem, which is like prior to the AI revolution. All of this stuff is done by a bunch of human beings who go and grab these reports and they put them into Power BI or they put them into Excel and they compile all this data, but it takes them weeks to do it. And then that report gets to the C-suite around a month to three months. So it's either a month or a quarter from when that data's relevant. And then the executive needs to make decisions that are future-looking based on that very dated data. Well, with AI, you can build. What I'm calling transparency layers or consoles that take in all of the data from different places in the company and it's agnostic on whether that data comes from an API or that data comes from an Excel spreadsheet or that data comes from somebody typing it in. And it compiles all of it together and shows various different views, whether that's graphs and charts, whether that's reports that executives would be writing, all of that, but at the click of a button, real time immediately, so that from here forward, executives are not making. Decisions off of very old data. And then they can even take that and expand it and look forward. So now we're we're because we're getting real-time data, we can do real-time planning. So that means that massive companies who have these big initiatives can now real time see all of those initiatives and pair them against exactly where they're at right now and what they're on track to hit and what they're not. Does that make sense? Jay Webb: Yeah, it does for sure. Like let me ask a follow-up question, but Scott Hanson: Yeah. Jay Webb: before I I do, let me just read another spot for Common Room. Common Room pulls your CRM, your product data and real world signals like web activity and other buyer intent signals into one cohesive view. One of their products, Person360, unifies account profiles across companies and keeps everything updated continuously. With Common Room, your team can stop guessing and start knowing. click on the link in the show notes. To learn more. let me ask you this, because obviously the company that you just alluded to is quite a large company. and you just mentioned some others that obviously will have the the same problem or do have the same problem. How many initiatives at a time are they managing? And did you get a sense of that? because it sounds like there could be a lot. Like, in other words, you could have one executive managing 10 initiatives themselves, right? So it's not just like one initiative for for per executive, right? Scott Hanson: For sure. without giving too many specifics, I know exactly how many, but it's in the hundreds. I'll tell you that. And yeah, yeah, we'll we'll leave it at that. Jay Webb: And can you quantif are are you able to or have they been able to quantify what that means either in opportunities gained, opportunities that weren't lost? Like you know what I mean? In other words, is there a quantifiable value to put on what you've done for them? Scott Hanson: Yeah, absolutely. Yes, there's a huge quantifiable value. there's a couple that are a little abstract that I'll start with. One is as executives implement this, so so once you are pulling in your data, all the data that's available to you and the company, and it's being organized and aggregated in a very efficient way, Jay Webb: Mm-hmm. Scott Hanson: it gives insights into the dark areas of the company, the areas where no one is reporting, where someone might be slipping, where someone is. underreporting things that are actually happening. And so there's a lag here that companies, when they do adopt this type of thing, need to have a grace period, but ultimately you're gonna see all of the ugly stuff that, like, you know, if an executive comes and reports to you as a CEO, like, my my department's amazing. I'm I'm grading myself at very high because of this, this, this, and this. And you're like, why are we always dropping the ball on this? Like, why is our customer churn so high? Or why, you know, all these kinds of things. Well, odds are pretty good. Not that this person's being dishonest in any way, shape, or form. There's just a gap in what data is actually being tracked. So one is it's illuminating all of the areas in the company that is that are just dark. And immediately you can see, we're not tracking this block over here. And no wonder we're we're churning customers because this is directly related to how fast we're doing support tickets or whatever it happens to be. And so that type of tracking immediately. Lowers churn. That's one area that that's huge with these types of initiatives and reporting tools. Second is headcount. Like it or not, AI is coming for your job. It just is. And this is something that I wanted to get into on the podcast if we have time, which is how can executives, especially middle management executives, that these types of tools do replace the reporting aspects of their job? How do they become very can't think of the word right now, but how do they support themselves to not get replaced basically by AI? Jay Webb: Yeah. Yep. Scott Hanson: that's something I want to dig in with you because I think people really need to understand that and understand how to do it. so there's more