Episode #6 - The 2-Person Startup Revolution: AI, Agents, and Human Vision w/Bradley Friemel

Full Episode Transcript

This is the full recorded conversation from The Human Protocol Podcast, published by The Human Protocol Group. © 2026 The Human Protocol Group. Please link to and attribute the original episode when referencing this conversation.

THE HUMAN PROTOCOL PODCAST (00:02.612)
Hello, everyone. Welcome to another episode of The Human Protocol. You may ask, why now? We are at a real turning point where companies write algorithms, governments write law, and someone has to write the human story. And this is what this space is and will be about. Think about what if your startup team or company was just you and

a dozen of AI agents, for example. This really isn't science fiction anymore. It is the new blueprint for startups today and moving forward. So this is the Human Protocol, the podcast about staying human while building the future with AI. I am your host, Mykel Salomon. Each week, we explore the intersection of technology, leadership, and purpose.

on covering how values impact and endure in a world rewritten by code. Today, we are joined by Bradley Freemel, a seasoned software builder, fractional CTO, and tech visionary with over 20 years of hands-on experience. Bradley has lived through the rise and fall of countless JavaScript frameworks, and I think I've been going through a few myself as well, Bradley, and languages.

and held roles all the way from junior developer to software architect before stepping into executive tech leadership. Now, as a fractional CTO and founder, he's helping startups and solid founders navigate the next wave of innovation, one where AI agents, automation, and human creativity collide. His mission is to make sure software engineering doesn't get left behind.

but evolves with purpose in the age of AI. This is someone who's not only watching the AI revolution, he's building with it. And he's here to help us explore what happens when vision meets velocity and when startups scale with the teams of two plus AI. Bradley Lee, welcome to The Human Protocol.

Bradley Friemel (02:22.816)
Thank you for having me and thank you for the great introduction.

THE HUMAN PROTOCOL PODCAST (02:26.118)
Awesome. It is my absolute pleasure. And thank you for being willing to have this conversation today, because I know that that's a topic that you are very passionate about. And I think this will be a very interesting conversation as we talk about how this new world being written by AI will have an impact on entrepreneurship, right? And startups moving forward.

So to kick things off, I want to kick things off with something that I ask every guest on the show, Bradley, which is a question related to when we strip the title, your LinkedIn bio, everything we know about Bradley. Who is Bradley at his core?

Bradley Friemel (03:14.466)
Yeah, at my core, would say I'm a problem solver. That's if you really focus down on everything and strip away the titles, the accolades, the accomplishments, everything. It's all because I am obsessive about solving problems.

THE HUMAN PROTOCOL PODCAST (03:32.52)
That makes sense and thanks for sharing that with us. think that's something that we all should be passionate about, As challenges and especially in times of AI comes into play, we need more people like yourself that are really investing time on providing solutions to those challenges as they come along. And can you tell us a little bit...

broadly, how did your journey lead you to this intersection of tech, entrepreneurship, and AI?

Bradley Friemel (04:08.684)
Yeah, yeah, no doubt. It's been a pretty long journey. Over the past like 20 plus years, I've actually worked with some post-acquisition companies and was around a lot of founders early on in my career that kind of helped steer me. And I've typically...

been one of the more higher producing engineers, which also got the spotlight on me to get more time with those people. So they kind of mentored me and I have a handful of them too that I still keep in contact with. So that took me most of the way through my IC career and now transitioning into leadership. I had a private equity firm. I've been able to see the opposite side of the table on the buying side of all these, um, SMBs.

to be able to bring into the portfolio. So with all of that in mind, it got me thinking like the entrepreneurial route, pushing in that because as fun as it is to be like a high level architect and a CTO, if you're a true engineer, you still miss building. So that's where that entrepreneurial spirit ends up coming from.

THE HUMAN PROTOCOL PODCAST (05:20.264)
That's very interesting because when I think about what you just said right now, once you get at building something that stays with you forever, right? Even if you detach a little bit, if you're passionate about that something that you still, it will kind of chase you forever, right? That passion and that needs to build things and things and seeing things developing and so forth. Even if you get more into the leadership side of things.

But I'm curious about the fractional CTO part of your experience. And not directly related to your experience only, Bradley, but I'm curious myself because I think we're going to see a lot more of these fractional roles in the future as companies continue to evolve. What does it take and what does it look like?

in a life of a fractional CTO or a fractional CMO or CRO, for example, how does that work today?

Bradley Friemel (06:25.91)
Yeah, so think of it like you just have multiple projects. You consider yourself your own company and multiple projects underneath that that you're liaison for. So you have to keep great notes and good communication with your various team members so that they understand the North Star and where we're going. That's how I've been able to accomplish the fractional side of being a CTO for multiple IPs, a founder in a company and a stealth founder.

THE HUMAN PROTOCOL PODCAST (06:38.557)
Hahaha

THE HUMAN PROTOCOL PODCAST (06:55.924)
And that means providing services to multiple companies, for example, in your case as a CTO. And for those companies, that means that the services that a fractional is providing to them, it's enough. It's covering all their needs as they move along, correct?

