Using AI to Eliminate Busywork and Unlock Creative Time with Jason Glaspey

Disclaimer: this transcript was generated using AI.

Bjork Ostrom: One of the biggest mindset shifts I see successful food creators make is this. They stop thinking only about traffic and they start thinking about product because traffic is great, but real leverage comes when you get good at selling something you own, recipes, meal plans, a membership, a system that actually helps people. The problem is that creating a product is really hard. You have to figure out the tech, the structure, the payments, the delivery, and honestly, that’s where a lot of really great ideas die. That’s why I want to tell you about Member Kitchens. Member Kitchens makes it incredibly easy for food creators to sell recipes and meal plans at scale without having to build everything from scratch. You don’t need to be a developer. You don’t have to have a complicated setup. They’ve already built the infrastructure for you. You bring what you’re great at, which is creating food content your audience loves, and Member Kitchens handles the heavy lifting of turning that into a real sellable product.

And if you’ve ever thought, “I know I should have a product. I just don’t know where to start,” then this is your starting point. You can check it out at memberkitchens.com and start thinking beyond traffic and toward building something that actually grows with you.

Emily Walker: Hey there, this is Emily and you’re listening to the Food Blogger Pro podcast. This week on the podcast, we’re welcoming back Jason Glaspey from Fleet of Geniuses and Copy Club. In this episode, Jason shares his journey from an internet entrepreneur to AI builder and explains how creators can start using AI tools in practical ways to simplify their workflows. Bjork and Jason talk about the current state of the internet, why creators often feel stuck reacting instead of exploring new opportunities and how AI can help shift that mindset. Jason also introduces OpenClaw, an AI-powered tool designed to proactively help with tasks and projects. They discuss how it works, how they’re using it in their own workflows, and how tools like OpenClaw can automate repetitive tasks like grammar checks, link audits, and SEO reviews, so that you can focus on what really matters, making compelling, human-centered content. If you’ve been curious about how to use AI in your business without losing the human element, this episode will help give you a helpful starting point.

Without further ado, I’ll let Bjork take it away.

Bjork Ostrom: Jason, welcome back to the podcast.

Jason Glaspey: Hey, Bjork, how you doing?

Bjork Ostrom: Good. Last time you were on, the internet connection wasn’t as good. You were in Mexico. You were living this really unique, creative existence with your family. You are still doing a version of that. You’re in Oregon right now. And it’s been fun for me to know you through all these different phases of not only your personal life, but also your business life. And the other fun thing with it is you also had a touchpoint in the world of food. You had a food site and a food business online. So talk us through a little bit of your journey, what brought you here today as somebody who focuses a lot of their time and energy, their creative thinking on AI, but how did you get here? What did that look like for you to get here? And then we’ll dive into AI for creators.

Jason Glaspey: Yeah, thanks so much. Great to be back on this show. My path has been a little circuitous in the fact that I graduated college at the beginning of 2000, 2001, moved into the creative industry, thinking I was getting into advertising, but the internet was really blossoming. And I spent most of the first 10 years of 2000s learning how to internet and how to make internet and how to play on … Back then, the entrepreneur thing wasn’t really … You played with the internet more because it was a toy and it was an exploration of what’s possible and communication. And it wasn’t all about money back then. It was really about, I don’t know, people finding-

Bjork Ostrom: Discovery.

Jason Glaspey: Discovery. Yeah.

Bjork Ostrom: Yeah.

Jason Glaspey: And so in that decade, I created a website called Unthirsty. It was a happy hour finder. In 2005, we were one of the first companies to use Google Maps API on a third party domain with user-generated content. We had thousands of happy hour locations. So you could say like, “Hey, on a Thursday and this zip code, what has drink specials, a Wi-Fi and a patio?” And you could find out, and it was great. But there was no monetization for that type of thing back then. Then in 2009, beginning of 2009, the Paleo movement was taking off. And so I created paleoplan.com, which was a really early meal planning tool that helped people follow the Paleo diet. And I ran that for about five years. Really fun to just go from … At the time, nobody was writing anything about Paleo. And so we really were able to hold a really great SEO advantage because as Paleo became popular, we were ranking number one for Paleo recipes.