but I'll sounds like we we hit a point there's a question so let me pause. Jay Webb: Yeah, let me well I'm I'll you just throw me the softball, so to speak, f to to Scott Hanson: Yep. Yep. Jay Webb: to tee up the next question, which was I had had a convers or multiple conversations with folks either online or offline about how AI isn't necessarily responsible for let's say cutting five people on the team, but it is responsible for keeping them from hiring another five people, right? And middle management, I think, is what you're alluding to, which is in large part, and by the way, the role the part of the role that they hate the most. Right, which is the reporting and all of that stuff, Scott Hanson: Exactly. Jay Webb: that is now less of their job. And to your point, what you're just alluding to can go away. And I'll just even go a step further to give it a little bit more context to what you're gonna say is what I've been saying is is that there is starting to be com there is starting to be a skills gap. I think there's always a skills gap, but the skills gap in AI fluency and people who are you know, non-AI fluent, I guess if you will, and maybe not adopting it. And so what's going to happen is the people who are made more AI fluent are going to be able to keep their jobs and protect their jobs more because they're going to they're just going to be well positioned and and other people are going to be at jeopardy. And then those other people who are or are looking for a job versus somebody who is more AI fluent is going to be they're going to the person with AI fluency is going to be that much more competitive. So tell me what your thoughts are on Scott Hanson: Yeah. First want to define AI fluency. Everybody and their freaking grandma is using ChatGPT, chatting on the browser. That is not AI fluency. Just because you have AI giving you advice about your life and all these kinds of things, you're not AI fluent. AI fluency means you are using AI to build a tool that another human being is using to support themselves. There's building tools for yourself. That's awesome. That's a starting point. That's where you need to go. But it's a whole nother game when you're building tools that Other people use. Okay. So I would qualify that as AI fluent. And it's so much more attainable than most people realize. It is actually Jay Webb: Mm-hmm. Scott Hanson: one of the easiest things to do if you just spend your evenings chatting like you would on Chat GPT in Codex, in Claude Code, in any of the CLIs that are out there. So that's Jay Webb: Yep. Mm-hmm. Scott Hanson: the starting point here. What I want to convey more than anything to people, especially if you're listening to this podcast and you made it this far, it is not hard. It truly isn't. And what you can do to set yourself up for enormous success here is start to look at the aspects of your job that you hate. The part where you're like, God, I spend four hours a day doing this and it, it's like exhausting. It's a grind. It's gr it's it's grunt work. And see what it would take to have Claude do it. But don't do it on Claude on the desktop. Or on the browser, I mean. I so many people do that. They're like, yeah, I read my I fed my contract into there. I put my Excel spreadsheet and it was awesome. But then their workflow becomes like putting it on the browser. It's going to lose context. It's not repeatable. It's not an actual piece of software that supports other people. It's just your own little way of using clawed code and it loses context all the time, which means it's not actually that good. So what you want to do is do the exact same thing in clawed code. Just take a little bit of courage. chat with it. It looks weird. It's got code. It's got all these scary things. But if you just talk to it exactly the way you would in the browser, give it your file and tell it to do the thing, you will be amazed at what you can do. And it's really, really cool. And so that's baby steps, but that's where I would open this door. Jay Webb: Yeah, for sure. and actually do me a favor, walk Scott Hanson: Yeah. Jay Webb: me through your process for dissecting that, right? In other words, you give somebody that advice, right? Like, okay, think about something that you are that you can't stand doing. So I I'll give you a perfect example. Scott Hanson: Yeah. Jay Webb: So for me in my world, frankly, anything that has nothing to do with people, I don't like it. So Scott Hanson: Yep. Yeah. Jay Webb: so I like my podcast, I like interviewing, all this other stuff. Everything else in the background is quite painful. So right now I have a workflow That I'm in the process of setting up. It's not completely set up, but I'm working on it. I'm actually looking at my other screen right now. That basically prepares me for podcasts, right? So we do you and I did a pre-interview. that pre-interview transcript gets fed into Claude. Claude looks at it, summarizes it for me, just gives me context, gives me suggested questions, Scott Hanson: Mm-hmm. Mm-hmm. Jay Webb: emails me all of that information. And then once the recording is done, it produces the the the episode, right? It produces the episode, Scott Hanson: Yeah, yeah. Jay Webb: produces all the