Bradley Friemel (07:13.986)
That's correct. Because if you're like a small, an SMB or an early stage startup, I could argue that you don't really need a full-time CTO. need somebody to come in there, set the pace, help you build a small department, set the vision, set the architecture. But that doesn't take a full-time role to do that. Once the plans are set, CTOs typically just manage the engineering and try to think of cool new stuff to do.

THE HUMAN PROTOCOL PODCAST (07:44.412)
It's bringing the experience to an organization that is in very early stages. So building things from the ground up. That's what people like you in fractional roles are probably providing to this company. The experience that they need to start building really the basics of the organization or the processes that they need to. And then they keep moving from that point on basically.

Bradley Friemel (08:09.132)
Yeah, that's exactly right. And then along with that lines too, if you have experience like I do in &A, you can start the &A prep early so that exiting your startup is much easier, much more profitable.

THE HUMAN PROTOCOL PODCAST (08:15.294)
Mm-hmm.

THE HUMAN PROTOCOL PODCAST (08:21.524)
What about the solo part of the role? Because it's project-based, right, I guess, and at some point you detach from those companies, right? So I guess the fractional part of it is probably not for everyone, if I would guess, right?

Bradley Friemel (08:34.766)
Mm-hmm.

Bradley Friemel (08:40.942)
Yeah, you gotta be accepting of change and loss.

THE HUMAN PROTOCOL PODCAST (08:46.386)
Yes. As far as you are in acceptance with what you're getting and what you're going to lose at some point, right? Because it's the detachment coming to play once the project finishes, mostly. Good. Very interesting. Thanks for covering that for us, actually, because that's a topic that I've seen a lot more and more. As companies start trying to build their own processes on the department, I see a lot more fractional.

comes into play. I was always kind of curious and interested to see how that part is working today. So thanks for covering that broadly. And let's get into the kind of the topics of this conversation, which is startups of the future, right? And I know that you are someone that it's very passionate about. How do you see companies and startups being founded?

in the future, especially with AI providing all these kind of additional support that founders may need and they may not have the skills or may not have the resources for those type of skills needed. How do you see that uncovering?

Bradley Friemel (10:00.812)
Yeah, definitely. Definitely. So I see a trend now of lean startups. So think two to three people, five maximum, to be able to achieve pretty well industry toppling results. So from what I'm seeing, the two most required situations is you have a technical expert and you have somebody that's an expert in sales and marketing.

THE HUMAN PROTOCOL PODCAST (10:06.996)
Hmm.

Bradley Friemel (10:25.654)
Those two are the key roles for any of these new stage startups to get off the ground because you can allow, well, you would have the expertise on how to get money and you would have the expertise on what to build. So now comes into play. The reason why a two person startup is so legitimate nowadays is the advent of AI. Over the last few years, it's only progressively gotten better. And now with the Gentic AI process for coding,

you can do amazing work in a very, very little time and very, little taxing time, especially if you're running multiple agents at the same time. So that covers the engineering side.

But even on the sales and marketing side, there's already you can script out AI items to generate your sales script, your marketing scripts, your marketing materials, everything really. And then even from the CX and the support side, the engineers can develop very quickly a documentation book.

Pump that into a AI with voice system over it, and then have a fully integrated talk system where people can call in for help. And it's not the most difficult thing to do. It sounds like science fiction in the future. And shockingly enough, it's not the most expensive thing to do either. We're talking about a couple hundred dollars to accomplish all these tools, maybe 600 a month. So it's for what you gain and the speed and clarity you can gain.

from utilizing AI, it's unprecedented. And from me working at more larger and established companies, they cannot keep up with what a startup can do.

Bradley Friemel (12:17.464)
you have to think about it from their side. They have a big moat on technology, but that moat's getting dry. Now they have a big moat on customer support. Well, that's being taken up by automation and AI too. So now it really lends credence to these small, lean startups to push and punch back at the big established firms.

THE HUMAN PROTOCOL PODCAST (12:39.504)
And from what I see and hear here with you, Bradley, the excitement part of it, as you are working with the startup, startups usually have a lot more freedoms than large enterprises, right? To practice and to play around with tools, with processes, and you're probably seeing how faster are they getting at implementing many of these solutions that may take longer with large enterprises.

Bradley Friemel (13:06.862)
That's exactly right.

THE HUMAN PROTOCOL PODCAST (13:07.548)
and the difference may be striking.

Bradley Friemel (13:10.71)
It very much so is. And it even goes deeper than that too. So the larger companies, the established firms typically disconnected from their user base or their ideal customer profile. Startups are in tuned with it. So they tend to generate more direct solutions for direct support of their customers.

THE HUMAN PROTOCOL PODCAST (13:35.048)
That sounds about right. That sounds about right, because I'm also seeing that from what I hear from talking to colleagues and seeing how companies are approaching this type of solutions. This sounds about what we are seeing really across. And when we look at these startups of the future, and when we were discussing briefly, you and I, you see it as a very positive.

development as I do myself. And I think part of what AI is bringing to the table, it's really supporting our creativity, right? But this will require a mindset shift in a lot of people that were not used to build things themselves. AI and all these solutions that we're seeing today will give us an opportunity to develop, design and develop any idea that we may have.

And this is what you think it's also part of what the transformational period will happen, that we will see more businesses being created because AI is providing the skills for those solo entrepreneurs that they need to really kick off an idea and put it in place.