And we worked with nutritionists and food experts and a recipe writer. So kind of had a little food thing for a while. And then after that, I actually had a bacon company where we sold bacon online. There were all

Bjork Ostrom: New- Didn’t know

Jason Glaspey: That.

Bjork Ostrom: Amazing.

Jason Glaspey: Bacon farms that we loved and we would get them to send us bacon, but they didn’t really have online distribution, 2011, I think. So we made a website, BACN.com. We would buy huge pound orders from small farms that would send us 50 pounds of bacon, but not one person a pound. And then we resold bacon, but we actually sold more T-shirts than anything. And we had great branding and fun stuff. And we would ship bacon in a freezer bag, but with the freezer pack then wrapped in butcher paper in a big stamp that said bacon tastes good. And we just had a lot of fun with … We were the online neighbor butcher store.

Bjork Ostrom: Yeah, cool.

Jason Glaspey: I actually got to write a book about that. We were doing a talk at a conference and people loved it. And so I wrote a book about creating the bacon business because we launched it in three weeks. And again, I think it was 2011. So I’ve been making things on the internet for a long time. The last 10 years has been a lot different. The internet’s changed. What it takes to be successful has changed. You have to really commit, I think, now before. There wasn’t enough noise happening on the internet. So if you made something interesting, you could get attention quickly, but obviously there’s a certain amount of attention now that has been gobbled up, and so it’s really hard to break out. And I think to lead into the AI conversation, that’s not getting easier, unfortunately, to stand.

Bjork Ostrom: Yeah. Lindsay and I are having this conversation two nights ago, and she was talking about this concept that she saw somebody talking about that really resonated with her. And it’s these two modes that we can be in. One is discovery mode. And I just pulled this up from, I don’t know who the original source is, but this is a Substack, Vatsap, V-A-T-S-A-P. It’s from a Google search. But they talk about discovery mode, turns on when you feel safe and detect opportunities. Sudden oncoming across a tree full of ripe mangoes just when you and your group feel hungry is kind of the descriptor uses. But in discovery mode, you’re happy, you’re sociable, you’re more eager for new experiences. And when I think back to early internet, one of the things I told Lindsay that was so fun about starting Pinch of Yum 15 years ago is it kind of felt like that.

It kind of felt like we’re in discovery mode and everything is amazing. And you publish something on the internet and then somebody comments on it or it goes viral on Pinterest and you see a spike in traffic and it’s like, wow, this is so cool, this internet thing. And as it matures, it becomes less discovery, it feels like. And potentially it becomes more defense mode. And it’s when you think that there’s a threat around the corner or your body gets flooded with stress hormones. And I think what’s happened, especially in the last three years, is there’s this really strange, not strange, there’s a divide. And I think the majority of people, especially people who have been working as creators for the last three plus years, four plus years, they’ve worked really hard, we are starting to go into defense mode. There’s a threat and we are feeling that threat.

And that threat is AI in the form of AI overviews from Google, in the form of easy content creation with photos and written text, you can snap your fingers and have a blog that’s created with similar recipes that are ripped off from another … I mean, we had this with Pinch of Yum, because it scrapes it, it repurposes it and republishes it. And so there’s a lot of defense mode right now in the world of creators and in the world of food creators. Also, I’m feeling some of the discover mode where it feels like early internet to me with the things that you are able to do and the opportunities that exist. And my hope for this podcast is to lean into some of the discover mode to say, what are the things that are really cool, that are incredibly helpful, that will be really fun, that will light you up, that will make your life easier in the world of AI right now?

