Discovery Conversation Guide: Ethan (Red River Web Design / Red River Automation)

Date: Mar 18 2026

Warm up (2 min)

How'd you end up running both a web dev shop and an AI automation business? Was that always the plan or did one lead to the other?

Boston 2019-2025 - deep learning in book store, banker at the time but wanted to learn how it worked so self-taught python. year or two couldn tbuild stuff so learned javascript. n8n is useful for businesses that is good enough for what they need. just json

Understanding his world (5 min)

When your clients come to you, what's the problem they think they have versus what's actually going on? they don't know what they want. sometimes they do and its easy. but people want to use tech because they feel like they'll be behind. they already know what they want they just want someone to do it for them.
Where do projects get stuck or messy for you? Like where does a client need something and you're either not the right person for it or it's outside your lane?

pretty rarely. im a big go to the doc person and i use ai to explain it. my answer is ill try rather than no, because i rather learn and get paid for it. subscription to o'reily tech books.

When infrastructure stuff comes up in your projects, like deployment, scaling, reliability, how do you handle that now? Do you bring someone in or figure it out yourself?

figure it out myself. supabase for backend, it helps me out alot. im a frontend engineer rather than a true backend engineer. services to be helpful.

How he buys and refers (5 min)

When you need to bring in another technical person for a project, what does that process look like? Do you sub out the work, refer them to someone else, or just scope around it?

i like to referr them to the person im working with. i dont like to markup someones service. if they want something for that person, we collab. example bringing in cyber security expert. which added a month to the project but it was worth it.

What would make you confident enough in someone to actually refer a client their way or bring them onto a project? there's a level of trust. bni is an example. help with something small vs larger.
Have you ever had a situation where a project would've gone better if you had an infrastructure person involved from the start?

probably. the big application. tell me more. its a platform where students can do homework socially.

The money question (3 min)

When your clients need infrastructure or backend work, what kind of budget range are they usually working with? Like are they expecting that to cost $500 or $5,000? Projects range from $1000 but 2k is ceiling to $35k. $100/hr for dev, $150/consulting. usually increase to 2-6k.
If you were going to bring in a specialist for infrastructure work on a project, what would you expect to pay them? Hourly, project-based, revenue share, something else?

probably hourly or project base. i'd like to refer my client to the infrastructure specialist directly. agency model is okay but i never managed anyone so that would be a constraint. and we would collab on the project and be paid indepentently.

The golden questions (3 min)

If you had to describe what I do to one of your clients, what would you say? two lines o business, business to professional where youare coaching engineers in specfiici problems. professsional support and larger infrastructure consulting on the other.
What's one thing you think I should be doing or offering that I'm not? Based on what you know about my background and what you've seen me talk about. developer i could help them with a particular problem. its typically larger businesses that need infrastrature. bni isn't large businesses. what i can do for them in this room, solopreneur or small businesses. I had to solve this problem.

i want to get a graph rag project on my own. claude to auto generate graphes essentially a monte carlo.

Close (1 min)

What would a good first collaboration look like to you? Something small where we can test how we work together before committing to anything bigger.

if you join the group, the trust thing will build and build. even things like this are a way to get to you know. you better and better. i think it would be fun to have a regular thing.

need to buy learningwithoj.com

truth over feelings.

Conversation Analysis

Overview

The conversation was conducted in-person on March 18, 2026, at a BEN (Business Expansion Network) breakfast meeting. What was planned as a 20-minute discovery conversation extended to approximately one hour, covering all 12 questions and then organically expanding into deep technical discussions about graph RAG, memory systems, AI tooling, documentation practices, and personal backgrounds. The extended length is a strong signal: Ethan was genuinely engaged and saw O.J. as a peer worth investing time in, not just fulfilling a favor.

Key Signals

How Ethan describes what O.J. does (the golden question):

His exact framing: "Two lines of business. One is business to professional where you are coaching engineers in specific problems, or coaching them with resume or job interviews or social skills in ways that they can present themselves better. Professional support on one side, and then larger infrastructure consulting on the other." This is the cleanest two-sentence positioning anyone has given across all discovery conversations. He naturally separated B2C and B2B without prompting, which means the messaging is landing. He also used "professional" rather than "consumer," which is a more accurate framing than the standard B2B/B2C split.

The collaboration model is direct billing, not agency:

Ethan explicitly does not want a markup/agency model. His preference: "I would want to refer my client to the infrastructure specialist. They are paying them directly instead of paying me and then I pay a portion." He's done this before with a cybersecurity specialist on the education platform project. This means your $325/hr rate would go directly to you with no middleman friction. This is the ideal collaboration structure.

Where the infrastructure gap actually lives:

Ethan admitted the education platform (social homework app for K-12) "probably" would have gone better with an infrastructure person from the start. He's built multi-tenant architecture, cybersecurity compliance for minor student data, and 10-12 social media feature clones using Supabase as a backend crutch. He described himself as "a frontend engineer with the ability to add a Supabase backend rather than a true backend engineer." This is exactly the kind of project where your infrastructure specialization would add the most value, and he knows it.

Pricing and budget context:

His projects range from $1K (which he now considers underpriced) to $35K. He charges $100/hr for development and $150/hr for consulting. His sweet spot is $2K-$6K projects. He wants to shift toward more consulting for sub-$5K projects and keep development for larger ones. His clients' budget range can support a specialist at your rate, especially on the larger projects.

Trust-building mirrors the BEN structure:

His trust process: meet several times, see examples of work, help with something small before something larger. "Trust but verify." This maps perfectly to BEN's rhythm of repeated exposure. Your offer to help with AWS questions for free was the right play. That's the "something small" that builds toward paid collaboration.

The anti-gatekeeping resonance:

Ethan bridged from banking to tech as a self-taught developer starting at 22. He explicitly called out anti-gatekeeping: "Not everyone comes from that background and their opinion in the meeting is just as valid as yours." He described running interference on people shutting others down. This is the same value system that drives Learn with O.J.'s positioning and it came up organically, not because you prompted it.

"Truth over feelings" as a developer culture value:

Ethan framed directness as fundamental to developer culture: "Most developer types, like truth over feelings. I'd rather you tell me than not tell me." He connected this to why developers are uncomfortable with sales ("you kind of have to live by omission") and why honest, non-salesy business development works better in this space. This validates Kirk's "I knew you didn't give a shit about hurting my feelings" signal from a completely different angle. It's now a 3-for-3 pattern across Kirk, Mickey, and Ethan.

He tested the AI guide on the spot:

Ethan downloaded the free AI guide during the conversation and said "I will read this today." This is both free product feedback from a technical peer and proof that the lead magnet works in a live context. He's the first real person to test the download flow and confirmed it worked smoothly.

The education platform teacher is a potential mathbliss.com connection:

The teacher client building the social homework platform is exactly the kind of person who would resonate with O.J.'s education philosophy. O.J. asked for the intro and Ethan was receptive. This is a research connection for the learning platform vision, not a sales opportunity.

BEN membership decision crystallized during the conversation:

Ethan's closing comment revealed that O.J. had already gotten "$2K worth of stuff" from the group before even joining, and the conversation itself was part of that value. The referral concern ("is the referral thing even going to work for me?") was outweighed by the relationship and knowledge value. O.J. indicated she was going to sign up before 3 PM that day.

Extended Conversation Signals (Beyond the 12 Questions)

The conversation naturally expanded into several areas that produced additional business intelligence:

Documentation as a pain point:

Both O.J. and Ethan identified documentation as a universal developer pain point: nobody wants to write it, nobody wants to read it, but everyone suffers when it's not there. Ethan's client (the teacher) pressures him for visible frontend updates over backend/documentation work. This is a potential content topic and a real problem O.J. could address in training: how to bake documentation into agentic coding workflows using Claude's markdown files.