Bradley Friemel (14:54.562)
That's exactly right. And then that ties deeper into the human element now more than ever. So with that, you have some amazing tools that can do some amazing things for you. It can go off the rails quickly. So that's where it takes the creativity for like,

an individual to come in, explain to it, break it down logically and be able to implement. That's the key differentiator that I think is coming out with the advent of AI. It allows those types of people to free up some of those thought cycles instead of actually working on like what line of code is broken and stuff like that. And more into the architectural, like bigger picture creative aspect of it.

And from a software standpoint and me being a diehard engineer, that's the fascinating part. When you get down to putting your hands on the keyboard, the actual coding, that's just implementation. And I don't mind at all handing that off to something else to type for me.

THE HUMAN PROTOCOL PODCAST (15:55.912)
I wouldn't blame you. I wouldn't mind either. Right? So interesting, interesting Bradley But people will need to, I guess, embrace it, right? And I start thinking that the new way of working may be in the future you creating based on the ideas that you have. Being curious about what is it that I...

I may do tomorrow, what is it that I'm passionate about that maybe my full-time job was not providing that to me? And let AI being kind of your co-pilot, right? You as the creator and the co-pilot and the AI being your co-pilot to put that idea and design it and take care of all the technical aspects of it. I see these as a very interesting and exciting possibility in the future. And I know you do as well.

to have more people getting more into the entrepreneur side of things.

Bradley Friemel (16:56.866)
Yeah, I believe that it's the way to push society forward is to have high powered entrepreneurs start coming out of like new grads or even established players. This is the tool to do that. And with all these entrepreneurs coming out or even with the wannabes might make it big themselves just because they can topple industry or they're going to push industry forward. One of the two things that are going to happen.

But I really side more to the side of industry, like incumbents will get toppled. And if they're not considering, you know, some random startup, a threat, they should.

THE HUMAN PROTOCOL PODCAST (17:37.266)
And where is the human element part of all this development? We're still gonna be in the driver's seat. And I know you believe that 200 % yourself and you're very passionate about the creative side of things is still with us.

Bradley Friemel (17:45.442)
Yes. Yeah.

Bradley Friemel (17:54.178)
That's exactly right. If you take the best engineer or the best product person or the best sales, they're going to have some creative thoughts. They don't get to be the best without being creative in their solutions. This just unlocks their capability even further.

THE HUMAN PROTOCOL PODCAST (18:10.676)
It's kind of expanding and opening the doors for a lot of things that historically, if you did have the skills, you were able to do it yourself or have the resources. But now, even without the skills or the resources, you have that possibility. It can be possible.

Bradley Friemel (18:30.798)
Very much so, yes.

THE HUMAN PROTOCOL PODCAST (18:32.724)
But we're still kind of in the driver's seat. The idea is ours. The implementation and how this will be developed, it may be done by an AI solution, but you still need to be part of this equation because before you give your go ahead, you're still deciding what's working and what still needs work, right, based on what you're getting from AI.

Bradley Friemel (18:56.28)
That's exactly right. You have to, you have to think about it like AI is a golden tool. If you give that golden tool to an expert, they're going to build a masterpiece. If you get, give that golden tool to a novice, they're still going to blow it up. So you really have to have that human element dialed in on your deep level of understanding where you're going and those types of items to be creative enough to ask the right questions and the right prompting to the AI systems.

THE HUMAN PROTOCOL PODCAST (19:24.072)
And it's very interesting what you just said right now, because the level of expertise, especially when working with AI, will depend on the product and how well you prompt. That's the output that you're going to get. And the quality of it will depend most of the time, really, on the models that you're using and kind of the prompting. How good are you at prompting? Because that's still on us.

to put our ideas in a prompt and explain to AI what is it that we want in a way that it can provide us as good as positive output for us.

Bradley Friemel (20:04.482)
Yeah, that's exactly right.

THE HUMAN PROTOCOL PODCAST (20:06.194)
Yeah, so. So let's say that this future materializes, we have. More and more solo one or two people, companies being created, people passionate about developing ideas and having these solutions into play. Of course, we always have, as you said, we always need one that knows the business part of things, right? How to sell this, how to market, how to.

Talk to customers and the other side, usually the other person, historically has been more technical, helping setting up and designing and developing the architecture of the product. What happens after that? How the rest of the org will look like? It's just a bunch of AI agents after that?

Bradley Friemel (20:58.754)
Yeah, yeah, it really depends on the way you want to grow. So if you start hitting scale, then it's, could see there's a lot of different factors to consider here.

So it depends on your industry. It depends on your ideal customer profile. And it depends where you want to grow as an entrepreneur or a business owner. So I see really generally two paths. There's the path of entrepreneurs that want to exit and get acquired. And then there's the other side where they just want to run the business forever. And both sides are very valid, but both sides have very, very different use cases.

THE HUMAN PROTOCOL PODCAST (21:33.716)
Hmm.

Bradley Friemel (21:39.318)
So in the use case of the lean one that wants to exit, that's the one that I'm a little bit more familiar with coming from the &A background. So they typically will keep lean, get as much revenue or AR and MRR as they possibly can, and then they can go shopping around for somebody to buy them. And they look very good on paper because they have high levels of MRR or ARR.

and not that much of a head count. So they're easily integrated into pro for Leo type situation for private equity firm or investment firm or something along those lines.

