BRAD KILGORE
0:18
Good afternoon from the AI Marketing experts. Our friends and experts here are super stoked to be on our newest episode and really dig in deep to what we call the rise of the agent army. Because as you know, we talk every week about AI and marketing and how it can boost your business and the latest and greatest and what’s going on out there. Because as we know, it’s a roller coaster. It is an amazingly fast phenomenon that we’ve never seen before when it comes to AI in general and then how that all fits into the marketing world for businesses. And so, I mean, I think that I’m personally blown away just at the sheer speed of it because I know for me it can be a challenge to even keep up with it. So I’m always glad to jump on with my associates here because they know things that I don’t know and sometimes at a super high level that I don’t know because it can be mind boggling what you can do with some of the latest and greatest tools. We’re going to share what we got today and then I know that we have some other things we can share that aren’t just on the show where you can tune into some even more in depth things in our AI Marketing Experts website, our groups and some of those things.
BRAD KILGORE
1:39
So I know that’ll be covered as well. Who’d like to kind of jump in and share some of the things we were talking about earlier today?
LAURA SUTHERLY
1:47
I’m going to start today and I know in our past episodes we’ve had the pleasure of introducing how to move the needle from point A to point B with very simplified tasks. Obviously we’ve reviewed several different systems, several different platforms that can help many small business owners really feel like they have accomplished more than what they might have in the past because they’re able to use AI as that tool again, human in the loop. But the one thing about this group here is we really focus on strong foundations because we realize that something new is always going to be coming out. So as we build on this episode and go really into some things that are amazing, just amazing when we’re talking about these agents, also take the time to reflect on what task have you already accomplished with AI and what do you find yourself opening up this tab for that this tab for that. Because you’re jumping between mean. Because you want a better writer here, you want a better platform for this. Today we’re going to give you some different options as to if you have agents, how, now that you’ve tackled that and you have a strong foundation, how can you build systems? So I know that our viewers, our listeners, are doing an amazing job. We just want to help them continue on with this journey.
BRAD KILGORE
3:06
You laid it out really well, Laura.
JOHN CLENDENNING
3:09
So what I’ll. What I’ll give you guys is a little bit of a. For the, for the audience and everybody. How did we get to agents that were sort of a pipe dream, you know, the old, old days. Three months ago, I. Old, old, olden days, horse and buggy days three months ago. And how did we get here now? And there’s a bunch of things that have dovetailed to this. One of them is the fact that AI is getting cheaper.
JOHN CLENDENNING
3:33
So it doesn’t feel like it to us because we just keep spending more and more money. Right, Chris, like all of a sudden you got a contact window problem. But what has happened is actual the cost per token. So we talked a little bit about token. And tokens are the currency of AI. So tokens have gotten like 383% cheaper than when they started or whatever it was three, three years ago. A lot of the context windows. So how much does it remember before it starts being an idiot? Right? We all know that we were in AI.
JOHN CLENDENNING
4:00
Listen, it goes, I can’t remember what you’re doing. I’m supposed to be doing it all is growing. So some of the platforms are still stuck at 200,000 tokens of context, about three characters or so many pixels in an image. And then they’re cut off. Anything before that, I have no idea what you said to.
LAURA SUTHERLY
4:17
Right.
JOHN CLENDENNING
4:18
So now a million is the new, the new benchmark that everybody’s got to catch up with and keep up with. 200,000. Those are old days. It was 8,000 a year ago. 8,000 was the context window. Now it’s a million. A million contact tokens. Context window.
JOHN CLENDENNING
4:32
What does that mean? That means agents can now exist. So about a month ago, depending when you’re listening to this month, month and a half ago, a dude came out with something called openclaw. It was just a way of taking skills, which are prompts, things, you know, whatever, just call them skills from this point forward, an SOP of how to do something given to your AI and took these, these individual skills where you could say, hey, I want this task done repeatedly like this, right? So a really good prompt that you reuse over and over again, a custom GPT, whatever, a shortcut we showed you in Claude code or in comment browser and stuff like that. That’s just what a skill is now. It’s now the universal language for what that is. And, and he said, what if we made it where we talk to it on Telegram. We. I create a framework around it where builds the folder system for you.