shorts, schedules all the shorts, starts rolling them out, right? So it's all of these things that basically saves me time. I'm able to do more podcasts in a day. all of those things, right? But what Scott Hanson: Absolutely. Jay Webb: is and we can even use that large company Scott Hanson: Yeah. Jay Webb: you just mentioned because I think that is just a great example of the opposite of just an individual, but where you're where you're creating agents for very large organizations and hundreds of millions of dollars in revenue. What's your process for helping people Break that down and figure out okay, here's the initiatives that you have, walk me through that. Just what does that Scott Hanson: Yeah. Jay Webb: look like? Scott Hanson: Yeah. I I think the example you just shared is amazing because where's your one of your zone of geniuses? Well, it's in people. And Jay Webb: Mm-hmm. Scott Hanson: that's the thing that you're really, really good at. What we can like branch that to would be domain expertise. So what is it that a middle management executive has? They have domain expertise. They have some sort of understanding how an internal business system works that if you just tell Cloud Code, run my company, it doesn't have that. You know what I mean? You understand podcast, how you want it, how you want your workflow to work that allows you to get more sponsors, that allows you to get these great guests and all of these kinds of things. That if I just tell Claude Code, make me a podcast, it's not gonna do anything. So, what we want to get to is what is what's your domain expertise? What's the thing that you know that nobody else knows? Or that it doesn't have to be that nobody else knows, but that you know really well. And we're gonna build systems around that. So first you're gonna feed. That context into whatever you're using. We're gonna say Claude Code for now. You're gonna give Claude Code that context. You're gonna ask Claude Code to go research about this thing. You want it to get, you want to just build a bubble of like this is the stuff we're focused on. And if you're using Claude Code versus chat, all of that context stays localized on your computer for it to reference over and over again. And how we set it up properly and Claude.md, and there's like you can get in the weeds a mile deep, but. The point here is just having that documentation. Let's say it goes and does that research. That document it produces, whether it's a.md file or a PDF or HTML, lives on your computer and it's going to reference it each time afterwards. So you don't have to keep filling that context and it doesn't have to hold it. It will just go look at it when it needs to. So that said, we're going to find out what your workflow is. And the best way to do that is to just like literally walk your machine through your day. Day is A, B, C, D, D, F, G. Okay, this is my normal day. And then we start to look at what are the things that Claude can take off of your plate. So in that process, I imagine we can identify six different areas that Claude could take off of your plate. And what we want to do is try to focus on them one at a time, not try to eat the whole elephant. So if we could break down your process into like the key nuggets, like the six nuggets that are like really important in your day. On this process of developing a podcast. Well, what we can do is map out a workflow or like a happy path of exactly how that should flow. And then my recommendation for people is to build an interface. So you can actually have the agent just go do that workflow for you. But we're back to the black box again where you can't see what's going on. You just get an output and you don't know where or why it gave you what you gave you or what broke. But if you build that workflow into just a really simple Interface that lives on your system that just walks you through each step where you can see things, that's a good way for you to play with building a piece of software, modifying and adapting a piece of software, and actually seeing the software do a job for you. And it's cheaper. So, how is this cheaper? Well, what we don't want to do is have a g agents own the whole process end to end because they start to do screwy things. After you run an agent, if you just tell an agent, go do this workflow 15 times. It's going to deviate, break, and do all these other things. But if you have an agent build a piece of software that does the workflow for you, or hits an agent at a certain point in the workflow. So, like for example, at this company, the Berkshire Hathaway, the reports are agent generated. But all of that data comes in through very standard basic ass software. You know, it's just a machine that you pull a lever and it gives you the exact same output every time, no deviation. But then how that output is explained or or analyzed, that's the word I'm looking for. Agents Jay Webb: Yeah. Mm-hmm. Scott Hanson: can do that, but it's just that piece and they have just a very specific job right there, and that's it. And then it's back off the agent, goes back to the software again. Well, if you do Jay Webb: Yeah. Scott Hanson: that with your workflow, you're gonna have an amazing output every single time. Does that make