So if that's the goal of that, then keeping it lean, keeping as much in AI as possible will be advantageous. But on the other side, if you want to run it as a proper conglomerate, then once you hit scale, you will have to start hiring people and experts to get into the company. Because while AI can do a lot for you, it can't do what a human can do.

THE HUMAN PROTOCOL PODCAST (22:35.7)
Mm.

THE HUMAN PROTOCOL PODCAST (22:44.161)
Yeah, I totally agree with you. Even if it's 90 % of what a human can do, still that rest of the 10 % is really what makes your company and your product a standout. It's really your vision that you're trying to put together and put it out there. AI cannot do that for us.

Because if not, everything will be the same. What's going to make you unique on your offering?

Bradley Friemel (23:12.686)
That's exactly right. It's kind of funny because in software engineering, we have a saying where the last 10 % of a feature is the hardest. And that's the exact same thing towards, yeah, yeah. It's the same thing with the company too. So in whatever position that you're in, if you're utilizing AI, it can get you 80, 90 % of the way there, but that last 10 % is the hardest and it'll tax you. And that's the most creative part.

THE HUMAN PROTOCOL PODCAST (23:23.442)
I've heard that. I heard that myself multiple times.

THE HUMAN PROTOCOL PODCAST (23:38.791)
It's the hardest.

It's another you just mentioned that it makes me think to bring it down a little bit. It's similar to construction, right? And people doing remodeling. Usually that last phase of any project, is the finalizing all the details, that's the hardest part and the part that some people don't like to do. So when I see and compare to both, it's really...

that hardest part that the human needs to drive if you really want your vision to be what you expect it to be. But still, 80, 90 % done for you, that's a huge deal.

Bradley Friemel (24:23.734)
yeah, yeah, it should. The way, once people started utilizing AI more as a tool, it should allow for individuals that are high output individuals to at least get a little bit more of their work life balance.

THE HUMAN PROTOCOL PODCAST (24:40.538)
It makes me think what will need to happen, right? And let's say that we have an idea today. We are planning to build, found a company. First, everything start with an idea. Then after that, how are you going to manage the rest? It's...

Because for some people, they don't even have experience with any AI solution managing people or managing applications. And that's what I think, for the most part, the type of role like yours, like a fractional, will still come into play. Because you will still tell them how to design the company if they want to be an AI first, for example, on this type of future that we're just talking about. What are?

how they need to structure themselves and how they need to manage those solutions that will be doing for the most part this type of job.

Bradley Friemel (25:44.066)
Yeah, yeah, that's right. If you're founding a company and you have no applicable skills in scaling or software design or anything like that, it's going to be much more difficult and you're going to require more capital to hire that initial CTO, fractional expert to come in and break it down with you.

THE HUMAN PROTOCOL PODCAST (26:04.07)
And that's still leadership, right? That's where we need our leaders and every single leader to get on board with AI, because I'm still seeing a lot of leaders not totally understanding what AI means. And in this future, as company tries to implement all these solutions, needs to start with you as a leader for you to understand it first, because that's where everything really starts.

Bradley Friemel (26:28.866)
That's right. And it's the problem that I'm seeing too, with a lot of incumbents and larger companies is the productization of AI is just non-existent. They, a lot of companies have tossed in AI like we're AI first or they're dot AI companies. And they don't really understand how to utilize a productized real AI in their systems.

THE HUMAN PROTOCOL PODCAST (26:51.686)
And that's a problem. Because that's the basic understanding that you need to have at that level to see where AI can come in and be a potential solution for your business and based on efficiencies that you want to look for or whatever is the business goal that you want to achieve. Yeah, think a lot more needs to happen for leaders to really

get on board and understand it first. That's why I see as well.

Bradley Friemel (27:24.11)
Yeah, it might come down to just the fact of a lot of people get entrenched in their lines of thought and they think that most advancements aren't going to be as groundbreaking. But I firmly believe that we're on the precipice of AI really being truly groundbreaking.

THE HUMAN PROTOCOL PODCAST (27:44.692)
Yeah.

Bradley Friemel (27:48.846)
And then from there, it's going to be, we're going to look back on this in like 20 years and it's going to be akin to the advent of the internet.

THE HUMAN PROTOCOL PODCAST (27:59.876)
It's the acceleration of it. It makes me feel excited while at the same time concerned. And I'm not gonna lie to you, right? And I make technologies on my core. I'm very passionate about how technology can improve our lives. But also what makes me concerned about it's...

how a lot of people, don't even know how to use it or they don't even know what AI means. And that's really my concern. For people like you, myself, or everyone in technology, one way or the other, we are using AI today. But there's a lot of people that they really don't have a clue. And that's my concern because, as you said, this is gonna be groundbreaking very soon, meaning it's gonna be massive. It's gonna transform how

we perform our jobs today, and I think it's gonna be in the next one to five years maximum. This will be completely rewriting our jobs and our lives. And that's what concerns me a little bit, those that haven't even get started on it.

Bradley Friemel (29:08.77)
Yeah, it depends how you look at it.