The "Khan Academy for Engineers" vision resonated:

When O.J. described the long-term vision of a free learning platform, Ethan's response was "that would be sick." He specifically lit up about interactive real-world labs where engineers get thrown into broken systems and have to figure out what's wrong without being able to just copy-paste errors into ChatGPT. This validates the lab-based learning approach and suggests the vision has legs with technical peers, not just target customers.

Graph RAG and memory systems are shared interests:

A significant portion of the extended conversation was a genuine peer-to-peer technical exchange about graph RAG, memory graphs, vector databases, multi-agent handoffs, and distributed systems. This isn't directly business-relevant but it deepens the peer relationship and positions O.J. as someone Ethan wants to keep talking to, which is the foundation of a referral partnership.

Shared origin story pattern:

Both O.J. and Ethan have non-traditional paths into tech (O.J. from web freelancing and pet stores, Ethan from banking). Both are self-taught in key areas. Both describe the tension between wanting to build interesting technical things and needing to focus on business development. This shared identity makes the collaboration feel natural rather than transactional.

Hope Punk and creative writing connection:

The conversation surfaced that Ethan does creative writing on Substack and is interested in speculative fiction. O.J. mentioned the One Year Later project (journal-format AI story) and got genuine interest. This is a personal connection point that goes beyond business and strengthens the relationship.

The "learningwithoj.com" domain idea:

Ethan organically suggested buying "learningwithoj.com" as a redirect during the conversation. A small tactical win that came directly from the peer exchange.

Cross-Interview Pattern Updates

How people describe what O.J. does (4 conversations in):

Person	Their words	Framing
Kirk	"She's not some consultant, she's your team member."	Relational, embedded
Mickey	"Thorough, deeply investigates, doesn't jump to conclusions."	Process and rigor
Ethan	"Professional support on one side, larger infrastructure consulting on the other."	Clean B2C/B2B split
Susan	"Helps people market themselves to the current environment."	Career positioning

Nobody uses the word "mentor." Nobody says "coach." The language clusters around peer, team member, and specialist.

Directness as trust signal (now 3 for 3 with explicit language):

Kirk: "I knew you didn't give a shit about hurting my feelings." Ethan: "Most developer types, like truth over feelings." Both describe this as the reason they trust O.J. and the reason developers in general prefer honest communication over sales polish.

Anti-gatekeeping (now 2 explicit, 1 implicit):

Ethan: "Their opinion in the meeting is just as valid as yours." Explicit. Mickey: Valued learning across different collaboration styles. Implicit. This theme is worth elevating in content and positioning. It's resonating with people who came into tech through non-traditional paths.

The self-taught/non-traditional path pattern:

Kirk, Ethan, and O.J. all have non-linear paths into their current roles. This shared identity creates instant rapport and is part of why the "experienced engineers who haven't gotten the title yet" positioning works: it speaks to people whose credentials don't match their capability.

Actionable Takeaways

Follow up on the teacher intro. Ethan agreed to connect O.J. with his client building the education platform. This is a research conversation for the learning platform vision and potentially a mathbliss.com connection. Don't let it go cold.
The AI guide got its first real test. Ethan confirmed the download flow works. Ask for his feedback in a follow-up message. His perspective as someone who does AI automation for businesses is exactly the audience the guide targets.
The collaboration structure is clear. Direct referral, independent billing, no agency markup. If a project comes through where Ethan needs infrastructure work, the mechanics are already agreed upon. No negotiation needed.
BEN membership follow-through. O.J. indicated she was signing up that afternoon. Confirm this happened and track the first referral exchange.
The documentation-in-agentic-workflows topic is content gold. Both O.J. and Ethan identified this as a real pain point. A Hands-On Tuesday post about baking documentation into Claude Code prompts would resonate with this audience.
"Truth over feelings" is a brandable phrase. Ethan named it, Kirk demonstrated it, and it maps to the brand personality keyword "Real." Worth incorporating into messaging or content.
Buy learningwithoj.com. Ethan suggested it. It's a natural typo redirect.

Audio Transcript:

All right, so we're gonna do the the discovery conversation with Ethan. All right. So... So I we already did the intro over the phone, so I know about your business and stuff like that, really exciting stuff, what you're working on. Um, and I wanted to ask, uh, how did you end up running both a webdep shop and an AI automation business? Was that always the plan or did one lead to the other? So when I started learning to code, I actually, I was in Boston from 2019 to 2025. And I loved going into Harvard Bookstore, MIT's bookstore, like some of the ones on the university campuses, and everything was about deep learning. And I didn't know anything about it. I thought it was fascinating. And I was a banker at the time, so I had nothing to do with technology for work, but I just wanted to know how it worked. So I taught myself, Python, so that I couldn't then use tensor flow, and I could use, you know, some of the, like, psychic learn, and some of those libraries for data science. And then doing that a year or two, I was like, yeah, I can't really build anything. I can't, like, build an application and sell it. I't start a startup. And that's when I learned Java scrubs. So it's, that was why I am interested in AIs, because I've been keeping up with it since 2019. Um, and so with in it, in, and that kind of thing, I think I brought that in because it's useful for businesses that don't want to pay a large amount for a very custom, large piece of software. It's like, this is good enough for a lot of what they need, and I can understand it pretty easily, because I've already done the hard work of learning to program. This is just Jason, so it's not difficult. All right. Wonderful. Um, so, when your clients come to you, what's the problem they think they have versus what's actually going on? That's a good question. I would say the biggest problem that they, they come to me, they don't really know what they want. Sometimes people do, and that's nice. Like, someone shows up, they say, I want this exact thing. Can you build it? Usually, yes. And there's a job there, but a lot of the time is people want to use the technology, 'cause they know they'll be behind without it, but they don't know what they want. And so doing, like, a diagnostic, um, but a consulting up front is typically the solution to that. Yeah. It's I mean, Most of the clients that have big projects already know what they want. They just need someone that can do it. Like, for example, I have a teacher who wanted to start a platform that forced her students that she could sell to schools and basically transition from being a teacher to being an entrepreneur, but she didn't know how to code or anything tech wise. So she knew what product she needed to build, exactly. She had every single piece of it mapped out in her head, but she couldn't build it. And so that's where I can... All right, great. Um, where do projects get stuck or messy for you? Like, where does a client need something and you're either not the right person for it or it's outside of your lane? Honestly pretty rarely. I am a big go to the documentation, kind of person. Um, I also use AI a lot to explain things, but I don't understand it. So if I if I have if someone says, hey, can you do this? my answer is I'll try rather than no. Because I'd rather just read the docs and increase my skill set and get paid for it. So it kind of wins that way. And so I don't charge people for reading documentation. But or I have a subscription to O'Reilly. I have them on my subscription, so I can read any of O'Reilly's technical books. And so if I'm like, oh, I don't really understand, you know, some sort of niche database thing that I should know, I can just find the chapter in the right book and read it. All right, so, uh, when infrastructure stuff comes up in your projects like deployment, scaling, reliability, how do you handle that now? Do you bring someone in or do you figure it out yourself? I figured it out myself. Honestly, I used Superbase mostly for the back end, which is nice, because as a back end as a service, it helps me a lot with the things that I'm not... more of a foreign engineer with the ability to add a superbase backend rather than like a short backend engineer. But I found those services to be really helpful. All right, great. So, um, when you need to bring in another technical person for a project, what does that process look like? Do you sub out the work, refer them to someone else, or just scope around it? Uh, typically, I like to refer them to the person I'm working with, and that way, I don't like to mark up someone else's service. Um, so I will just introduce them and if they want. If they want something from that person, then we just start collaborating on that part of the project. For example, I brought in, on that same platform, a cybersecurity specialist because this is information that's user generated from minors. So it needs to be projected at a pretty high level compared to a normal web application. And so, I brought in someone that I had met through another networking event, who is a cyber security. He ran a cyber security company. And so he did a consultation for us. And a report, and then I fixed all the things in the report. Which ended, like, a month to the project, but it was worth it. That's great. Uh, all right, so what would make you confident enough in someone to actually refer a client their way or bring them onto a project? Probably, I mean, there's a level of trust. Like, B&I is a good example of having met someone several times, seeing examples of their work, is good. Having them help with something small before helping with something larger, that kind of thing, would be really helpful, because I believe someone, for the most part, when they say that they are good at X, Y, Z thing, but I also like to see examples of their words. Trust but verify. Yeah, trust but verify. Exactly. Love it. All right, so have you ever had a situation where a project would have gotten better if you had an infrastructure person involved at the start? Probably. Probably that big application I'm talking about. I learned a lot of that as I went. Tell me more about that. Yeah, so it's a, um, it is a platform where students can do homework that is social media themes. So I have a clone, a club. It's in Light Lake, because it's very, very scaled down clone of TikTok, Instagram, Facebook, DMs, direct messaging, what else? I mean, there's like there's 10 to 12, but we're using none of them right now. And so, like, a blog, and just all of the things where students can do homework, and instead of having, you know, write a 500 word essay on Romeo and Juliet, it's make a 3 minute podcast on Romeo and Juliet, or what would they DM each other, or make a TikTok? And instead of it having to be on TikTok on the internet, actually published for everyone, it's on our own little private system that's only visible. to that system. So just the teachers and the students can see it. That's interesting, because then you're leveraging the social pressure without the risk. Right. I love that. They love being able to make things, they don't do their homework, apparently. At least in this teacher school. They just, like, the kids don't really care about... Did you do homework? Yeah. Yeah, I mean, I loved school. I'm a nerd, but, you know, they just didn't want to do it. And so when she started saying, do this or that, then the kids lit up. And so then she wants to turn this into a business and sell it to, you know, 100s of schools. And so I've made it multi tenant, actually, like, several layers of multi tendency to make sure it's secure, and, but yes, they can make the homework, send it to the teacher, The teacher gets almost a social media experience themselves, and the student gets the experience of creating something in a safe environment. That's great. Um, I don't know, uh, is there any way that you can connect me with that person? Because I'd love to get a teachers perspective. Yeah? Yeah. Yeah, yeah. I can I'll talk to her. If she's... Because if she's working directly with students, I have a friend from, like, middle school who's a teacher, but she's a middle school teacher, and I'm looking for, like, um, more just not just my friend. You know, someone that's out there doing this kind of stuff and also doing things that are outside of the traditional education system. I think that that's amazing and I would like to pick her brain. Okay. Yeah, I'm sure she'd probably appreciate any advice you can give. Yeah, I mean, not just for, I just want, I, I, honestly, I just want to hear her perspective on how she's, uh, looking at conveying information more effectively in this day and age. Okay. That's really what I'm looking for. They're not really like engineering jobbers. Okay, yeah. Just more research, I guess. Gotcha. Okay, I'll ask her. Awesome. Cool. She'd probably be really excited because she's trying to get traction, but she's still working. So it's hard when she's not full time. Yeah. Yeah, yeah, it's hard. Try to balance it. That's why I had to do that too. Like, um, I basically had the idea to do my own business forever. Because I started out like freelance web stuff back in the day. But, um, I got W2 jobs and then like in 2023, I was like, I want to do this something full-time. I don't know exactly what. Right. So let me save up some money and when I get to my two-year runway, I'm going to do it. And then that's what I did. Yeah. So, you gotta have that cushion, because working and trying to do it at the same time is almost impossible. Yep. Yeah, yeah. All right, so we're at question eight. I still got time? I'm say we're nine, almost 10 minutes. Oh, I'm good. You just keep going. All right, awesome. Um, so, when your clients need infrastructure or backend work, what kind of budget range are they usually working with? Like, are they expecting that to cost 500 or 5 K or higher? Yeah, so, I mean, my projects typically range. I have done things as little as a thousand. I'd like to keep things $2,000 and up right now, so I found that $1,000 projects are really $2,000 projects. But my largest project is 35,000. And so I typically, I budget just so you know, $100 an hour, and so that's either estimates, or I just, at this point, I've switched over to charging