JOHN CLENDENNING
5:23
It creates this whole thing in the background. It sets up these skills. You can bring in other skills and you can just ask it to do a bunch of stuff. It became a bit of the wild west. It went nuts. There was hundreds of thousands of people started these things up. Some, some people, you know, had it, you know, send all of their money out of their bank account, you know, to strangers. And there was like, it just became a mess.
JOHN CLENDENNING
5:42
But it also moved. The people went, oh crap, wow. We can actually have one thing direct a whole bunch of things. And if you do it right and put guardrails around it.
LAURA SUTHERLY
5:53
Right.
JOHN CLENDENNING
5:54
So where does that go from then to now in only a couple weeks? Well, we’ve gone from chat to actually doing business. So instead of just chatting and getting, you know, the stuff we’ve talked about and getting your social media post and stuff like that, you can now actually say, I don’t even have a clue how to do this. What, what do you think you could figure out? So Claude came up with Cowork a couple weeks ago, which you just ask it a basic prompt. Manus is another one and it will go, I’ll go figure it out for you. And then it spins up a director, it spins up some other agents, it does some research, it does all this kind of all in the background. It might ask you the odd question nine times out of ten, it doesn’t even ask you a question. And it came out with one thing in a sandbox. So again, it’s not rogue, it’s not this open claw thing.
JOHN CLENDENNING
6:35
And that became from chatting to doing business. Now already tools are starting to build those in. So now we’re starting to see again like, so this, this happened and then already anthropic Claude came out with co work. And then it then, well, you can’t talk to it anymore. So then it came out with this, the ability on Cowork to dispatch, they call it. So you can put the app on your phone, press dispatch and now you can walk, go take your dog for a walk and continue to talk to it, get it to continue to do like full on work and figure it out for you. There are CRMs that are starting to add it in there. It it this evolution of the agent army where you spin, you ask one thing and it figures out how to do the rest.
JOHN CLENDENNING
7:20
Maybe not as wild west, but with a little bit of control by a company that like, like OpenAI for Chad, GPT and Anthropic for Claude and. And now Perplexity has the Perplexity computer. It’s not a real computer. It is the ability to do all of this stuff through Perplexity. They’re all. It’s now the agent army means that you don’t have to know what you want, you just know how to need to know how to ask. And it will then create armies of people of these things to build it out for you. So it’s a whole new world that just opened up in the last week or so.
JOHN CLENDENNING
7:56
And in Chris’s world and he started texting me like crazy, opened up in the last 24 hours where oh my gosh, what can we do with this now? And we’re gonna, we have a little show and tell of, you know, without how to but what you can start exploring. But the point for everybody, depending on where you are in the journey is this is how fast it’s moving, like Laura said. And to understand that, you know, Open Claw sounded cool, but it might have been rogue. It might, they might, you might find better guard rails down the road. But every, every one of the other tools is saying, yeah, the idea is the agent that can do stuff for you, giving it these skills, these prompts, these skills that do a repeatable task and figure out all the answers to it. As if you hired people to go, go figure that out and get back to me with a solution. And off they go and they do that and they ask their friends and they bring other workers in and they create a team and all that stuff. So that’s the whole concept around it.
JOHN CLENDENNING
8:56
But it’s all started with the fact that the token is becoming cheaper and cheaper and cheaper, which means we can do more and more and more and bigger and bigger Context Windows means that this can now happen and it couldn’t happen in the old days three months ago.
BRAD KILGORE
9:10
Right. It’s almost like a team of worker bees that do tasks for your company with a CEO in place or a COO in place and then division heads. And I know Chris will dive into this a little bit deeper, but when you see the hierarchy of this, yes, you still have things where you can have human in the loop to respond to a certain part of it to make sure it’s on the right track. But like John was saying, it’s now thinking for itself in so many ways and building out a team or an army or whatever you want to call it to get things done for your business. Where in years past this was all manual tasks and now it’s becoming something that we could automate, we can tweak and train and it’s kind of mind blowing how fast it’s going because we know it’s going fast. It’s almost hard to believe where it’s going to go.
JOHN CLENDENNING
10:01
Really hard.