So if you look at it from society as a whole, I could see how that would be a little bit of an issue. But if you look at the early adopters, there's the opportunity to be made for them. Eventually everybody will come back or they will figure it out. They will have to, or they'll get left behind or just straight be laggards in the curve. But the ones that are early adopters like us and other entrepreneurs doing small startups, they're going to be the ones to win most of the early opportunity.

THE HUMAN PROTOCOL PODCAST (29:42.842)
It's going to get ahead, right? Basically, if you want to get ahead, jump on it. Start using it. Start informing yourself. Ask questions. Be hungry about trying to know how this can benefit not only because of the job that you're doing today, but what may be the job of the future for you. And that's always the message that I want to tell everyone, right? Just...

try to understand how this work and how you can start implementing it even in your personal life. There is so much benefit from having this technology helping you in your personal life. A lot, a lot of it. And I always, yes, just like you Bradley, I always try to focus on the positive side because everything has always a negative side, right? Any technology, anything that is disruptive at its

core like this is, of course, there always going to be pros and cons. But if we focus on the pros, we also see how AI is helping the fight against cancer, for example, solving these very complex issues that we have had for forever. So there is so much to be talking about and cheering about with this new technology.

Bradley Friemel (31:09.584)
yeah, it even goes further down than that. So even on the health and wellness aspect of it, you can get in there, understand how to optimize your body to lose weight, gain muscle, and you get all of the benefits of having basically a nutritionist right in or right at your fingertips on your phone. It is truly amazing what

THE HUMAN PROTOCOL PODCAST (31:30.516)
Hmm.

Bradley Friemel (31:34.094)
AI can do for you and even your own personal life. And anecdotally, I love baking bread and I can't tell you the amount of bread recipes that I've gotten from AI. Yeah, actually, I got my very nice sourdough English muffin recipe off of AI.

THE HUMAN PROTOCOL PODCAST (31:46.184)
Good ones.

THE HUMAN PROTOCOL PODCAST (31:53.812)
But I'm sure you put your flavor on it, right? You kind of make it broadly, right?

Bradley Friemel (31:58.324)
yeah, yeah, yeah, for sure.

THE HUMAN PROTOCOL PODCAST (32:02.292)
Yeah, it's like a personal assistant. We did not know we needed until we have it.

Bradley Friemel (32:09.176)
Yeah, yeah. Well, it's, it's the same arguments that could be made around when that the internet's bombed up because people were afraid of it. People didn't know how to use it. And now it's what 20 some odd years later, almost 30. Yeah.

because it came out, what, like late 90s where it started hitting scale? Yeah. Sorry, early 90s. So yeah. And now looking back on it, it's like, how did we live before the internet? And I think we're at the next transition with AI. And except I don't think it's going to be 20, 30 years. I think it's going to be like five, six years.

THE HUMAN PROTOCOL PODCAST (32:48.604)
I think so too. I really think so too. It's going so fast.

Bradley Friemel (32:53.676)
Yeah, I mean, it's I started using chat GPT three and now compared to their new releases and all of the Claude and Kero and agentic stuff. It's like this is a new thing. It is nothing like it used to be.

THE HUMAN PROTOCOL PODCAST (33:08.764)
It's nothing. It's nothing. it's, how are you going to keep up with everything going on? Cause I'm having a problem keeping up with all these tools, all these new models, right? And it's like, how are you keeping up?

Bradley Friemel (33:23.564)
Yeah, that's, I'm not, nobody can. I, I definitely though pick on the things that I find interesting and then I'll go and play with them to at least understand on a deeper level. Like that's how I got fairly acquainted with claw code. And then I jumped into the Kero beta and started utilizing them and understanding how they're utilizing agentic coding. And it's pretty fascinating, but

To stay on top of everything in this industry is just good luck.

THE HUMAN PROTOCOL PODCAST (33:54.932)
Every day there are news about AI, this AI product, this AI solution. It's like when I talk to my friends, tell them, okay, it's hard to keep up. You probably, you're gonna find the one that you like the most and you're probably gonna stick to it until you hear something else about a new product or another one and you give it a try, but to keep up with all the news and all these companies doing some sort of AI.

Bradley Friemel (34:04.034)
Mm-hmm.

THE HUMAN PROTOCOL PODCAST (34:23.368)
delivering AI solutions and products in the market today, it's hard to keep up. It's almost impossible.

And the same thing is when the business are trying to analyze requirements on what type of AI solutions, they're going to go through a bunch of it that they will need to take a look at it and see what's good or not, basically. What is one that you're using and that you like a lot? One AI solution, for example, or a couple, if you may.

Bradley Friemel (34:55.532)
Yeah, I have a combination that I've really enjoyed here lately. So in one of the companies that I'm helping out and actually that's still founder and

I find that if you utilize Claude code and Kero, it cranks out a pretty great architected solution. So I basically go through and write my requirements and for my feature or my phase, for instance, I take it into chat GPT to further expand it because I feel like, I've been finding that the five version with thinking.

actually does a pretty okay job of breaking things down logically for a PR basis. And then you take that, the individual PRs, because they're chunked out into items, go through Cloud Code because it has a wonderful planning tool within it.