hourly, $100 an hour for development, and 150 hour for consulting. Hey. So, yeah, but most of the time, it lands in a $2,000 to $6,000 range. Although, I want to increase that. my goals is to build, um, do more consulting for the small projects, for 5,000 under should be more consulting, and then, you know, larger projects, I still happy to take on as a developer. All right, cool. Well, we'll see. It's kind of whatever the market needs, right? Yeah, yeah. It's one of the market needs. And I think, like, two, I think you're on the right track. Like you're, it sounds like from what you're describing your business, you're, you're right in the tech wave, you're doing the right stuff. Um, it's just a matter of like, um, scaling your impact. Something like that? Like, like, trying to take on those larger jobs without draining your resources. Right. Yeah, that's what it sounds like to me at least. All right, so. All right, so we got that one. All right, so, uh, If you were going to bring in a specialist for infrastructure work on a project, what would you expect to pay them hourly project-based revenue share something else? Um, that's interesting. I mean, probably either hourly or project based, but again, I likely would want to at least refer my client to the infrastructure specialist. That way, you know, they are paying them directly, instead of paying me, and then I pay a portion of it, 'cause I'm happy to just work for my hours. I mean, probably would be a good idea to the agency model, where someone pays me, and then I contract the workout, but I don't have, I've never managed anyone, so that would be, um, a constraint of mine would be learning to do that part of the job. So I'd probably, like, refer someone directly, and then we would collaborate on the project being paid independently. All right, cool. Awesome. Uh.. All right. So this one's an interesting one, because I usually ask this to my clients, or people that have worked with me, that know me, like, in depth or whatever. But I would love to get your take on this question. If you had to describe what I do to one of your clients, what would you say? Hmm, I would say that you have two lines of business. One is business to... professional? I wouldn't even say business to consumer, but business to professional where you are coaching engineers in a couple different lines, helping them either self-specific problems, or coaching them with, say, resume or job interviews or social skills in ways that they can present themselves better. So, this is essentially professional support on one side, and then larger infrastructure consulting on the other side. Probably what I would have called it to someone. Which is great, because we could collaborate with that. Yeah, yeah. I think that's great. And that's, you got it. Good listener. Okay, cool. I'm just saying, really the purpose of that question, when I ask clients is I'm looking for the language that they're using to describe me so I can use it in my marketing. Gotcha. Because if I, like when I started out, I was like, mid-career engineer. No one really calls themselves a mid-career engineer. Right. Um, so I was just like, all right, how do you describe yourself and then how would you describe me so then I can like, communicate what I do so you can see if I'm even the right fit for you. So that's kind of the purpose of that question. So, and this one's kind of weird too, because I don't, because you kind of know what I offer at, like, a high level. But what's one thing you think I should be doing or offering that I'm not based on what you know about my background and what you've seen me talk about? That's interesting. I'm wondering, because, obviously, you help with someone like me who's, like, a developer. You could coach me and help me with a particular problem or with a job interview. You can also help businesses, like, in the room, but typically it's larger businesses that are gonna need your infrastructure services, I imagine. Maybe I'm wrong about that. A lot of the businesses in BNI are not large. They're, you know, more, like myself, or it's one person, or they have maybe five employees. They're not going to be 30, 50 employees. So I'm just wondering how you can communicate to the room, what you can do for them, in this room, who are more on the sole apreneur, to very small, small business versus, like, a medium sized enterprise. On the infrastructure side. Or on the coaching side, Michael, we were talking about, you know, training their employees to do something. Yeah. Now, that's a good point. That's pretty much the same kind of question I had for Lance at the visitors' day. I was like, you know, it's a really cool group and I like it. Uh, but I don't know if anyone could even refer me to anyone. Like, and he was kind of saying, you know, I could talk to Santiago, CIO, um, and he said, you know, he's looking for engineers that know Kubernetti's and I'm like, you know, I know it, but I also know people who know it. So it's like, that could be a cool thing. Right. And the other part, too, is, I could connect with people who are kind of recruiters, who are talking to engineers looking for jobs. True. And, uh, so that's a connection, too, if they know someone who's a recruiter out there. Um, You know, there's different ways of kind of getting, uh, my name out there so I can help people. Right. Um, but I like, I like the, what you're, uh, I like what you're saying here because it's like, all right, some, Some things could be tuned for larger businesses. Um, and seeing how that can fit in as like a consultancy. Right, wait, I've had to solve this problem myself with the group, because I, like, if I was building what I would really be excited to build, I would love to be building, like, like we were talking about with the Gramp rag. Like, I really want to get a graph rag project, because I'm fascinated with that, and I think it would be incredibly useful. I mean, forget what the name of it, but something just came out where someone was using claws to power generate graphs and then use that graph to create, like, multiple simultaneous simulations of situations. and essentially do, like, a Monte Carlo, but at a graph level, instead of one, you know, down a pointer, one row at a time. I like, that's fascinating. So I love that kind of stuff. But most of the people here, probably, wouldn't need something like that. Yeah, yeah. And it's like, well, like, I would love to make a brack rag of someone's entire company with every stakeholder, every employee, all of their skills, it records every email, every text, everything, because then if someone leaves, a key person risk, they have, you know, knowledge or apostasies that maybe aren't written down. It is actually documented. The AI knows it now. Yeah, yeah. Like, that would be very useful, but not for a two person company. It's a 50 person company. I hear. I'm totally, like, the stuff I want to work on with that stuff is, like, distributed. systems. Yeah. All about distributed system, consensus algorithms, stuff like that is super fascinating to me. But I only really got to touch that stuff at American Express. Right. And it's like, and even then, people were like, do we really need to do this? And I'm like, please. Please let me do it. Let me do this. This is cool. I like this. But like, um, but at the end of the day, like the worst project I was on where it was like, you know, it's fun. I love it, but cost benefit analysis shows it really doesn't justify the engineering cost, so we got to cut it. Right. And then get moved on to another project where it can justify the cost. So the business pressure of, you know, what an engineer actually wants to work on versus what the market will bear. is like the tension point that we have to deal with. Indeed. It really, it's frustrating, but, um, you know, I, I think, like, the best way to do it is, like, maybe just, like, free time, you're doing that project. And then maybe it can turn into a product that you sell later. I'm thinking about doing it for this group, because everyone, every single time, if we go around, someone has to be introduced to a property manager, like, by default. It's like, I wonder what we could do with something like that with all of the, all the businesses in Florida, all the businesses in one city here in Florida, just something that's like, a private tomb, this group, index that is, you know, graphical and rag, so that they could talk to an agent. Like, you know, I have this business, who would be the most useful, like, who do I need to know? And it would be like, there are 18 people in your region that would be good to know, and here's how they connect to each other, and to you, and here's what you can ask them, and just like, Yeah. I was thinking that same thing when you were when you were showing off your AI bot thing? Yeah. Um, I was thinking, you know what would be really helpful in that report is 2nd and 3rd connections. Yep. And I have like a version of that in mind where it kind of, it does a search and it sees if there's anyone like related to that person. Right. Um, and stuff like that. But nothing like what you're describing where it's like this huge thing where it's connecting all these dots or whatever. It's just very, you know, janky, right? Like multi-hop reasoning. Yeah, yeah. anything like that. But, um, but yeah, having those those connections that that are there. but not a parent. Right. And also, not only the connection, like surfacing, the connections is useful. But identifying the patterns between those connections... And then extracting that pattern, could be really useful from, like, a business intelligence perspective. Yes. Um, and, you know, LMs are really good at patterns, finding patterns. Right, right. So that's a really good application for that. I think I'm gonna build that just for funsies. Yeah, do it. Just like build a thing and then, you know, have fun with it and then you'll probably end up turning into a product or something and then you can like have that as a service offering. You'd be like, hey, you know, I have an AI, uh, I don't know, chief of staff or something. I don't know, something like that. Like, that can just... That can, like, do all the stuff that you're talking about. Yeah, that's true. Uh... All right, so I