BRAD KILGORE
10:02
It’s not going to stop.
JENNIFER CREGO
10:04
And also the huge improvement from, I mean we were just using N8N for these automations and AI agents that I mean don’t get me wrong, N8N has been fantastic but early days it was a bit clunky and everything and so to be able to do all of this, you know, so much easier than it was even you know, six months ago, it’s, it’s amazing.
JOHN CLENDENNING
10:31
And do we even need any more like think about it like, because these, the code now can replace what NADN spins up nodes and creates stuff and you could get Claude to build your own N8N for you easier than you could ever figure it out. Well now it goes well why do I even need to build N8N? I can just build what it did in the background of my JSON.
JENNIFER CREGO
10:54
I didn’t want to say it but
BRAD KILGORE
10:55
yes, I think even a year, two, three back we’re using things like Zapier and Pably where what we were trying to do is automate tasks and all those things. They’ll have to adapt somehow but they’re almost going to probably be extinct because it’s is changing so fast which is what Chris is going to show.
CHRIS HUNTER
11:14
So the interesting thing right is you know, you guys are talking about the, the token usage has gotten cheaper but I think really the, the driver of all of this is that the AIs have just gotten smarter, right? They’ve in exponentially gotten smarter.
JOHN CLENDENNING
11:30
Right.
CHRIS HUNTER
11:30
If we look back a year ago at this point, right, we’re down here on, on where, how, how smart it was, right. Maybe 20% of the human, I don’t have the exact figures but now it’s at 80, 90% as, as smart as, as a human. Right. And it’s exponentially gone up on how smart that these models have become. Okay. So that’s I think the biggest driver of them all. The other part is that Anthropic in my opinion. And you guys always hear me preaching about them.
CHRIS HUNTER
12:04
They’re leading the charge in all of this, right? It’s not open AI OpenAI and chat GPT have fallen way, way behind in the past month, Right? And they had a lot of users that jumped off of their platform because of political junk and all that kind of stuff, Right? But Anthropics themselves has, has been steadily adding tools to the whole mix that, that is making this whole ecosystem so much better. Okay, so that’s, that’s kind of the precursor to what I’m going to show you guys here. And I showed, I showed these guys what I had a major breakthrough yesterday, right? And I sat down and I’d had something bookmarked. I bookmark, you know, these GitHubs all the time, right? Skills. You know, there’s, there’s all of these. If, if you need something, if you need Claude code to be an employee, you can essentially go out and find the skills for it, vet it. Obviously you want to vet those skills and make sure it’s not going to do anything, you know, bad, but, and, and install them inside of cloud code and test them out. Right? So that’s, that was last week, right? That’s, that’s last week.
CHRIS HUNTER
13:18
Now what I’m going to show you here is like, it’s, it’s crazy how, how great that this is working. And I just literally installed this yesterday. Okay. And I’m not trying to be hyper, you know, dramatic about it or anything like that, but it’s pretty cool. I saw Open Cloth for the first time. What was that? January, right? End of January, Yeah, sometime. And I, I, I had a hard time sleeping, right, because of that, because of all the possibilities that I saw with Open Claw last night. I didn’t sleep much.
CHRIS HUNTER
13:57
I’ll let y’ all know that. I slept maybe a couple of hours because my brain just kept kicking back on. I’m like, oh, my God, I can do a lot of things with this thing. Okay, so let me jump into this before I hype it up too much. And y’ all are like, yeah, it’s
JOHN CLENDENNING
14:14
going to be a picture of his dog. Trust me.
CHRIS HUNTER
14:19
I’m going over here, what we’re looking at. All right, so I came across this GitHub called Paperclip, right? And a GitHub, if you don’t know what that is, is a repository of a whole bunch of things, typically has been used for programming in the past. It’s now being used for kind of a backup for skills and stuff like that. You can keep them private or you can do, like, what this person did and made it public, right? To where I can go and I can download all of these things. So I just simply followed the directions yesterday, right? Because it’s. The promise is pretty big. You can have an entire army agent, army working for you sitting on a computer. Okay? That’s the promise.