And it’s not to ignore the things that are negative. It’s not to ignore the realities of AI overviews and all of the other variables that exist within the world of AI and are a department within the umbrella company of AI, but it’s to look at the ones that are discover mode that will allow us to have some really cool opportunities. And so it’s like opportunity-based view of it. Do you feel like that resonates with you and what you’re seeing when you think of those two ways that you were looking at the landscape, analyzing things, knowing that you have a background both in content creation as a creator yourself within content-based businesses, but then also somebody who’s seeing a lot of the benefits of the practical uses of AI?

Jason Glaspey: Yeah. I would say that one thing that’s different is right now there is a sense of discovery, but there’s also this sense of, well, at least with some people, I think there’s a sense of there’s a clock ticking and nobody really knows what the clock is or what happens when the time changes. But I feel like there is this sense that things are changing and they’re changing quickly. And I like to say the acceleration is accelerating, it feels like. The speed at which change is happening is actually-

Bjork Ostrom: Speeding up. …

Jason Glaspey: Increasing.

Bjork Ostrom: And that’s uncomfortable because you also don’t know what you’re speeding towards.

Jason Glaspey: Yeah, exactly. And so I think that there is a little less optimism as there was in 2005. And part of the lack of optimism is we’ve seen how corporations kind of took the soul and joy out of the internet. It is more addicting than ever, but less satisfying, I think.

Bjork Ostrom: Social media being a great example. Facebook, you’re using Instagram, you’re using any of these platforms. And I think we all know and recognize the potential downside of mental health and divides between people, noise. Sure.

Jason Glaspey: There’s so much coming at us. And the internet’s had 20 years of AB testing, so it’s pretty good at knowing what will stop our subconsciousness and pull our attention. And so we’re walking around just getting ripped in all these directions. So I’ll pause there to not get into the negative, but actually talk about that’s an advantage of where AI can come help and right away is slowing down. There’s so much happening, but AI can help you. I’m using AI to help me understand and manage the noise so I don’t have to cognitively be pulled in every direction.

Bjork Ostrom: Can you talk about that? Yeah. What’s an example?

Jason Glaspey: So I’m using OpenClaw and I’ve been in love with it since I downloaded it the first time. I think it was January 20th was when I installed it. So just over the four or five week mode.

Bjork Ostrom: I’m maybe a weekish in.

Jason Glaspey: Okay.

Bjork Ostrom: And I’m right there with you.

Jason Glaspey: It is unbelievable. And it’s scary and I know that I’m being a little loose with some of my security patches and whatnot.

Bjork Ostrom: Yeah. But it is … Can you talk about … So it’s OpenClaw, previously called Cloudbot, then MultBot, and then OpenClaw.

Jason Glaspey: And it was Claudis to begin with.

Bjork Ostrom: Sure. So can you talk about at a high level, what is OpenClaw and should everybody listening to the podcast right now be using it?

Jason Glaspey: Yeah. I’m going to answer that first. No. No, they shouldn’t. But it would be a huge mistake to not identify what this is demonstrating. Right now, OpenClaw is a … You can run it locally on your own machine or you can run it in a cloud, like on a virtual server somewhere. Right now, I have it running on a digital ocean droplet where I have a Docker container holding it and it has a bunch of other things it can do. But then I communicate it to it through a couple of different channels, and one of which is Telegram. You can also use Slack or Discord. I actually built my own dashboard so that I talked to it with my own dashboard, and that dashboard shows me a whole bunch of other things. But the key point that what OpenClaw really is doing is it verified a pattern, and that is always on twenty four seven, always accessible agent that can do certain things and that has some understanding of memory about you.

And when you start adding those things together, they start compounding with value. So you can give it access to your calendar and your email. And so you can say things like, “Check my email and check my calendar. What do I have going for the day?” Or even better, you have that set up as a routine that automatically occurs and then sends you a message. What’s different from that than some of the other systems that have been available is you could kind of pull this off in Notion, you could kind of pull it off and Google Suite if you hooked all the right things together. But in all those situations, it was somebody else’s decision and business model that was determining what you were allowed to do and how you were going to be able to do it. And what OpenCloud gives you is it’s open source.