guess, uh, question 12. Uh, What would a good 1st collaboration look like to you? Something small where we can test how we work together before committing to anything bigger? Probably something like that, yeah. I mean, again, it's... Especially if you join the group, we were gonna get to know each other. So the trust thing will build and build and build. But yeah, I mean, even just things like this is a good way of getting to know you a little bit better. Like, I think it'd be fun to have, like, a regular meeting after one of these, like, every few weeks or something, and just nerd out about whatever's going on. If you're into that kind of thing. Oh, totally. Uh, there's a Michelle Baklos, I don't know if you know her. uh, like tech connector. She started as a react front end developer, and now she's like a director of GSI or something, I don't remember her company. But, um, I met her a long time ago, and, uh, she hosts this... thing called You Should Know, or something like that. And it's, um... As you can tell, I'm not really good at names, but ASL was my 1st language, so I think it gestures. So the words are kind of the same. I see. But, um, the whole point is you show up as you're an engineer or developer in the community and you show up and there's these lightning talks for 5 minutes and they time you. And you just talk about nothing pitching, nothing selling. Just what's a cool technology you're working with? What's something you built? What's something you want to show off that other nerds would appreciate? And it's really fun and she does it every month. So, um, I'm not saying adopt the lightning talk format, but like, I enjoy that meeting. So if there's something like here, if we come back and people are just kind of sharing cool things that they're working on or things that they discovered or things that they read about. Yeah. I'd be into that. I'd be into that, too. Dan, did you met Dan briefly, right? I'm bad with names. What does he look like? He's wearing a black shirt, glasses, about my height. Uh, he was the one I was talking about going to lunch with. Oh, yeah, yeah, yeah. Yeah, so he wants to start a lunch club. That would be like, that could fit into that. And just like, what are you working on that's cool? And it's not meant to sell to the other people in the lunch club. It's just... Just sharing information. and learning. I think that would be amazing. Yeah. Yeah. But, um, yeah, I'm I'm pretty much on I'm pretty much joining, even though, like, um, like kind of we talked about earlier, the, I don't know if the referral thing is gonna super work out for my particular business, but, um, you know, I think I can get value in other ways. And, you know, also, I think that the people that I'm working with, because I'm, like, doing the, my, my 1st deal that I close B to B is AI Assistant Development, training, or an India team. Okay. And, uh, so I think, like, because I'm talking to these founders, who also need services. Right. I think that I could kind of get... Go back through. You know, and I think it could work out that way. Right. Oh, yeah, another thing I would like your information on, since you're doing like the automation stuff. On my website, I just put this week a free AI guide. Yep. Um, and if you click the little thing, it goes to the thing, and you put it in your email, and you could get a, um, PDF. It's 10 page PDF. Oh nice. And it has like an overview of like the AI landscape and some models and like how you can apply it to your team and stuff like that. And then like a little thing at the end is like, if you want to work with me, here you go. If you can take a look at that. Yeah, it looks like I can look at it. I would love your feedback on anything I can make it better. Yeah, let me see. It's learning with OJ, right? Yeah, learn with OJ.com. OJ. Sorry. No, you just gave me an idea. I should buy learning with OJ. and have it redirect. Yeah. That's why these meetings are super valuable. And I have, like, so many domain names. I have nothing on it. Yeah, buy them. So where is the, um, the little widget? If you screw up to the top, see the little star? Oh, the AI free guide. I want to do my work email. Cool. Let's take a look. Yeah, and I think you'll get an email, and then you can download it. Cool. Hopefully it works good. I mean, think you're probably, like, the first real person to test it. Cool. Oh, this is sick. I worked really well. Nice. Yeah, I will read this today. Yeah, awesome. Yeah, any feedback you can give, um, you know, tear it apart. Like, you're not gonna hurt my feelings. And that's one thing that one of my clients said, too, was one of these questions, the client questions are a little different, but I was like, um, you know, It was something like, um... Uh, was there anything that almost stopped you from reaching out? And he's like, no hesitation. I saw how you solve problems, blah, blah, blah. And I knew you didn't give a shit about hurting my feelings. And I'm like, yeah, because it's not about personal attacks here. And it's, and if you deliver information, like, feedback in a way that's not, like, oh, you're stupid, you know, like, because you're not. You're literally just figuring it out as you go, like, everyone in life. Right. Um, and if you just have a little bit of respect for another person, it doesn't cost anything. You can, like, deliver feedback that's less polite, I guess. I think it's that most developer types, like truth over feelings. Yeah. Like, I'd rather, like, if I had something in my teeth or, like, something that, like, you know, I would rather you tell me than not tell me. I would rather your, if I have a logical error, I love when people are like, No, you're incorrect about this, and here's, and then here's why, and can change my mind. Doesn't mean, I'm not gonna change my mind unless you actually prove it to me. Yeah. But I'd rather have... Don't tell me. tell me what you're doing. Don't just tell me, but if you can prove it to me, then I will happily change my opinion, because I'd rather know what's true than what I, you know, be mistaken. Yeah, I think that's great. I think most developers, these the ones I know, seem to be kind of like that. Yeah, I think that, and I think that's also why, um, you don't see a, a lot of people, developers like super zazed about sales, because it's kind of like sales is like, well, you kind of have to maybe live by omission, some salespeople do. But I like to do the approach of just like, like you said, like, you listen to the problem, you go, all right, here's what I offer, and here's who the people I know offer, you know, take it or leave it. Right. Or I'll learn about it, and we'll circle back around, see if I can actually help with this one. Yeah, yeah, yeah. Yeah. Yeah, and you can, like, kind of springboard from there as opposed to, like, the sales, like, car salesman thing. Right. Or they're, like, trying to, like, sell you a lemon, and you're like, come on, man, like... Right. You know, I get that's your job as selling lemons, but stop doing that job. No, stop it. Yeah, yeah, yeah. Just stop. I was in sales for a long time, and that was the first job I had out of college all the way through. I was, I think I'm... Yeah, well, honestly, like it is, but it's like, if you can just be honest with someone and if you can genuinely say, you know what, I don't think this is a fit for you, you really need to go here. Yeah. the miracle on 34th Street playbook, like this. We don't have it, but go here. If you can genuinely do that, then when you do say, you know what, we do have this, I think you're a perfect fit, then it's genuine. Yeah, it lands, yeah. And then I don't think sales is very difficult. It's just about being honest and helping people solve their problems. Yeah, well, you're one of the good ones. Well, I tried to be. Tried to be. The 1st bank I was at was not like that. The 2nd bank I worked for was exactly like that and it was a good fit. Yeah, that's what I'm saying. Like, it's it's not common. Um, unfortunately. Uh, but I think, you know, it's good. I really like that you're that you're doing your business and you're doing the right way. Thanks. Yeah. I mean, I'm new. I have a lot to learn. didn't realize you started so recently. That's incredible. I'm, like, super, super new. Uh, newbie, but, uh, I've been, you know, kind of doing, like, freelance stuff the whole time, but as a casual thing. Yeah. And not really focusing on marketing and, you know, stuff like that. Um, but now I'm like all in on the business thing. It's the full-time thing, and, uh, it's really fun. I mean it's a whole different world. Yeah. And I have to really resist the tool, the pull to build tools all day. Right. Because I'm like, ooh, I don't have someone putting a deadline on me. I could just literally build whatever I want. And I'm like, this could be useful for my business. I could justify doing this. But in reality, I have to be, uh, out there talking to people, you know, putting my name out there and stuff like that as opposed to being there, building tools or working with teams building tools. But I'm trying to, like, shoehorn in there, like, part of the job is me being on a team building stuff. Right. Or showing people how to build stuff. Right. Um, I think that that would be, if I could do that, the majority of the time, and then kind of expand to where I'm like, I have a staff that can do the other stuff. Right. That would be great. That would be great. Um, but I don't know. That's like far away. That's like a super dream. I don't even know if that's even gonna happen. Would you work towards it? Yeah, that's that's the kind of, I guess the big picture vision is like a Con Academy for Engineers. Let me be sick. That's what I would love. I would love that. I would love to just have a free resource with maybe like a paid, like upgrade thing. Yeah. Um, to just go there and just know that you're gonna get something that isn't gonna waste your time. Yeah. It's not gonna be full of a bunch of boring corporate jargon. Maybe even have the platform be like sold to companies to do their training internally. So that when you're watching these stupid ass videos, and you're, like, at this company, and they're talking about stuff, you don't even care about. Instead, it's teaching you