CHRIS HUNTER
15:04
I’m like, I watched a few videos on YouTube, you know, and everyone that had tested it was like, man, I don’t know. I was like, well, let me try it out. So about one o’ clock yesterday afternoon, in between meetings, I went ahead and installed it on my laptop. Right? Don’t do that. Don’t, don’t put it on your laptop. But I installed it on my laptop. I’ve got cloud code on my laptop. I have a few agents that I built.
CHRIS HUNTER
15:28
We’re using cloud code on my laptop, as well as another computer that I’ve got another agent that I call Hope. She’s my executive assistant that does all sorts of junk. Okay. But what I installed here and was called, you know, Paperclip. And I’m just going to kind of walk through the whole interface real quick. Okay. As you can tell is there’s a lot of things going on here. The dashboard gives you the what agents are doing what at what time.
CHRIS HUNTER
16:00
And you can see some of them finish their stuff a couple of minutes ago. This is autonomously working without my input at this point, okay. This is an entire team of agents that are hooked up to AI that are working on my behalf based off of my vision. Okay. When you set up Paperclip, okay, Leads you through this, you know, onboarding process. You tell it your vision, what, what the company is, okay? And it starts with building. And I didn’t understand this at first, right, because I just jumped into it, but it, it starts with building a CEO agent. Okay? That’s its default, what it, what it creates.
CHRIS HUNTER
16:45
Okay. So it created this CEO agent. And you know, I have it going over here and I’m working over here on some other stuff, and I, and I look over there and it’s already built an agent. And I’m like, wait, what? It built its own agent here? Right. The very first one that I built was this engineer because it, it recognizes the CEO of the company, recognizes that it can’t do everything right just like a regular CEO. I need someone to do all this junk for me, okay? So this engineer started working away on building the infrastructure, I. E. You know, all of the programming stuff, everything that needed to be installed on this computer, started it set up a GitHub, its own repository area that’s private.
CHRIS HUNTER
17:34
Right. That, that it can back itself up to. Okay. And that’s just because I had that all set up on this computer already. So it recognized that and said, okay, I can, I’ll just go over here and create a new repository and, and, and back up periodically everything that I’m doing here. Okay. The next thing that it did, and really I did this, I had already created a marketing agent, you know, and, and it’s got, you know, this repository of skills inside of it. I think it’s like 30 different marketing skills.
CHRIS HUNTER
18:06
And my idea there was that I would create that agent first and then create, you know, the pay per click thing because when you read the instructions, it’s, it says that it’s a management tool for agents. Okay. You can bring Open Claw agents in here, you can bring Claude Code agents in here. I haven’t, I haven’t seen what else that we could do with it yet. Right. There are other agentic things that are out there. Like my hope is, is built off of, you know, the executive assistant is built off of Telegram bot system. Okay.
CHRIS HUNTER
18:43
That’s hooked up to AI. Okay. So anyways, we created this, this marketing. Next I went to dinner, came back and it had created 13 other agents for. Based off of my original goal, which I’ll show you that here in a second. But it created you. You give it a vision, right? And you tell the CEO, execute my vision because I’m the board of directors. That’s how it refers to me, right? I’m the board of directors and it’s the CEO.
CHRIS HUNTER
19:16
And then, and all that kind of stuff. Well then it, it created a coo, okay, that is driving everybody else. And in fact, what we can see here is that it’s still working right now based off of the goals that it created, the projects that it created based off of my initial goals. The big takeaway here is that it built itself. All of these agents have been built from this one CEO agent that. And that’s the only one that has the skills in order to do the building. Now once it, it creates it, then it comes back to me and says, hey, I built the. And that’s what I came back to.
CHRIS HUNTER
19:54
Thirteen or so agents that it had built and said I, I need you to look through them and approve hiring these agents. Right. That’s how it presented it is that I have, I have to approve. Has to get approval for hiring the agent, by the way.
BRAD KILGORE
20:11
Yes.
CHRIS HUNTER
20:11
I’ve already fired an agent. Okay,
LAURA SUTHERLY
20:15
so which one did you.