You can take that plan then, and then put it in Kiro's prompt and it'll generate, use cases and stuff like that within the design phases and then all the way out into the implementation phase. And it's, it does as good of code that I would build naturally. And I mean, honestly, probably better because I would personally cut corners. And this is like optimizing for like usability cases that I would just gloss over in my initial

builds.

THE HUMAN PROTOCOL PODCAST (36:24.72)
Isn't that amazing? I mean, what may be taking hours or weeks for you in the past, you are having something built with quality in a matter of seconds or minutes, let's say the longest. I mean, that's a lot.

Bradley Friemel (36:38.318)
Yeah, it's I I am almost looking at it like it's magic

THE HUMAN PROTOCOL PODCAST (36:46.628)
Sometimes that's what I feel myself as well. I say, I have an idea and it tells me step by step how I need to approach this if I want to be successful. Before that, we were trying to use Google, finding articles, right? Reading through, putting things together ourselves, very time consuming. Now the thing in a matter of seconds is just tell us most of what we need.

Bradley Friemel (37:01.72)
Mm-hmm.

Bradley Friemel (37:11.384)
Yeah. And from the engineering aspect, I can't tell you the amount of times I've been on Stack Overflow early in my career. I haven't been on there in years now because I'm getting tailored solutions to my exact problem with the AI agents.

THE HUMAN PROTOCOL PODCAST (37:26.8)
It makes me think broadly, how are we, and if you're concerned, and what's your concern, and how are you handling that today? Questions and concerns are on protecting intellectual property, for example. I know companies are doing a lot more in data governments and privacy, right, because of these exact concerns, but for companies that have an IP or a product,

Bradley Friemel (37:44.354)
Mm-hmm.

THE HUMAN PROTOCOL PODCAST (37:56.04)
How do you see these AI solutions and providers taking care into consideration concerns about the piece of code going to the internet, for example?

Bradley Friemel (38:07.682)
Yeah, for sure. So that's a pretty valid question, especially when you're dealing with like HIPAA and items like that, because it's not just the source code and the IP, it's the information that the engineers or supporting staff shares even on an AI system that could potentially leak out.

THE HUMAN PROTOCOL PODCAST (38:14.333)
Mm-hmm.

Bradley Friemel (38:25.87)
And, but from what I'm seeing is a lot of them have this enterprise layer that has an SLA on there where it's like, they won't use it to train stuff on. It's, can be a concern, but you can also then have retaliation based on that. And if they're doing it on yours, then they're probably doing it on others. So then it's probably a class action. It, it would behoove the AI company to not do that.

THE HUMAN PROTOCOL PODCAST (38:49.3)
Mm-hmm.

Bradley Friemel (38:53.774)
Or if you're deeper into the compliance with like SOC 2 or GDPR or deep into HIPAA, then you can actually build your own AI systems that's localized off of various models through systems like AWS and bedrock, hugging face, stuff like that.

THE HUMAN PROTOCOL PODCAST (39:13.254)
and it stays within your tenant, I guess.

Bradley Friemel (39:15.694)
Yes. Yeah, exactly. It's tinnitus.

THE HUMAN PROTOCOL PODCAST (39:19.944)
Got it. Yeah, I think that's kind of the path forward to address some of the concern because if these companies want large enterprises that are very concerned about IP, they need to put in place safeguards that guarantee the code is not going to be exposed or not going to be any leak. And would you feel comfortable with having AI?

revisit your whole application and give you feedback where it needs to be improved.

Bradley Friemel (39:55.778)
Yeah, mean, that's, I hate to use this as a cop out, but I'm an engineer. That's part of the process. It's a review. We constantly do that for everybody. It's always constructive for the most part.

And it's supposed to help you be better. And that's the way that I've always looked at giant reviews like that. And with the AI system, it's just like a rubber ducky principle that you have something there that can bounce ideas off of you. And instead of talking to your rubber duck to explain your problems, this one talks back and tells you how to solve the problem.

THE HUMAN PROTOCOL PODCAST (40:33.446)
and where the problems are, right?

Bradley Friemel (40:35.617)
Mm-hmm. Yes.

THE HUMAN PROTOCOL PODCAST (40:38.29)
Yeah, I think that's also a positive development. Having AI can review your work, even if it's a piece of code or whatever it is that you're doing and give you feedback before you even publish it or make it available.

Bradley Friemel (40:54.382)
Yeah. And it's, interesting if you use a couple of Gentic models, you can actually use one against the other one to review its code and stuff like that. It's pretty neat.

THE HUMAN PROTOCOL PODCAST (41:05.22)
It is very, very interesting. And I'm curious, Bradley, what would be your advice to those thinking on getting started on building a company? What do you think are the first few steps that people need to have in mind from having an idea to actually put it in practice, right? And how they can leverage AI today for that to make that happen.

Bradley Friemel (41:34.712)
Definitely. So my AI workflow, if I wanted to found another company would be, would, you, the start, you can come from two different directions to generate a pretty cool idea.