the stuff that you want to do, and it's interactive, and you actually learn, and you're working on maybe, like, kind of the ideas are, like, from the infrastructure side, it would be really nice to have real world labs. Yeah. Not, here's a clearly, uh, fake scenario that's super easy to solve. Throw them into the fire. Build up to it. But, like, give them at least that end goal of, like, you log in and you literally don't have a guy. You just log in and then you have to kind of figure out what's the problem. You just see the thing breaking. And then you're like, all right, I got nowhere to go. And that would give you a lot of real world experience right away. I sure would, especially if you don't have access to things like, you know, go to the console, copy the error, sit at the ChatGPT. What is this? Like, if you don't have that, it would really require you to be like, oh my God, I'm going to, like, look at the documentation. Yeah. No, it seems like nobody likes to look at the documentation. No one, no one, not only do they not like to look at it, they don't like to write it. Not bad. That's the hardest thing I had on teams is, like, come on, man, get the stuff out of your head, get it into a dock so someone can, like, update the read me at least. You call me out. I'm just saying. My big project, I don't have a single thing written. I have the cybersecurity response report written, but other than that, It hard. You have to decide, am I going to spend my time building the thing or am I going to write about building the thing? Well, it's also, and not to say it's client pressure. My client's very sweet, but, like, if I'm, like, I spent all day to day working on documentation. Well, what about the feature I want? What are you doing? Like, she likes, like, even if I'm working in the database, she, she, I can, she's never mean to me, but, like, I can definitely, like, can tell that's, like, where's the visual update? She's a very visual person, so... Yeah, you need to show the result. I always have to do something on the front end to at least allow her to see what I did on the back end, because she can't see it. Yeah, I know, and it's like, how do you how do you communicate the value of having well-written documentation? Right. Like, it's almost like a hope. Like, you're like, okay, I spent an hour writing this dock. What if the project is scrapped tomorrow? Right. You know, what if we don't even want this feature? Well, what if I die? So I just, yeah, that's what I'm saying. or you won the lotto and you're no longer working. Like, either way, like, you're no longer maintaining this thing, and and maybe the documentation helps another person coming in, but it doesn't help if that thing isn't there anymore. Right. So if you're building things that are kind of moving fast and being cut, the documentation doesn't really make sense. Um, But... You need... It's like you said, no one wants to read the docs, but if the dogs aren't there, there's nothing there to read. Right. So it's like a chicken in the egg. Yeah. It'd be nice if it, like, cursor or anything where you're using a gentic coating, would also write the documentation and then reference it, and then update it, or, you know. That what I do with my prompts. That's what I do. Okay, that's great. All my prompts have, like, a little thing in it that says, you know, build this part of the feature, because I break my features down into smaller prompts so I can check in between. And then I'm like, do this part of the feature and check the dogs. Like, there's clod.md or some platforms have agent, MD, you have skills, MD. There's all these like markdown files. It's like this big joke of like, if you look at open source projects now, it's all marked down files in like one script. Yeah, yeah, honestly, it's like a super joke. Yeah, and that's one of the reasons I want to help people with Claude specifically, 'cause it's like, well, the skills MD files is what they really need, and they're not going to be able to do that very well. Like, yeah. Yeah. Yeah, yeah. And if you know how to prompt it, you can get it to write it well. Yes. So you can kind of, like, feed it in, but you gotta be the architect of it. You're driving it. You can't just be like, I need a caught MD and then have it like, you know, improv it or whatever. You gotta give it the you gotta give it the constraints. Um, but yeah, if you, you know, the, your idea of like doing the, um, documentation as you're building, you know, that's something that I was doing with my prompts, but it would be great if you didn't have to prompt that. If it could just be like, I already know that I need to update dogs as I'm building. Right. I already know I need to check Dodge as I'm building. Right. That would be cool if it was just baked in. as part of the agent. Well, this is, and this sounds like, I actually, I run a lot of this stuff through, like, 3 or 4 different AI tools. I go through, like, Groc, because it'll be different than open AM. Yeah, yeah, all the models are different. And so I like to get, like, a consensus, but I was, like, how could we, and I'm nerding out? Like, who would it be if I could basically take the entire code base every time I'm, you know, finished working on something and put it into a vector database so that the AI can then reverence the entire code base more efficiently than just grabbing? Yeah, yeah, yeah. In real time. And then it could create documentation for me and all that. I was like, that would be so cool. And it's like, that is so redundant that it's like not worth it, but I would. But that would be cool. It would be cool if it could just, like, know. Like, instead of like piece by piece going through it, it could just have that big picture view and then navigate it. That would be awesome. Like everything I've ever said to it should, you know, instead of just being in the context, like a rolling context window, it should be vectorized. I would love that And then, you know, a non-technical person could go to their admin panel, like their page, and be like, what's going on with this part of the project? Well, even we've done this and, you know, March 28th, last year and did this and this and this and this. And here's a documentation, and just, they could talk about itself. Yeah, yeah. Not really self-aware, but... closer. Yeah, closer to that. No, I think that's amazing. And I also, like, just to kind of get meta a little bit, I really like, I'm kind of detecting, like, I like how you're bridging the gap between tech and non-tech. I think that's awesome. And I feel like that's because of your background. Yep. And I think that's cool because you're like bringing, like, you know that pain point. You know exactly where that is. being non-tech getting into tech, um, and then I think that's super valuable too. And, um, you know, a lot of times in engineering spaces, it's like, no, I've been coding since I was, like, a kid. You want Atari or whatever. And you're like, all right, come on, man. Like, that's fine. You know that's cool. It's cool you were nerding out at, like, you know, when you were a total, like, toddler. Yeah. But like, not everyone comes from that background and their opinion in the in the meeting is just as valid as yours. Like, just because you had, like, you were good on Atari, maybe this person who came in later is super great at this new technology. Like, listen to them. Because that would be like what I would end up doing is like running, you know, interference on people that were, like, shutting other people down. And I'm like, no, anti gatekeeping. Like, we gotta, like, they have a valid opinion too, on this tech, and they should be able to contribute. Yeah. Um, so I think that that's, I think that's cool that, that's where you're coming from. Yeah, yeah. Well, I mean, I started at 22 and I had done some hypon programming. I was like 12, but like for a day, and then I, then I was like, I'm gonna go outside. Yeah, I'm going to actually live. I picked it up 10 years later, but it's so much more interesting than banking was. Yeah, yeah. The finance world is like weird. It is, I love finance, and I love talking to people, but I hated the job. So other than those two parts of the job. So loving keeping up with, like, what's going on in the markets, but I wasn't really being paid for that. I was paid to, like, go get relationships. Customer kind of, and customer service. So it was relationships, and that's how I got paid, and then customer service was, was the, about the job. And it's like, well, this isn't really, you know, I'd rather be building something than sitting here at a desk. Yeah. Yeah. Dealing with other people. Well, I like people. just that I don't like, I didn't like being tied to a desk, because I didn't sit there and wait for someone to come in. Yeah, I had to, you know, yeah, I can talk about it, but it was... Yeah. I hear you. I'm kind of not built that way either. I couldn't do a banking job. Um, I couldn't do a food service job. Like I got fired immediately from my 1st real job at like Burger King as a teenager. Oh, yeah? Yeah, I was like, apparently, I kept like giving people refunds on their food. And they're like, you can't keep giving away free food. And I'm like, the motto is have it your way. Like, come on, man. I'm literally doing the motto. Right. And they're like, if you do it again, you know, you're out of here. I'm like, another person came up with their order being wrong, and I'm like, I can't help that the cook is, like, effing up back there. Right. I'm gonna give them a refund, and then they fired me. And I was like, All right, whatever. I don't even care. That's all right. I'm gonna work for this pet store. And then I, like, work for the pet store, and then I did the freelance web dev at the same time. So that was really cool. Yeah. Um, yeah, just like more wired for, it is much more valuable anyway, or the web stuff than, yeah, Burger King. So like, you're just wired for that. Yeah, and that exactly. Like, it was like, I looked at my my one paycheck, my half of a paycheck at the Burger King and I was like, all right, that's like nothing. He's like, I think it was 525 an hour. Oh my God. It was like minimum wage. Right. So, and then my first, um, actual sale was 50 bucks, and it took me, like, 