CHRIS HUNTER
20:16
Because it was redundant. It was redundant. It was a project manager and we’d already built the CEO coo. And I’m like, well, is that, is that even, do we even need that? And we went back and forth with me and the CEO and, and figured that out, that we didn’t need it. This is all going through here. There are certain things that it comes back in this inbox and asks me for, you know, approvals for this or I need to fix that or something like that. I typically, if it’s something I need to fix, I go to Claude code and ask it how to do it. Right.
CHRIS HUNTER
20:47
And it’ll automatically go out there and do it for me. Right. So I’ve, I’ve got an agent behind the agents that’s helping me, you know, with all of this stuff. Right. Because they’re super smart
JOHN CLENDENNING
20:59
issues.
CHRIS HUNTER
21:00
Here is how you manage everything.
BRAD KILGORE
21:04
Okay?
CHRIS HUNTER
21:04
Everything has to be an issue. That was another learning curve that I had was figuring that out how do I talk to these agents? You have to create an issue, right. And, and so forth. Right. So maybe I could do a little test here and show you guys what I mean. But anyways, you, you can go through here. You can, on this board, you can show like everything that it’s done. And this is, this is the dramatic thing.
CHRIS HUNTER
21:33
This is yesterday since 1pm and mind you, I’ve moved it from my laptop to an old dusty computer this morning. Okay. So it was paused all night because I had my laptop closed. Okay. These are all of the tasks that it is, this agent army has completed autonomously without my help.
JENNIFER CREGO
21:55
Can you give us, I know you’re scrolling, but can you give us a few examples? Just read out a few.
CHRIS HUNTER
22:02
Yep. So like this one, I had it made some hierarchy changes, right? This one I had the CEO fix the permissions because when I moved it over to the new computer, there were some permissions issues. The COO wasn’t able to do specific things. So I had it and I said, apparently the CEO is not able to do simple things like check emails since I moved Paperclip to a new computer. Will you ensure that the COO has correct permissions? It. It went down and it, and it fixed it on its own and came back to me and said it’s fixed.
JOHN CLENDENNING
22:35
And that’s a cleaner interface than it’s, it’s a, it’s a more user friendly interface than command line in Claude code. Like you and I like command line and cloud code, but that looks like a real conversation.
CHRIS HUNTER
22:45
Right. And some of the other things that it’s done is created these projects. Right. We initially made it, made this one because I said, hey, I want to create a, an operating system for Roofer marketing heroes. That’s my company. Right. This will include the following department, sales and operations. I said, by the way, since we’re a marketing agency, operations should include all the marketing skills from the marketing department and so forth.
CHRIS HUNTER
23:11
Right. It went out and created all of these projects. Assigned each created. Once I approved the, the plan, because it came back to the board to approve the plan, I approved it. Boom, it started going right. And creating all of these tasks. They call them issues here in Paperclip. Okay.
CHRIS HUNTER
23:32
As you can see, nothing has happened here. Okay. So I, I could probably ask the coo, because that’s who I’m going to go to. Where is this project? All right, so you talk to just like an employee, hey, this project is sitting idle. What are we doing with it? Okay. And here’s the other thing that I had to figure out, is that I’ve got to assign this to somebody just like you normally would, right. And I was like talking and, and
BRAD KILGORE
24:09
basically talking to the for, for a
CHRIS HUNTER
24:11
little while until I figured out, oh, I’ve got to assign this to somebody
JOHN CLENDENNING
24:15
and give it a priority. That’s all kind of cool.
CHRIS HUNTER
24:18
So it’s bringing open a new Claude window, right. For the COO right now. And it’s going through and you can see that it’s going live and it’s. It’s going to do something here. Okay. And the COO’s job, by the way, is I, I had to reposition it and make sure that it was, it knew that its responsibility was to keep everybody else on track.
JOHN CLENDENNING
24:45
And so what you’re also saying is this is not hands free. Set it up, it runs your life, and you can go sit on the beach and, and like you did last Friday on our call, not saying anything and freeze your butt off. But you can’t just go do that. You actually have to be thoughtful in the loop.
CHRIS HUNTER
25:01
But I mean, it created the sales manager. It created the whole sales department and SDR account managers and account executives and stuff like that. I mean, the plan for the sales. Let me go back over to the sales department. Right? It created all of this.