So you can either go with something that you're passionate about or familiar with in your realm of expertise and then expand out from there. or you can actually go and say, like, for instance, start utilizing chat GPT with, you know, the latest model to be able to bounce ideas off of that, to further flesh them out. Like for instance, you can honestly step in there and be like, I want to understand, uh,

industries that have low technical adoption or low supporting, and then it can return a handful of industries. You pick one that you're interested in and just keep going down the research rabbit hole until you feel confident enough that there's profitability within that vertical.

So, and with all of that, you can honestly utilize AI to do market research and product research, research around incumbents, all of it to understand. So with that too, then on both sides, now we have, let's say we have a baseline idea of going into a specific industry. It's like, great. Now you can do product market research, utilizing AI to understand the incumbents and their feature set, because that'll set the gold standard for what you can build.

or what you should build and kind of what the market is saying. So from there then all you have to do is come up with your product plan since you have your features and go from there. And once you get through with that then it's starting to get the rubber meets the road. So...

Bradley Friemel (43:25.088)
From there, I would default back to utilizing systems like Cloud Code and Kero to start generating the website and or the marketing website, the infrastructure, CSC process and et cetera, so that you can start pushing your code and your project and everything live. And once you have something to show, then you start contacting potential customers.

and trying to get your early champions to work with you through that growth period.

THE HUMAN PROTOCOL PODCAST (43:57.466)
Awesome, that's amazing. Thanks for covering that with us. Are you using today any exciting tool that can help you with building your website, for example?

Bradley Friemel (44:06.894)
Yeah, yeah, no doubt. I've utilized Cloud Code to build a marketing website pretty successfully. And it actually operated fairly well. And I've used Figma Make, actually, to generate my own personal website. And it is very nice looking. Yeah, it took me maybe 15, 20 minutes.

THE HUMAN PROTOCOL PODCAST (44:26.44)
Figma make.

THE HUMAN PROTOCOL PODCAST (44:32.518)
I'm going to look at it myself because I haven't heard about that one. You see, there are so many that sometimes it's hard to keep up. But if you're recommending it, for sure I'm going to take a look at it. Definitely. Figma make.

Bradley Friemel (44:46.414)
Yeah.

Yeah, I would highly recommend it if you have a portfolio site or just a generic or a sales page that has very little interactions on it to utilize them. It's and then all you have to do is redirect your domain to point at their addresses and stuff. It's not very difficult to do. And even on my site, I was able to configure form spree so that you can submit something and automatically email to me and linkage out and everything.

I think maximum, it was 25 prompts back and forth, maximum.

THE HUMAN PROTOCOL PODCAST (45:23.43)
and it builds your website.

Bradley Friemel (45:25.186)
the entire thing and it looks incredible. It better than what I would have been able to build as a web developer in a week.

THE HUMAN PROTOCOL PODCAST (45:27.398)
with integrations and everything.

THE HUMAN PROTOCOL PODCAST (45:35.848)
Wow. That's amazing. That's amazing to hear. And for you guys watching us today, this is the type of benefits and why we are excited about AI, right? It's about developing and allowing you to kind of, you know, put your dreams and build something that you are passionate about. And usually when you're trying to start an idea or a company, one of the first things that you think about is your website, because that's kind of

going to put you out there. And these tools are really building in 20, 30 minutes, as you said, from your own experience. Before, if you don't have any software development skills, you wouldn't be able to do that yourself.

Bradley Friemel (46:20.686)
Yeah, that's right. It's like, even if you're not trying to start a software company and you're starting to, you're starting a brick and mortar like tire shop. Yeah, you can do this and you can save that five, $6,000. You'd have to chart or you'd have to go to a web developer and get charged for them to build for like 20 bucks.

THE HUMAN PROTOCOL PODCAST (46:27.984)
Any business, any business basically.

THE HUMAN PROTOCOL PODCAST (46:43.654)
It is amazing. Awesome. Thanks, Bradley. And we are approaching the last part of our episode today. And it's also a group of questions. It's only one topic, but it's divided in four to five questions about the same overall topic. And it's really your advice, starting with...

regular individuals, human beings that are not using AI today that don't even know what AI means, what would be your advice to those?

Bradley Friemel (47:25.422)
My advice to the people that have little to no experience with AI would be to first go to ChatGPT. Go on there, Google ChatGPT. There's a free client where you can just start typing questions.

And type questions that come to your head and see what happens. Like even for instance, it has so many amazing results. It's like earlier I said that it helped me generate some bread recipes. And on top of that, it's, I asked it to give me a movie that I saw in the late nineties and I couldn't really remember anything of it. And it was the first hit that came back. And it's like, that's the level of amazing opportunity that you can see.

right off the jump and then from there hopefully it'll give you a little bit more creative sense around it.

THE HUMAN PROTOCOL PODCAST (48:19.412)
And a lot of people are using AI today and they don't even know that they're using it. So when you go, for example, to Google or any of those web searches, they have already AI embed on their responses. Before, when you search, it used to be multiple options. You have to go through a scheme through and find the articles that you were interested on. Now it's really answer driven. Basically, AI will compile from what you're asking.

and give you the answer right there. This is how these things are evolving. But what Bradley is saying is really how the door will open for you in the whole AI era, because working with ShadGPT gives you a taste of what this thing can do. And you will only start understanding better how this solution can enhance our capability and can help you.

and moving forward. So totally agree with you. ChatGPT or any of these other tools are great to get started with. What about for companies and enterprises broadly? What would be your advice for companies trying to navigate the whole AI world as far as understanding how AI can help their business?