30 minutes to make. Right. And I'm like, dude, like, I can make a lot of money doing this, and it's fun. Right. You know, it's, like, really fun. I got a lot of nose, though. That was not so fun. Like going around and being like, hey, I can build your website live coding it. They're like, that's cool, but, you know, nah, we're not gonna do it. Um, and then I got the first sale, and I was like, yes. And I kept them at that rate for like years. Yep. And, you know, because I feel like, you know, you took a chance on me, so you get to pay 50 bucks instead of, like, you know, 100s or whatever. Um, and then 1000s later. But like, uh, yeah, it's pretty cool how things have evolved. It is. You know? Yeah. And I totally get the whole, like, Yeah, I wanna work with the technology and think about interesting problems, like, you know, craft rag and stuff like that. Yeah. As an agent systems, like what you were talking about. Yeah, multi-agent systems in the handoff. That was a big friction point that I ran into. was like, 0 my God, my agent is like super behind or whatever. Um, But you get smooth it out and stuff. And then there's also these cool things with like memory graphs. I don't know what memory graph is. It's like a, because one of the big problems with AI is, uh, there's a lack of continuity, and they're trying to solve it with memory. Like, a clod release a memory feature for Bree now. It used to be just for paid people. Okay. And it's allowed, it allows you to, uh, allows you in one context to for the, uh, AI to reference the other chats. Yep. and be able to pull things in effectively. But it's super janky. And sometimes it'll be like, I don't even know what you're talking about. you're like, come on, man. I literally gave you the URL of the chat, or I gave you this thing that you should be able to reference. Yep. So it's not really good. So, memory graphs are one way that I was looking at these research papers of being able to give continuity to AI agents. This is, like, a neo for J graph? It's like, uh, yeah, yeah. Because this is what I'm talking about with graph rag. I think we're talking about something similar. It's very similar. Um... Uh, distributed, um... Let me see. I can't see it, but it's fine. I can email you. Yeah. some cool stuff I was looking at. But, um, it basically there's like 3 There's 3 main things that this researcher team found that were really effective in handling memory at a better rate than what is currently out there. And there's ones like, um, orchestrator, where you have a centralized orchestrator that does stuff, but it has, you know, certain things that it sucks at. There's attributed one, and then there's a third one. don't remember. But, um, the point is you could pick one of the 3 or maybe you could do like a weird blend with like different things. And you can kind of get the best of all worlds. Yep. But they kind of just tested it, and they showed you, and, you know, they didn't prescribe, like, this is the best way. They're just like, if you go down this path, it's more effective, and these are the three ways that you can implement it. Yep. And it was really interesting to me. So I was like, All right, let me try to, like, implement something like that. So that's like one of my side projects is like trying to build something like that on my little thing and see him and then test in, um, like, clawed or Gemini or whatever. I usually use Claude Gemini and ChatGPT or open AI stuff. or Kodaks and Turmoil thing. I've heard codex is really good. It's way better than chat GPT. Yes. Not as good as Claude. Right. He is still the superstar. 4.6 as well. Do this, like, super great. It's so expensive. It really is. I used to, you know, cursor used to be 20 bucks a month. And it's just like, this is great. I can do unlimited amounts of work, and now it's like, every hour is like 2 or 3 bucks, which is good if I'm getting paid. I have a few projects where I'm doing a revenue share, and until they're making revenue, I'm not getting paid. Yeah. And so those are like. You're like draining some money. Yeah, I know. I know. And then you get the daily and weekly limits. Oh, my God. Uh, but you can go between Sonnet and Opus. Yep. You know? you know, high reasoning, task, opus, and then the rest is on it. Yeah. And that helps with cost. And I used to use cursor a lot, too, because it has like auto and it kind of kind of pick. And it keeps us down. like auto anymore. Like, I don't I haven't used it in like a bunch of a bunch of months. But I used to use it a lot. Uh, in Uh, 2024, 2025, early 2025. I was using it a lot because my job kind of like had it. And I was like, yeah, I like the dips and the interface is cool. VS code, whatever. But I haven't really used it for my projects. I literally just have like my IDE and then I have plot in in a terminal. What IDE are you using? Um, I use uh, neo them. Okay, I've not heard of it. It's cool. It's like a lot of uh, and then I use VS code too. Okay, yeah. Classic. I kind of yeah, it's really good. You can't, and there's a bunch of extensions and it's really cool. You can really make it your own. But neo them, I like because I had a bunch of stuff. It's like Bim and Bim is like this kind of like command line. Okay. ID. you can set up these like macro kind of things. And you can do a lot of stuff where you're literally just on the keyboard. You don't need a mouse. Yep. And you can do things in, like, one key. So you could go really fast with it. I like that. Um, but there's a little bit of a learning curve, and some things are like, you know, I'm just gonna, I don't have the extensions or the willingness to build something in my vim environment. Right. I'm just gonna use VS code and get the extension and just run. Right. Uh, so that's why IBS code on the side. Yep. But, yeah, what do you use? I use pretty much cursor. I do like Claude code, but sometimes I like being able to see everything in the BS code. I grew up on BS code. All my like demo projects, all the YouTube tutorials in BS code. So it's like that, that's where my brain works. It's the dips, right? Right. I love the dips. I love the dips of cursor too. The dipping. And so I will sometimes open up clot code in the terminal inside of cursor, but you can do it in BS code. Yeah, yeah. And just do the terminal stuff in the terminal, but then being able to see the architecture and everything and the dipping. do that too. Yeah. Yeah, I like you go up terminal and then terminal's on the bottom or you can move it around, and then you just do pod, and then I also have different profiles for my Clyde. You can set up, I think it's Claud... Shit. I don't remember, but it's like pod Project Dirt, some environment variable, and you can set up like your, if you use ZHS, or bash or whatever, you can. I just did go into bash. I'm like, damn it. Yeah, you got to go into the profile and you can set it up to where it, you can have different pod profiles. So I have like a work one and I have like personal ones and I have whatever ones. Yeah. And then I can just do Claude one or Claude 2. And then it launches a little thing. And you could also do things like, I learned pretty recently, and it's, like, pretty embarrassing, 'cause it was, like, Claude came out in, uh, pie day, 2023. Did it really? Yeah. Wow. And I didn't join that early. I was like in 2024. And it was only because someone at my work was like, Claude is amazing. And I was like, all right, let me check it out. Because I was on ChatGPT. Right. Um, and like you, I was kind of into the whole machine learning thing. Yeah. We were doing Python... I hosted one of my coworkers at the time in 2018. Uh, I was working on a mid-level, uh, Telecommunications construction company? Yep. So we were building apps that were like, where the internet wasn't. Gotcha. So we had to like build it weird. And it was really fun learning. Um, but this UI, UX designer lady, Silky, which is an awesome name. I like Silky Smooth. And, um... I know, it's great. And Silky Lopez. And she wanted to do this women's coating coffee event. Yep. And we got the CIO to super support it. And, like, it was a budget and all this stuff, we had catering from Chipotle. And we just invited everyone, not just women, but it was like, you know, like women in tech and everyone. Yeah. And I was the first person to do a workshop, and I did a 3 day workshop, so every Saturday, 3 days. Perusing Python. She named it. And, uh, I was just like, yeah, let's start out. You don't know anything about Python. And then by day three, you're building a chat bot. Right. In 2018. Yep. And because I was super into it. I was like, this is amazing. You know, machine learning, you could do machine learning with Python. Right. So that was really fun. NLP. Yeah, yeah. And, uh, you know, the bat we built was nothing like what we had now. Right, it's like super, super primitive. But it was cool. Like, I like to see, I like how everything went. Yeah, no idea. I did something similar in, um, this had to be like 2021. So this is like right before Chat GPT came out. A buddy of mine was like, oh, you know what we should do? We should use, because I had just taken like a course on natural language processing. Yeah, we should try to get it so we, like, you can input like a book, the text of a book, and it will find all the nouns and extract them. And if it's like, this is a instant of that, it would create a class of that, and then put the thing as one of the instances. and basically have, like, working code as the output. And I was doing it, it wasn't using any sort of... No, it was just like, yeah, it was literally like the jankiest thing, but it would work. Yeah that's cool. And I was like, this is the future. And then, you know, that GVD comes down. I was like, It is. It's the future. Yeah. It's happening. But I think it's cool. And I'm really excited to see where it goes. I mean, there's a lot of doomers that, oh, it's all the stuff. But, you know, I'm an optimist, I guess. I mean, I like to say I'm a realist, but really I'm an optimist. like a, there's this, uh, uh, I don't know if you read, like, fiction or anything like that. Yeah, I love sci-fi and stuff. We'll have to talk about that at some point. Yeah, but there's this thing called Hope Punk. Okay. And it's the opposite of grim