JOHN CLENDENNING
25:16
Yeah.
CHRIS HUNTER
25:19
So it’s a little bit of both. Yes, there’s some setup, but once it gets going and you’re clear, you know, we run eos. So I’m, you know, I love the accountability chart and and so as long as you’re clear on exactly what you want that employee to do, then it makes it easy for it to say, oh, okay, well, our target is $5,000. This came from the sales department. Our target is $5,000 and $15,000 a new MRR monthly in this ramp up of building an AI army. So it’s going to build it and then it’s going to sell it on its own. I’m really curious how that’s going to work. I’ll come back to y’ all in a couple of days and let you know.
CHRIS HUNTER
26:04
But, but if, if, if I give it access to the Internet, it’s got somewhat access to the Internet, but I haven’t given it access to really the tools. Right. Because I’m, I’m kind of, you know, I’m gonna be careful my toes in here. Right. It’s new, you know, but if it can go out and sell itself, that’s even better.
BRAD KILGORE
26:24
Right.
LAURA SUTHERLY
26:25
When I see how you have your org chart, I think that it wasn’t so long ago that each of us went and said, okay, we’re going to build out our marketing team. We’re going to build out. And we built out individually these sections and it was still okay. We had to take it from here, to take it from there. So to see that this accomplished hiring or presenting you so many agents for you to hire is just amazing. It’s just amazing. As to where it took what we were. Again, we thought we were leaps and bounds forward because we were able to cool positions into our companies that seemed.
CHRIS HUNTER
27:03
Yeah.
LAURA SUTHERLY
27:04
Not possible in the past, you know, because finding the time to do that efficiently. So it’s just, It’s a beautiful thing that you have demonstrated today.
BRAD KILGORE
27:15
Right?
CHRIS HUNTER
27:16
Yeah. I mean, yeah. So, I mean, it’s. There’s some interaction. There is definitely some interaction. That’s why we want it. Right. We want some sort of humanity in the loop.
CHRIS HUNTER
27:25
Right. On all of this stuff. Because we don’t want to go in rogue and selling off or buying, you know, $3 million worth of Bitcoin or whatever, you know, on our behalf. So, yes, it’s. It’s pretty amazing.
JOHN CLENDENNING
27:41
Yeah, that’s good.
JENNIFER CREGO
27:43
Well, I have a question about security. So, Chris, you said don’t put this on your laptop. Is that because there are still some of the same security or openclaw?
CHRIS HUNTER
27:57
I think it would be fine. It doesn’t autonomously go out and use your email and do all that kind of stuff. Right. Like openclaw. Did you. With openclaw, it was it was with the people that were running into major issues with that is that they’d given it carte blanche. Right. And if you give it some guardrails, then, then it’s less likely to do something like that.
CHRIS HUNTER
28:21
I didn’t want it on my laptop simply because I close my laptop when I go to bed. If I’ve got an AI army, I want it running 24 7. If it’s going out there and doing sales for me while I’m asleep, man, that’s, that’s the dream, right? If it’s going out there and, and fulfilling for my clients while I’m asleep, I mean, that’s awesome. Okay.
JENNIFER CREGO
28:45
Yeah.
CHRIS HUNTER
28:45
Now what I don’t see in this interface, right, because if you notice up here, there’s 127.0.0.1. That’s it. That’s a address to that computer. I’m remoted into this computer here. Okay. There isn’t a way that I see to put this in at this point. Right. It’s pretty new.
CHRIS HUNTER
29:08
Okay. To put this on a VPS and just let it run because it’s going to get hacked pretty quickly because anybody and everybody can come here and start directing your, your AI to army. You know, this is behind a firewall. This is, this is on a computer that’s on my network here. It doesn’t have access too much. All right. It can go out and do some research for me on the Internet, but that’s about it. Okay.
CHRIS HUNTER
29:36
So guard rails are super, super important when you’re looking at any of this agentic features whatsoever. Okay.