You have your hands on all of the levers. There’s nothing black box setting that you don’t get to control. There’s no business model that’s forcing you to use it a certain way. I think that is all a big part of why this took off. And the developer behind it is a really well-beloved and appreciated, respected developer. So when he released it, he got a lot of attention right away and it matched a bunch of patterns that we’re looking to land.

Bjork Ostrom: So people who use WordPress would be able to identify with that idea of open source. And WordPress can be kind of confusing because there’s an open source WordPress and then there’s the wordpress.com. And for the vast majority of people who are listening to this podcast, you have a WordPress site that is hosted, that’s the part that you’re paying for, is hosted on another service. You pay for the hosting, but the actual software that you’re using, which is WordPress, you don’t have to pay for. You just download it, it’s integrated into hosting. There’s all these different ways that you can interact with it. OpenClaw is very similar in that the code itself, you don’t have to pay for it, but you have to figure out how to host it, and there’s different ways to do it. I have it set up on a Mac Mini that’s on my desk.

And the interesting thing is I don’t have a monitor on it, I don’t have a keyboard on it, I don’t have a mouse on it because all that I need is for that Mac Mini to be connected to the internet and always on. And then the amazing thing is I can talk to it and it can respond back to me. And this is where it feels very different, is that I can then ask it to actively do things. And this is where I’d be interested to hear about some of the things that you’ve done. I’ll share a quick one as a concrete example in our world. So for a long time, we’ve been trying to figure out how do we organize our photos and videos, right? We have 500 videos, we have thousands of photos, over a decade plus of content creation. Right now, they’re in iCloud, so they can easily be accessed by all of our team members who have a phone that’s connected via the same iCloud account and they can upload stuff as needed.

And so I asked Lindsay, this was yesterday, and the other interesting piece with this is I’m doing this all via voice and I have a walking desk, not on it right now, but it feels very future where you’re walking, talking to the computer and then tasking it to do things.

But I asked Lindsay, “Okay, what’s your kind of vision for this? ” She sent some texts back. I took a screenshot of that and then I just said, “Hey, I’m working on a project. I have a massive amount of media and I’m sending this via Telegram. It’s video content, photo content. I’d like to start with the video content and organize some things. Here’s the message from Lindsay. Please take some time to thoroughly review this and let me know if you have any follow-up questions to get clarity on it. ” And its response is like, “Hey, this is a great project, very doable with AI.” And it gives you a recap of the goal, the problem, and then it says, “Can AI do this? ” Yes. The approach depends on the answers to these questions. It gives me some questions and then it says, “Here’s how I’d approach this.

“ And it gave me the list of how it would move forward with it. And it says, ”The hardest part is just getting access to the media files. How do you want to get those to me? “ It’s like, ”Oh, wow. So my job is just making sure you have access to those, and then it gets to work.” And that piece, the gets to work piece is what is really incredible. And so that’s a use case of how we did it. One more real quick one for the long … This is a very Minnesota thing, but for the longest time, one of the things that has just been a persistent little program that runs in my brain that I’m thinking about is what is the temperature in the winter and what is our humidity set at? Because if the temperature gets too low and the humidity for our in- home humidifier is too high, all of the windows get condensation on them.

And so you have to be adjusting the humidity down based on the temperature. And so I asked, I was like, “Hey, here’s the problem I’m having.” And it was like, “Hey, we can do an integration. You pay a $5 fee to Google to get access to the device, and then there’s an API call. I’ll check the weather every four hours, and if I have the API information to that thermostat, I’ll adjust the humidity based on this skill that I’ve created.” And it’s like, that’s a small thing, but for my brain to be able to release that for every winter, for the rest of my life will feel really good. So can you talk about maybe some of the use cases that you’ve seen with it to help people understand just the capabilities of what this can do?

Jason Glaspey: Yeah. I think it’s interesting to highlight that AI could do a lot of these things before, but it was stuck inside the web app for Claude or an OpenAI download tool. And sometimes those apps could even have access to your desktop and write files and even run a script. But what OpenClaude does is it just creates this complete wrapper around it’s always available, it can run things at any time, it can set a cron so that two o’clock in the morning it wakes up and does something, it can be responsive and it has a disc that it can write to so that as it gets information, it can keep that information close hand.