sense? Okay. Jay Webb: Yeah, I love that. I love that. It's fantastic. Actually, let me ask you this. So the the example that I just gave you, d should I because right now it's essentially a workflow in Claude Code, right? It's in or or Claude Claude Chat, basically the the the the desktop, right? Sh Scott Hanson: Yep. Yep. Jay Webb: are you saying should part of that be moved into software, some sort of an agent, I guess if you will, off of that Scott Hanson: Yep. Jay Webb: is what you're saying? Tell me about that. Scott Hanson: Odds are pretty good it already is software. So if you're telling if you're using Claude Code and you're telling it to go do something like, hey, run my podcast workflow and you've turned that into a skill or something like that, Jay Webb: Mm-hmm. Yeah. Scott Hanson: it will have likely already built software on the back end. You just don't see it. You just don't know that that's what it's done. So then the question Jay Webb: Interesting. Mm-hmm. Scott Hanson: is like which pieces of the output are constantly not looking or feeling the way that you want. Jay Webb: Mm-hmm. Scott Hanson: And that's where I would. Have it create more of an interface style workflow. And that interface, Jay Webb: Hmm. Scott Hanson: ideally, you click a button, does a thing. Then it, when that output happens, you do another thing. You can keep all of this in the CLI. And I do for some of my things. Jay Webb: Mm-hmm. Scott Hanson: But by and large, creating that interface is going to give you so much confidence in building a tool that Jay Webb: Mm. Scott Hanson: other people can start to use. Jay Webb: Yeah, that man, that's fantastic. one of the things that you said is, you know, that as you're as you're building your own workflow and and thinking about the work that you do as an individual day in and day out, one of the recommendations I have, and I'd love to get your thoughts on this, is that, you know, you could actually ask Claude to just interview you about what it is that you're doing on a on a daily basis, right? And maybe Scott Hanson: Yes. Jay Webb: it's thinking of questions that you're not even asking yourself. Is that is that fair to say? Scott Hanson: 100%. And there's actually a bunch of skills and very smart people who've put that together. the one I would recommend is have your agent go to GitHub and grab Matt Pocock. M-P-O-C-O-C-K. Matt Pocock's skills. He's got, I think it's 26 skills. He refines them constantly. He's well known for this. He's got 200,000 stars on GitHub. It's like one of the most popular repos on all of GitHub. And you have your agent install those skills, and on it it has it's called Grill Me. And you every time you want to do a new process or you want to do anything like that, you run the grill me skill and it's gonna take you through a very specific workflow to ask you questions that are relevant and to get to a certain deal. And then there's a next step after that. Matt Pocock has a quick less than 10 minute video that walks you through exactly how to use those skills. It will 10x your quality of code, your ability to use agents, all of that stuff immediately. Jay Webb: That's fantastic. I'm gonna go check him out for sure. Matt Pocock, P-O-C-O-C-K. Scott Hanson: Correct. Jay Webb: Fantastic. Okay. Let me ask you this one last question and then we'll get into the Rapid Fire Five, which is how do you how do you start to if you're an individual who's listening to this and you know, you can even use the CO as an example that you that you mentioned, or maybe not, maybe down from Down from maybe even a smaller company, how do you start to build trust in all of this? Right. Because there's so much about, you know, even in the own their own interfaces, it says AI can be wrong. Or and you hear people from Anthropic coming out against themselves, a lot of cases in terms of stuff the Scott Hanson: Yeah. Yep. Jay Webb: stuff that they can do. So tell me about that. How do you how do you start to build trust and build confidence in in the AI agent building? Scott Hanson: Absolutely. I think that's if I could put a theme of what I'm trying to help communicate on this entire podcast, the value I can give here is Jay Webb: Yeah. Scott Hanson: go basic, get Claude Code. It's a good interface. It's not the best, it's not the fastest, it's not the cheapest. But we're speaking of specifically Claude Code, which is the desktop app from Claude. Download that, click on the code tab, and then just do one thing. One thing. Have it give you a daily brief. Of what's on your calendar or have it one of my favorites, this is one of the best, easiest ways to improve your life right out the gate. Use a agent to record your calls. So whether that's fathom or Fireflies, or if you do Zoom, the Zoom cloud, whatever, and connect, ask Cloud Code to connect to the API for that cloud recording software. So if it's Fathom, We're gonna download, we're gonna have it download the transcript from all of your calls and ask it to analyze and take actions from those transcripts. So, what that's gonna do is you can build a little piece of software that every single time you do a call, it takes those people in, ingests the transcript, puts them into a CRM, sends out emails. you