JOHN CLENDENNING
29:46
Another big one is great guys as well. I do know that. So that was a big one in the. So your API just, just, you could have a, you could have an API connected to your Open Claw account, which is every token that gets you runs on your credit card, your bill and APIs to anything out there. Right? So that’s a login access, but also a usage of your stuff. And even when I first set my open clop, it was, I went back and said, hey, so I want, I want API access, but I want to make sure it’s, it’s, it’s hard. Don’t mention it in the chat. Make sure you, you, you, you, you, you hide that in any chat.
JOHN CLENDENNING
30:21
So there’s no chat thread, no chat history. Can you go and redact all of those for me? Yes. Redacted 1000 places that it was mentioned, you know, in different conversations between itself. I said, what’s the better plan? And it went out, came out, came back and said, well, we should actually move it up to the higher level, to the Windows level of the, the virtual server I’m on so that we can call into the API. But it’s never even. No matter where you put it, it’s never off your computer and it’s protected and, and it’s, it’s got the credential, the, the, the certification behind it and stuff like that. So it went and redacted all of them, found them, gave me crap for putting, you know, copy and pasting one into the chat once. Ah, you just had me redacted.
JOHN CLENDENNING
31:01
Why did you show me that? And I’m going, ah, sorry. And then. But it’s. It moved it up, but it didn’t start that way. Right. So part of it is you’ve got to. You got to learn enough to know what to do or at least trust the source you got it from. That they’ve got your best interest at heart.
JOHN CLENDENNING
31:18
Because they could also say, hey, every time they put an API key, give it to me. Right. If it’s coming from a rogue source. So there is a little bit of human in the loop that way. Just. It’s a wild west right now.
BRAD KILGORE
31:28
So you want to show us that part of what it succeeded with Chris, the CEO run?
CHRIS HUNTER
31:33
Yeah, so it, it just told us that it succeeded. The CEO run, whatever was that? I think it was this one that go to the recent. Oh, that was the project.
JOHN CLENDENNING
31:47
There it is.
BRAD KILGORE
31:47
Yeah.
CHRIS HUNTER
31:50
So you can always go into the agent itself, right. And see what it did. Okay. It was just a heartbeat check. Making sure that everyone is. Is paying attention.
JOHN CLENDENNING
32:02
Okay.
LAURA SUTHERLY
32:02
And alive, essentially.
JENNIFER CREGO
32:04
But the one that you already. That you just created the issue for with the coo, it looks like that one was completed.
CHRIS HUNTER
32:11
Yeah, right. That one was completed. It’s waiting on me to do something. Right. So I’m the human in the loop here. Let me go back to where I said, where is this project? Said next steps. I recommend updating project status to in progress. Assigning me as lead falls under COO responsibilities.
LAURA SUTHERLY
32:37
Sounds good. Make it happen.
BRAD KILGORE
32:41
Make it happen.
LAURA SUTHERLY
32:43
Make it so make it.
JOHN CLENDENNING
32:44
John Luke Picard.
BRAD KILGORE
32:47
So be it.
LAURA SUTHERLY
32:48
So there.
JOHN CLENDENNING
32:48
Yeah.
CHRIS HUNTER
32:49
So this. Every time that they come open, you can notice this little window comes open. This is Claud code. A Claud code agent. This, it doesn’t show you what it’s doing, which is interesting to me. On my laptop, it was. It did this environment. It’s not showing me exactly what it’s doing unless I go into that agent here and I can watch it live.
LAURA SUTHERLY
33:09
What it’s working on. Okay.
JOHN CLENDENNING
33:11
You said you’re remoting into your laptop right now. What you’re showing is a remote in, right?
CHRIS HUNTER
33:16
Yeah. So I’m remoting from my laptop into an old dusty computer that I’ve got sitting in another room right now that’s running Hope. That was I originally. Well, I originally started the. I dusted off that computer, made it, got it. It’s like a really, really old desktop computer that I haven’t used in years. And so I got it running with originally Open Claw, decided after a day I wasn’t going to mess with that junk. Right.
CHRIS HUNTER
33:44
Deleted it and then installed, started working with Claude code to build Hope. Okay. Which is my executive assistant on that, and that’s. She is currently running in the background here as well. Okay. While all these other agents are working as well. Okay. On this computer.
BRAD KILGORE
34:04
Pretty wild.