Bjork Ostrom: Meaning the actual compute … In the case of the Mac Mini, I think it’s easiest to visualize it. Just like you save a file on your computer, when you ask it to do something, it saves that to a file, which is on the computer. And so what’s interesting is that you can switch … Right now, I’m using Claude to run it, but you could switch to ChatGPT, OpenAI, and it’s switching the reasoning, it’s not switching the memory, which is also really an incredible feature of it.

Jason Glaspey: We coupled from the corporate business model of come to our website and use AI and we will make it easy for you to now

You have that same technology, but it’s available on your terms. And so some of the things that I’ve been working on to make little automations is the most exciting thing that I’ve done is whenever you have a conversation with AI, it is stored as a session. I said this, it’s replied, I replied, it used a tool, it figured out some things, it did some thinking, and then it replied, and all that’s stored in a file on my computer. And it turns out that those conversations really do represent the work that I do now because I very rarely go to a website, log in and click their interface. I’m reducing the number of websites and tools that I use to begin with. And now when I use them most often, I’m connecting via an API. So I’m sending a request from my computer to a services computer with a special API token that’s like a password.

And an example of that is I can send something to GitHub and say, “Put this on my website.” And I am not going to the GitHub.com website, I am sending a signal to GitHub with all the information for it to do what I want to do next.

Bjork Ostrom: And are you doing that all from within Telegram?

Jason Glaspey: No, I’m doing that from a number of places, but I do do it through Telegram.

Bjork Ostrom: Sure.

Jason Glaspey: But like I said, I also-

Bjork Ostrom: You could do it through Slack as well.

Jason Glaspey: Yeah. And I’ve built a custom dashboard that actually, it’s a website that is storing all of my sessions and allows me to choose them. And I’ve created other agents, which is just, I can have my primary agent and then I have a chief of staff agent. And the chief of staff agent is different than my primary agent in that it has a different system instruction. It’s called an agents.md file. So that explains who you are and what you do and where the tools are that I want you to use and how to use those tools. And it has its own memory. So this agent, my chief of staff, behaves like a chief of staff. Then I have a front end designer and they have my style guides and our code and how we do our front end. And so when I talk to my front end designer, I’m only talking about front end things.

And you can do that with OpenClaw. I built a dashboard that allows me to switch in and out of them more easily and all of these conversations are being saved to my server. Then I have a reminder every five, 10 minutes, I think it looks at my sessions and have I added anything to any existing sessions? Has a conversation continued? Or have I opened and started new sessions? And in each one, it looks at what are the files that the agent is reading to determine which project is this in?

What did I do?

Bjork Ostrom: This is a session within Cloud or ChatGPT?

Jason Glaspey: Either.

Bjork Ostrom: Sure.

Jason Glaspey: So right now any session-

Bjork Ostrom: But when you refer to a session, it’s like a chat thread within whatever LLM you’re using, my guess is majority Cloud or ChatGPT.

Jason Glaspey: Yeah. Especially with OpenClaude, I’m primarily using Claude.

Bjork Ostrom: Yep.

Jason Glaspey: And all those sessions are stored. And so now I have a system that opens my session folder, looks to see if there’s anything new, if there’s anything new, it analyzes it and understands what was the session about. What project was I working on?

Bjork Ostrom: And it’s storing that in- Did I

Jason Glaspey: Accomplish anything? And then it goes and it updates my project file for that project. So it’s retroactively managing my project without me remembering I did that task, click, I checked it off the list.

Bjork Ostrom: Yep. Is that like literally you have a list of things that you’re working on for a project and it’s monitoring those? Or are you just saying hypothetically it’s kind of like keeping track of what’s happening with the project?

Jason Glaspey: A little bit of both, depending on the project. Some projects are really linear and it’s like, “Do this task.” And in that case, I just need to know that I’ve done it and my AI can look at my session and identify that I did it and then go check the box.