Bradley Friemel (49:42.658)
Yeah, definitely. So there's two aspects of that. You have the customer facing and the internal facing of AI. And for a lot of companies that are just using AI as a buzzword, I would have to advise to tell them not everything needs to be AI. So it needs to have the right use case and the right items to be able to even be useful for your customers.

THE HUMAN PROTOCOL PODCAST (50:01.79)
Mm-hmm.

Bradley Friemel (50:09.888)
Like for instance, analyzing reports, great for customers because a lot of customers don't understand financial reports. Personalized shoppers, another great use case for that. Upsell potentials, another one. So that's some just minor examples off the cuff of true productization that could actually benefit customers instead of just randomly putting AI into something.

THE HUMAN PROTOCOL PODCAST (50:17.459)
Right.

Bradley Friemel (50:39.254)
just to kind of catch the wave, so to speak.

THE HUMAN PROTOCOL PODCAST (50:41.502)
Mm-hmm.

Bradley Friemel (50:41.944)
But from the internal aspect is pretty interesting because I've been, I'm obviously talking about how to help engineers at Ignosium, but along with that, you can actually help customer support. And that's, I've been seeing as the biggest bang for your buck because you can go and generate all of this amazing documentation, but you know, as well as I do, not everybody's going to read all that documentation. This makes it to where you can type in your question.

gets trained on your documentation and then comes up with a tailored solution for your exact question. So that enables the support team to then go into these support calls with much, much higher level of confidence and honestly give the customer a better experience too.

THE HUMAN PROTOCOL PODCAST (51:29.844)
Absolutely, and I'm happy to hear this example because customer support is my thing, right? And I'm seeing it today as our co-pilot. And that's what I tell my team and all the colleagues that I talk to. This is unleashing capabilities and a potential that we always had in our people, but really at a scale. Because when you are working on issues, you already have a summary of your issues.

can really link and find cases on your database that are related to this, potentially be related to the same issue, can offer you KB articles based on the issues that can provide the solution to your issue. And all these is ammunition, as you were saying, even if you need to come to a meeting with a customer, our support people, they have all the ammunition or a lot of ammunition that they can have to better support the customer.

during a call while troubleshooting an issue. So it's definitely a huge advantage as I see it today.

What about an advice to these AI companies broadly providing and developing AI solutions? What is something that you would like to tell these founders? Hey, we need these products, but this is what you should be thinking about when developing them. What would be that advice to them?

Bradley Friemel (52:59.738)
It should be simple. Just please do good. Because a lot of these can go off the rails into some squirrely places. Yeah. And there's a ton of funding around these. So just please do good.

THE HUMAN PROTOCOL PODCAST (53:06.984)
very quickly, very quickly.

THE HUMAN PROTOCOL PODCAST (53:17.948)
And from my end, if I may add something, my please do good would mean test, test, and test again. Don't let this go by itself because that's a recipe for disaster. And we like it. We want more of solutions, but we want you to test as much as possible, right? Always. For anything in life, but especially with AI, that's kind of my please do good. Make sure you spend time on testing, right? Yeah.

And what about for your advice for governments and policymakers? What do you want to see on that side of the aisle?

Bradley Friemel (53:59.694)
Yeah, no doubt. That gets into some pretty interesting territory. So for policymakers and items like that, obviously they're going to have to have a lockdown version because it can't be tampered with. it becomes interesting because...

innately each model is biased to something or another. And that's kind of not what you want to see within policymaking is a bias to one or the other. You kind of want to see very even thinking around that representing your constituents. So while I think AI is good for a lot of things, I'm not sure if it's good for that yet.

THE HUMAN PROTOCOL PODCAST (54:44.244)
Absolutely. From my side, I would like to see a little bit more of the government really stepping up and understanding what this is about. And not only in the US, but globally trying to see some standards about what quality assurance means, right? For these solutions, are some of the smart regulations that should be in place?

So everybody can play using the same rules, right? Basically, those are kind of the things that I'm looking for, especially in the area of data governance and privacy. And last but not least, what would be your advice to a young broadly, but today or in the near future with AI? What would you tell a young broadly as you look what will happen with AI in the future?

Bradley Friemel (55:45.73)
Get ahead of it, stay on top of it, work hard. That's it.

THE HUMAN PROTOCOL PODCAST (55:50.18)
Awesome. It doesn't get all right. Work hard. That's going to be our lives. Even if we have AI, we're probably going to have to work harder in many things. It's just a different kind. Absolutely. I love that broadly. I love it. Really thank you for sharing your time with us today, your insight and your vision broadly for what's possible when we really think we're really

Bradley Friemel (55:54.062)
Yeah.

Bradley Friemel (55:58.562)
Mm-hmm. Yep. It's just different kind.

THE HUMAN PROTOCOL PODCAST (56:19.118)
startups of the future is in times of AI. And if today's episode really gave you a glimpse of the future, please share with your family, your loved ones. We have episodes every week, and I hope that we can see you again. And one more time, Bradley, thank you for your time. Really appreciate it.

Bradley Friemel (56:43.448)
Yeah, thank you.