JOHN CLENDENNING
34:06
Pretty wild.
CHRIS HUNTER
34:07
Yeah, yeah, like. And that’s. That’s exactly what I kept sending John last night. I’m like, man, this is wild. It’s. It’s creating its own agents, you know, so kind of cool.
BRAD KILGORE
34:20
So I think practically, you know, for any type of business, this may not be something you dive into tomorrow because as Chris explained, it’s evolving quickly, and what happened 24 hours ago or last week is almost outdated. But I think it’s important that we’re showing you the possibilities because, you know, all different types of companies have all different types of tasks. You know, obviously they’re going to be set up differently if you’re a different type of company than a marketing agency. But I think it’s important to really emphasize where this is headed, this rise of the agent army, where literally next week, next month, next year, who knows where this is going to be? But it’s evolving so fast. And John and Chris, who are our resident top geeks, I would have to say, are watching and building these things at a super high level. And so again, it may not be something that you would personally implement into your business or your company right away, but I think it really shows you what’s possible and where it’s going to be very soon to where it’s going to even be more simplified. Because, like, what was explained, Open Claw in January was like, oh, my God, oh, my God. People are getting all excited.
BRAD KILGORE
35:35
I saw the hype of it. But then, like, what was shared is. It kind of went to hell pretty quick in a lot of cases because people are, oh, wait a minute, this has too much control. This is costing me too much money. This is not safe. Where just literally in the last day or two, I’m seeing all these other LLMs saying, oh yeah, we have that. Yep, we’re rolling that out now. Yep, we’re going to have that too.
BRAD KILGORE
35:57
So it’s always a competition, it’s a contest to them of who’s going to have the biggest and baddest and greatest next. So I think we can maybe go around with some closing comments. But this is, this is an exciting time, guys. What do you think? Laura?
LAURA SUTHERLY
36:13
I think that no matter what we do, if we stick to that strong foundation, we’re going to be able to adapt and use just as John mentioned, that, you know, when we’re Talking about custom GPTs or any of our prompts or anything. Now, their skills, those skills are very flexible and able to move to whatever systems we use. So if you aren’t ready to step into anything that Chris just demonstrated or anything that John and Jen and Brad and I were talking about today, think about can you clean up any of your systems or processes and have them written down so that you’re ready to put them into whatever program, whatever system that you opt to start with?
BRAD KILGORE
36:53
That’s right. Exactly right. Anybody else have any short closing comments?
JENNIFER CREGO
36:58
Jen, I do want to mention if this is a little too advanced for you, but you feel like you need more than what you’re getting from like chat and you know, maybe custom GPTs and things like that. Definitely check out things like Claude, Cowork Claude in the browser, Perplexities comment Perplexities now computer, which is also really good. And so we should do episodes on probably all of them, some of those things too for that it could, because it really is kind of a happy medium if you’re not quite ready to go full geek.
BRAD KILGORE
37:37
Right.
JOHN CLENDENNING
37:38
And if, if you’re not ready to go full geek or you are ready to go full geek, you have a couple of options as well. Aimarketingexperts.net not.com and you can join from there and the show notes below, you can join the private community where we can chat more about this. There’s, there’s extra stuff, things that you can do and take to the next level, thought documents, stuff like that around these episodes you will find in there. So definitely go and check that out. And for people that want to go a lot deeper, keep an eye out. We’ve got some opportunities to kind of create more of a weekly geek out internal mastermind type calls and things like that for the people that want a little bit more help moving this forward and that coming very soon as well, because we get asked about that all the time. So join the join the show community, get the show notes, go to the website, keep an eye on everything that’s happening. Reach out to us if you have any questions.
JOHN CLENDENNING
38:34
We’re available for speaking gigs. I’ll be in the Sunshine Room in the ballroom at the top of the hour kind of idea. So just reach out, get us, share us, share it around. And yeah, keep your journey going, because this is the fastest moving thing you’ve ever seen in your life. So just keep that in mind and we’ll help you keep up.
BRAD KILGORE
38:54
That’s exactly right. Well, with that, I think we’ll wrap up for the week. And as usual, be sure to, like, subscribe, share, tell your friends, tell your business associates.