Bjork Ostrom: Sometimes

Jason Glaspey: It’s much more flow where I’m just like, “I think this is what I want to do next and I spend some time researching it. ” Well, then it’s like, “Okay, you found all this information. We stored that in your resources for that project so that when you need that information later, it’s not lost.” But I didn’t have to go save it. I didn’t tell it, “This is good stuff, put it in my file.” It just said, “You were doing work, you found this good information, we stored it for you. ”

Bjork Ostrom: And that comes from you giving it the directive to do that. And I think that’s the really important piece with this to help people understand my primary use case at this point weekend is just interacting with via Telegram and building things that are called skills, but a really practical skill, an example would be you could essentially look at any opportunity that you have with the things that you’re doing on repeat and saying, “Is there a way that I wouldn’t have to do this and AI could do this? ” And an example would be, and the context for all of this is like, how do we get to do as much of the human stuff as possible and as little of the non-human stuff as possible? And in a world where AI will continually get better and we are content creators, it is more important every day that what we are actually doing is creating compelling human-forward content and doing less of the things that computers can do that don’t matter if we the human do them.

And so an example would be, let’s say you have a blog post and one of the things you do is you want to audit that post to make sure it checks all of the things that you consider to be best practices. One of the things that you could do with OpenClaw that is different than a tool that like ChatGPT or Claude, which we are more used to working with in an interactive mode as opposed to a proactive mode where you tell to do something proactively and it does it. So with OpenClaude you could say, “Hey, monitor my website and anytime that a new post is posted and it’s important to know you’re saying all of this just in natural language via something that feels like messages on your iPhone.

Anytime that you see a new post on my site, I want you to follow these steps to audit it to make sure it includes all of these things that I deem to be important.“ You define that list and that list could also be co-created with OpenClaw. ”Hey, what are the other things that I’m missing from an SEO perspective? Oh, could you also do a review of typos and see if there’s any grammatical errors?“ All of these things can be built in and then it can automatically do that, which is to me, one of the amazing things is that automation and proactive piece with this. And then once you start to do that, suddenly you’re like, ”Oh my goodness, there are all of these opportunities in my life to do these things that doesn’t matter if I do them because they’re behind the scenes, it’s not public, it’s not a piece of content.” And I would assume the first part is automating the things that you are doing that don’t matter if you’re doing them.

And then you start to look ahead and say like, “Okay, what are some of the cool things that I’m not doing that I could be doing that would have an impact?” So what did it look like for you as you kind of started to step into it and maybe if you put on the hat of a creator, What are some of the things if somebody does say, “Hey, I want to set up OpenClaw, number one, how hard is it to do that? ” And then number two, what are some of the really easy first things that you can do as you step into using a tool like this?

Jason Glaspey: Yeah. To answer the how easy is it to set up Open Claw? If you’re comfortable geeking, you can have a version working in a half an hour. But the danger is you will never want to stop playing with it because I use it to help me refine the system of using it. And there’s a whole new paradigm there.

Bjork Ostrom: For sure.

Jason Glaspey: I can be talking to it and I’m like, “I want you to be able to do this. I can’t do that. ” Okay, well, figure out how to do that. Okay, now I can do that.

Bjork Ostrom: Yeah. It’ll say, “I can’t do that in my current state.” And then it will say, “Here’s how I could do that oftentimes.” And then the options are either, “Okay, go and do that, ” or you give it the access it needs to do that.

Jason Glaspey: And one of the things that is kind of tricky to take a tiny tangent is that you mentioned earlier, what would it look like to do this? And it’s like, blah, blah, blah, blah. Would you like me to get started? And man, what a disorienting question. Most of us are not ready to go from, I had this idea to now someone’s like, “Can I just make that happen and it’ll be about four minutes?” Give me four minutes and I’ll figure this out for you. And that level of speed is emotionally dysregulating, I have found.