dark. Yeah. And, uh, the whole concept boiled down to a sentence is the glasses have full. Fuck you. Yes, I love that. I love that. And I'm like, yeah, that's how I feel. I needed, like, what book would be, like, a good representation of that? Ah, dude, uh... Let me, let me look at my notes. Let me look my notes. Uh... I love this sort of, I want to say I agree with everything, but, like, the A 16 Z, like, techno optimist perspective of, like, I think technology's good, and it makes things better, and I love it. Well, I don't see it, but maybe it's on my phone. Give me luck. Hmm... Oh, yeah, Becky Chambers. Yeah. Um... Let me look up her book that I read recently. Becky Chain Chambers. I think she has a podcast too. Mm hmm. She's really good. She's, like, award winner. She's really great. I haven't read a lot of her books, but I, I, if you want to know Hope Punk, I think she's like one of the, she's one of the Hope Punk people. like it. Um, what's her book? That's really good. Uh, a long way to a small, angry planet. Okay. Okay. So that's, like, a good one. Um, and there's probably others, too, that are really good. I can look, but that's the one that came up in my little search thing. Cool. Um, in my history, my recent history. So, I think that would be a really good one. And I was like, yeah, this is a cool book. I like it. And, um, and I like to write too. I don't really have, like, I'm not good at it, but I still like to. Yep, you know. I'm that way. Like, I would enjoy it. I just, yeah, have to do it. So it's not my thing. Yeah. Yeah, I did creative writing in college to get a credit. Yeah. And I met these poets, and, uh, they were just so wild. Like, they were just, like, so wacky and weird, and I just loved hanging out with them. So, and it was so different from like hanging out with the tech people. Right. Or, like, I kind of I kind of grew up with, like, so many different kinds of people meeting, uh, different aesthetics. Yeah. Like, uh, it was like nerdy. And then I, like, rebelled as a teenager. We went like super goth and edgy. And then I went, you know, back to kind of nerd and then, you know, more professional, you know, and kind of grew up. So, um... But the poets really, like, made an impression on me, and they taught me how to write poetry. And I was like, wow, this is cool. And then I was writing, like, little fiction, short stories, and it was cool because, like, you be like, I'm trying to find a store, like, I have this idea. I wanna read. And then you can kind of just write it. And then you satisfied that weird thing. So you're like, I don't know, it's like a weird thing that you're doing. like a plot outline for a book that I want to exist that doesn't. And I was like, okay, they're gonna scratch my edge. Nice, nice. Yeah, yeah, I would do it, like, even if it, like, I do it, even though I'm bad at it. And I'm starting to put my stuff out there. I have like a little substack and I get like a couple readers at like my stuff. Um, like I did, uh, I did like this story about hacker and an embodied AI. And they go through this stuff together. And it's like a year journal of what they're doing. And that was really fun. I got some, like, readers that actually subscribe from that. So I was like, ooh, nice. But, you know, I'm not like, I'm not like putting out novels or anything like real. These are just fun little projects I do on the side. Yeah, it's fun though. And addition to, like, building tech. I wish there was like, I wish you didn't have to sleep. Yeah. I always I always say, I want a 40 hour day. You know, like, eight to 16 hours of work, but then I also want, like, 10 hours of free time and end up asleep that I don't have to set an alarm. Yeah, yeah. I would, if I didn't have to sleep, I would be like, I could just cut that. Like dreams are cool and I like like the feeling of waking up and like that comfortable of like, but I'll trade that for more time. I love sleep. I sleep like 9 hours a night. Nice. would love to not eat it. Yeah, that's why that's where I'm at. That's where I'm at. It would be nice if it wasn't a requirement. I could, like, be like, I want to do a lazy Sunday and have a nap. You know, it'd be nice if it was just an option instead of a obligation. Right, yeah, 100%. But, yeah, so, I mean, uh, I guess we're like an hour, like 50 something minutes. end up being an hour. Yeah, yeah. one is great. But I really appreciate your time and nerding out here. I got a lot of good information and just, like, hearing about your business and your thought process and just knowing, like, you're one of the good ones out there doing stuff. Yeah. It just makes me feel really good. Yeah, likewise, I hope you join the group. I really... Yeah, yeah, I'm gonna I'm gonna... I think, I don't know if Lance is the right person to message, but I'm gonna check my email. I think someone messaged me about membership. It was either Lance or Bryan. Yeah, yes. Someone emailed me. So, um, probably before my 3 PM, I'll go on there and like sign up. And then, uh, and then see how that goes. Because I'm in. Like, I was in the 1st, uh, visitors day. But the only question was like, is the referral thing even going to work for me? Right. But even if it doesn't, you know, what is it, like 3 grand or something like that? Two? Okay, so it's like 2 grand and I already got like 2 grand worth of stuff for my business right now. So it's like, you know, awesome. Yeah. I honestly, like, what are you trying to do? Like, if I sent you a referral, would you prefer another engineer or would you prefer, like, a business that needs infrastructure? Like, how, you know, what's your... It's so hard. They asked me that yesterday. and I'm like, dude, I'm on both. Well, I'll just put it this way. The engineer is the long term. Yep. The business is the short term. Gotcha. The business is a higher ticket item. Yep. And I could do it. Yep. I can, like, help them do their thing and get them out the door. And get the money and then use that the funnel my dream, which is helping engineers. Right. And then the engineer is less money, but it gives me data. Yep. And it's the thing I actually want to be doing. So that's kind of, you know, I guess you could make that determination who you want to send me. Yeah, both. If you come across both, I'm totally down. Infrastructure, I kind of specialize in cloud migrations. I'm big in the cloud. I jumped on that as soon as that came out. Um, and I know the major platforms, like, uh, Azure, I hate Azure, but, you know, Microsoft is in every major enterprise. They're like, ooh, it's Microsoft. I gonna jump on it. Yep. But then anyone who gets their hands dirty knows, like, Azure sucks. Like, it's not good. Um, but I can do azure. AWS is great. I love AWS. And Google Cloud is cool too. It just doesn't have as much market share as AWS. And it's a lot easier to, like, get people on AWS because everyone's kind of on it. Yep. So, um... the popular stack wins because it's popular. Yeah, it's like a feedback. So I kind of like, if I'm positioning myself, it's EWS. But I know the other two. Cool. Because I can't do that. at all. So that'd be great. Yeah, yeah. If you ever have any... I mean, even no charge. I'll just help. just help. Like, um, if you have any, like, AWS, like, specific questions or whatever, uh, and I would say, like, the boundary line is, like, uh, like, if you want me to actually, like, build a thing or sit with you and build a thing, then I gotta charge for that. But if it's like, hey, I got a question, and I just need a... the right direction or I need like, have you seen this before kind of stuff? That's no child. I'll, you know, I don't care. I'll even do a little bit of research right there and like be like, hey, this is what I think. But yeah, that's kind of like, that would help me out a lot. Right, yeah. And then from your side, it sounds like, you know, if I'm interacting with these founders and stuff and they need like custom solutions, AI automation, stuff like that, I'm training their engineers, but they also want something bigger. I mean, I know your guy, Ethan. Right here. Like, even he can do it for you. And I can like funnel the newer way that way. I appreciate it. Yeah, it's kind of been the table of two product lines. You're kind of like yourself. I have the big ones, or I have, you know, something more like a training. I think that the small websites have probably won't do that much longer. Just 'cause it's, like, the software really goes, which is pretty much larger. Yeah. Using it for a website, their performance really well, but... And I think it's smart to pivot out of that because we're getting into a world where everything is bespoke. Right. And why, why would you need someone to come in and build a smaller thing when I could vibe code it? essentially. And this is essentially what I want to teach people, except I don't want them to get in the weeds of, like, this cursor, and here's how these files were, because they don't like that. Just, like, use clod, and they'll do it. It's called artifacts and things like that. That would be really cool, 'cause if you can make it so they're developers without the ID tooling, that would be cool. Well, I think then Claude is moving the direction. Yeah. There's a lot of, like, I just learned about Colorado effects the other day. I was like, you can just ask it, hey, build a website for this, and like, you know, represent this data visually in an interactive way, and be like, done. I was like, well, I used to be able to do that. That used to be something I would have needed to do. Yeah. And you could do it by talking to it. Right. You can you can just be like, hey, I need this thing. And you can even go in the voice mode and it talks to you back. It's pretty crazy. I do, like, I'm like talking to it. I have a standing desk, and what are my, I don't do this all the time, but I have the, uh, the headphones that have the microphone, and so I'll just, like, talk to it, and I'll, like, pace around in the office, and it'll be, like, build this, do this. Yeah. Yeah. Yeah, I think it's great. But, yeah, so thank you so much. I really appreciate it. Yeah, anytime. And honestly, hopefully we can do this more often once you've joined.

20260318 Discovery Conversation 03 Ethan With Red River Web Design