can expand on this huge, but start small by just having it ingest the transcript and give you some ideas. But from there, you're going to very quickly be able to go, wow, I can manage every conversation I have immediately at the like when it happens, I can have actions taken and I don't have to physically take those actions. So I'm not going to personally update my CRM, have Claude do it. I'm not going to personally send the email, have Claude do it. And that is one simple workflow that for me has immediately changed the game. Jay, you and I, right after we got off the call, Claude analyzed the Fathom Transcript and sent a note to Graham, my my co-founder at tenfold, we organize this with you and then you, here I am on the podcast. So it's e it's so efficient. Jay Webb: Yeah, and that example that you just gave is frankly one of the first ways that I got started, which is it started to, to your point, ingest the transcript from my client kickoff calls. And what happened was is that I would start a search and before I even was able to walk out of my office, frankly, the agent already had a company list built and a number of different things already created for the search. Right. And it was just as simple as just listening to the the call. So I think that's I love that. It's a very good example. And as far as building trust, it's like when you have an output where it's successful like that, right? Where and it's not touching anybody else. In other Scott Hanson: Yeah. Jay Webb: words, just touching your own workflow. And you look at it and you Yeah, this is this is great. And all of a sudden you build more trust and you can start building more things and more things. And yeah, for sure. I love that. Fantastic. All right. you ready for the Rapid Fire Five where I ask you Scott Hanson: I'm ready. Jay Webb: Five questions, you give me five answers. All right, cool, cool, cool. one more sponsorship read for from our sponsor, Common Room. Common Room is the AT AI go to market intelligence platform that defragmatizes your first party data and combines it with real world buyer signals. Context 360 enriches, resolves, and keeps it always fresh, for sure. yeah, and data, by the way, is is is is obviously as you might agree, extremely, extremely important as you're building agents. because we're Scott Hanson: A hundred percent. Jay Webb: Yeah, without it it's like, what are you doing? Scott Hanson: Yeah. Jay Webb: All right, here we go. Rapid fire five. First question, what motivates you? Scott Hanson: What motivates me at this present moment is that I am on the front of the wave and I know it. And it's so freaking cool to like be riding the wave of a massive s civilization level shift from the seat that I'm in. And it's just like I feel like I've got a lot of responsibility from that place and and every day I cannot wait to get up. Jay Webb: there with you. Name a big goal you have right now and when would you like to accomplish it? Scott Hanson: Big goal I have is to move from software into hardware. So right now we are building a massive piece of software in the trucking industry. So this is outside of the Berkshire Hathaway subsidiary I was talking about. We're building a a fleet management system, like a dispatch layer for all of the trucks. And at some point that's going to transition into hardware. And I cannot wait to start playing with robotics where I pair AI and all of these skills with AI into actual robots. And that will my mind's already blown that that can happen. And I haven't taken that leap yet, but I expect that to happen here in the next probably four or five months for me. Jay Webb: Wow, that sounds game changing for sure. That's exciting. You have to come back in the next, I don't know, four or five months to give me an update on that, Scott Hanson: Can't wait. Yep. Yep. Yep. Jay Webb: what's your most preferred way to learn new information to stay sharp? Scott Hanson: right now I'm finding that Twitter X has like if you follow the right accounts, there's like people who are sharing some of the best AI accounts out there. And if you follow Carpathy and even Claude Devs, like Claude Code Devs and just a number, this this guy, why am I blanking on his name right now? Tebow, something like that. he's one of the open AI lead developers. You follow those guys, you are going to be on the cutting edge. And I love just seeing what they're talking about. Jay Webb: That's fantastic. I'm gonna have to start following these folks too as well. Carpathy. I'm gonna have to get you actually what's what's the is it is that his last name, Carpathy, or is that his Scott Hanson: Yep. Yeah, it's it's Andre Carpathy. He's K-A-R-P-A-T-H-Y. He was one of the founders of OpenAI, and he also helped build Tesla, and he is now at Anthropic. Jay Webb: Amazing. All right, for sure. Definitely start following him. what's your first sorry, who's your favorite character of all time? Scott Hanson: Mm. It's been shifting. Right now it's Tony Stark just because I'm like I feel such a kindred spirit with him. And at the the his his final the way he went out was so heroic that it's like, he's got he's got this Christ energy about him that I'm just really down with. So Jay Webb: I think you might