Bjork Ostrom: I’m

Jason Glaspey: Like, “Wow, I just thought of this. Give me a minute.”

Bjork Ostrom: Totally, totally.

Jason Glaspey: And it also is one of those things where you can just be sprinting down rabbit holes because you’re making so much progress so fast.

Bjork Ostrom: For sure. And there’s something happening. I’m sure it’s a dopamine type response around- It’s like

Jason Glaspey: Being a slot machine. Oh my God, I want it.

Bjork Ostrom: Yes. I had this thought I wanted it to exist. And between 9:00 and 9:45 when we sit down and watch the start of Survivor, I’ve created the automation for adjusting the humidity in my room. And it’s sent me a text and said, “Hey, I shifted it down a little bit because it’s kind of cold.” And that is such a great feeling to have that, but it’s also disorienting. I think the other piece with it … Well, finish that thought and then I have a follow-up question.

Jason Glaspey: The second thought I would say is, man, AI is so eager to help that sometimes it pulls you into ideas that you don’t really need to be doing. So learning some discipline on like, “Actually, I know we could do that, but I only have so much energy and I don’t want to be managing that later. So let’s not do that is an important skill to develop.” As well as I have already observed in my own habits, there is a desire to build it all, build the perfect system you’ve waited your whole life for. Instead of like, how am I going to tweak what I’m already doing? How am I going to add a little bit more nuance? How am I going to get a little bit better results with a little less effort? Compounding systems that tweak which you already know and are good at have a higher likelihood of sticking and truly helping you.

Bjork Ostrom: That’s great.

Jason Glaspey: People not to just go hog wild and start building the perfect system. Start using it in ways that really matter and start clearing those paths for automation and getting yourself out of the way, getting yourself used to not having to object to certain tasks. That’s enough work that it takes a little bit of time.

And I would say, you touched on it, there are things that humans can do that AIs are terrible at, but there’s also things that you can do. Everyone has an unfair advantage that if they really knew how to exploit it, they’d be unstoppable. Maybe it’s because you live near the beach and you can get the best beach photos. Maybe it’s because you live on the mountain and you get the best mountain photos. It doesn’t matter. It doesn’t matter if you’re a food blogger, you have access to this one place and nobody else does. There’s something about what you have that is an unfair advantage. And how do you pour gas on that with AI? And there are parts of you that are just suffering because you hate them and they’re not your strength and you will never do them. So what would it look like if AI gave you a B minus version of that automatically?

Bjork Ostrom: Versus you either not doing it or you suffering through it, doing it, and then also getting a B minus. Or even if it was a B+. Yeah. And

Jason Glaspey: Then avoiding other work that you could be doing because you’re emotionally drained.

Bjork Ostrom: Yeah.

Jason Glaspey: So I really look to AI to be like, how can it help me lessen the administrative burden of life and work?

Bjork Ostrom: Yes, totally. And I think that’s the discovery piece that is so exciting where where’s the opportunity? And again, I think it’s really important to point out we’re not ignoring the fact that there are significant threats with AI, security, which we can talk about. It’s so all encompassing as a term and it means so many different things to so many people and carries a different weight for so many people. And I think it’s important to acknowledge that. 100% it’s true. It’s negatively impacted hundreds, thousands of probably even just food creators. And so it’s not ignoring that. We feel the impacts of that. The question is just like, okay, what are we going to do with it and what are we going to do about it? And I think one of the opportunities, which we’re talking about today, is to look at those things that can be outsourced off of your desk to somebody else’s.

Previously, that might’ve been somebody that you would’ve had to have hired and you might not have had the budget to do that. You might not have had the resources to bring somebody on and help with a certain thing, but now you can start to think about how can you bring those people on in the form of something like OpenClaw. And I want to hear about, for somebody who does decide to go about setting this up, and let’s say they get past the technical burden. One tip that I would give if you do decide to do this is the only way that I got through the setup process was using AI to troubleshoot the setup process. 100%. And it’s one of the things I’m trying to get better at is anytime that I get stuck on a thing, especially a technical thing, it’s then going to AI with a screenshot, with a description and saying, “Hey, this is the message I got when trying to set up OpenClaw.