be the second person in the first eight episodes or so that I've pre interviewed pre recorded that. may have said him. I'll have to go back and and look. I'll have to ask Claude to look for that. I'm pretty sure. Scott Hanson: Mm-hmm. Mm-hmm. Jay Webb: I love that. That's fantastic for sure. And you mentioned him earlier. last question. What's your favorite thing to do when you aren't working? Scott Hanson: when I'm not working, I love to go off into nature with my wife. I I love camping. I love I just recently got back from a backpacking trip with some guys where we hiked up a really big mountain. It was very scary, very awesome, and just love like full out of tech. Like just no phones, no cell service, any of that kind of stuff. Jay Webb: Amazing. How often do you get to do that? Scott Hanson: About once a month. I got a a group of dudes that there's four of us total that we go and do that. That is how we get away. So Jay Webb: Sounds nice. Sounds nice. You ready for the bonus question? Where my Scott Hanson: Let's hear it. Jay Webb: last question. Yeah, absolutely. Actually, no, let me see if I can find it first. Hold on. Matter of fact, while I look for that, what question would you like for me to ask the actually, nope, sorry, got it right here. Sorry. Okay, here we go. All right. here is this qu the question that came from Alison Barris. He's the CMO at Active State. if you had the chance to go to Mars or on a distant planet, would you go? Why or why not? Scott Hanson: Ooh. it depends on when and if I could get back. So if I can't get back and I'm I'm there, I would go towards the end of my life. Or what I would imagine hope to be the end of my life, which would be like, you know, my twilight years. if not, Jay Webb: Yeah. Scott Hanson: then no, I'm here committed to the to Earth, like committed to the family I'm I'm gonna be building and and the people and friends and community that I have here that I absolutely love. So yeah. Jay Webb: Fantastic. on earth is where you landed presumably one of your largest clients in a an asana. So there you go. Scott Hanson: Mm-hmm. Mm-hmm. Mm-hmm. Jay Webb: what question do you have for my next guest? Scott Hanson: Yes. I had one and now it's escaping me, so let me find it really quick here. Jay Webb: Sure. Well, while you're thinking, one one of my guests, it can range. One of the guests asked if you had walk up music like a a baseball player, you know, what would what would it be, right? Yeah. Scott Hanson: That's great one. That's a great one. That's a great one. I I if because I know a lot of people you you work with or you you have on here are executives and people who are running companies. So it's like Jay Webb: Mm. Scott Hanson: if you knew you could not fail, like if it wasn't possible to fail, what's the thing that you what's the step or action you haven't taken in your life that you you would take if you knew you couldn't fail? Jay Webb: knew couldn't fail, what's the action that you would take if you knew you couldn't fail? Wow, I like that. That's a that's a good one for sure. Luckily I don't have to answer that. So I'll wait for the next I'll wait for the next person to answer. That's that's that needs some that needs some thought. But you know what? The next question the next guest isn't gonna have the the benefit of time. I'm just gonna put them on the spot the way I put you there. So for sure. Scott Hanson: Mm-hmm. Love it. Love it. Jay Webb: Scott, this has been fantastic. I will say that I love digging into the weeds and learning new things. And one of the reasons why I like doing this podcast is I feel like the people that that do listen to this podcast also have that same mindset. growth, goat goats of growth isn't just growth in terms of revenue, but it's also growth in terms of spirit and mind and all that stuff. So Scott Hanson: Absolutely. Jay Webb: you definitely contributed to that today for sure. I'm so excited to get this out. what's the best way for people to reach out to you to say thank you. Maybe they have a question Scott Hanson: You bet. Jay Webb: I was unable to ask or maybe they want to hire you to to help them build their own AI. Scott Hanson: You bet. Jay Webb: Systems. Scott Hanson: Scott at tenfoldlabs.io. That's my email. Easiest way to get a hold of me. my agent will be listening and ready to grab your email, whatever you want to share. I am down to just give information. At this point, I know that the more value I can give people and just how to get started and and where to go, the more I'm doing my my service and good to humanity. So that's all questions are welcome. Jay Webb: Fantastic. And when you email him, make sure you put the goats of growth in the subject line so Scott Hanson: Please. Jay Webb: that yeah, that he gives your email a special attention for sure. So Scott Hanson: Absolutely. Jay Webb: I wanna thank you for that. I also wanna thank Common Room for sponsoring this episode. for everybody who's listening, please I wanna thank you, of course, and don't forget to subscribe, rate and review. it signals to folks that are going through their apps and looking for new podcasts that this one is certainly valuable. So I want to thank you everybody and until next time. Goodbye.