What do I do? “ ”Okay, paste this, put it into terminal, here’s what you need to do. “ And it’s like that’s how I got through it because I did hit a lot of walls. And actually while we’re recording this, my friend Preston messaged me because we’ve been going back and forth and he said a screenshot of the classic, ”Hey, I just came online. Who am I? Who are you? “ Message that it sends after a bunch of messages he sent that were like, ”Hello, hello, hello, hello, no response.“ And then you get the response. And what he said was, ”Wow, that took a lot of wrangling.” So I think acknowledging that for anybody getting into it, if you do feel that you’re not alone, you just have to use AI, Cloud or ChatGPT, screenshots, descriptions to help you get through the process. Eventually you get that message, which is just such a weird message.

“Hey, I just came online. Who am I? Who are you? Wow, what a weird thing. ”And then you say,“ Here’s who I am, ”and then you name it, here’s who you are, and then you start to get to work on stuff. And it’s like, that’s so strange. But once you get through that set of process, what are some of the first things that you could do? You’re a creator, you talk about this idea of offloading those administrative tasks. I can share some ideas, but what would you say for a creator could be some categories, some specific examples of things that they could look at doing?

Jason Glaspey: Yeah. Man, I’ll say first of all, what I do to install something like this is I use ClaudeCode and/or a tool like Cursor. I actually run CloudCode inside Cursor a lot. The advantage of one of those two tools, and either of them would work or WinServe or many others, but I’m going to just use those as examples. They have the ability to see your desktop write to your desktop, so write files and run terminal commands for you. And so when you’re trying to install ClodCode or OpenClaw, I opened CloudCode and said,“ Help me install this. ”And it did 98% of the effort because-

Bjork Ostrom: Can you say that again? Which program is able to see your desktop?

Jason Glaspey: So ClaudeCode, it is a programming tool. And you could also probably do this with OpenWork or Claude, I’m sorry, not OpenWork, Claude Cowork. But ClaudeCode is a tool that you install in your terminal, and then you use it in your terminal. But it basically brings Claude into your terminal, and then Claude can sit there, run commands in your terminal, access your desktop and do things. And you can open it just in one folder, you can open it in your desktop route, like your user folder. That’s what I do. I just opened my Jason Glaspey Mac route user folder and said,” I’m trying to install OpenClaw. Here’s the URL to OpenClaw. Here’s the URL to the docs for OpenClaw. Let’s figure this out together.

Bjork Ostrom: “And

Jason Glaspey: So- And I’m like, ask me questions when there are configuration options.

Bjork Ostrom: So I’ve switched, I’m using Claude more than ChatGPT now, probably within the last two to three weeks. 95% of my use when I’m in Cloud itself is within the chat interface. Well, no, 5%, I say 95, 5% is in cowork when it’s something that’s a little bit more involved. And these are just the tabs that I’m switching between. And then there’s ClaudeCode and I’m never in CloudCode. So can you describe, again, I want to make sure that I understand that. When you’re saying that it’s able to see your screen, that’s within the terminal application that you’re using or that’s within CloudCode the desktop app?

Jason Glaspey: It can see your terminal.

Bjork Ostrom: Oh, see Terminal. I see. Okay. Yeah, yeah, yeah.

Jason Glaspey: Now I can. There are ways to allow all of those devices to actually see my Mac or a browser, but those are slightly different than what I’m talking about.

Bjork Ostrom: I see. I see. So it’s able to hook into terminal and you’re able to work together within Terminal. One of the things that’s potentially a little bit intimidating is in the setup process for the Mac Mini, you are using terminal to set it up, which is essentially a way to interact with code with your computer. I’m not a big terminal user. It was a lot of copy and paste for me, copy, paste and screenshots. And for you, being more technical, you’re hooked into terminal and so there’s a little bit more of a collaboration happening there within terminal. Is it asking you before it’s executing if it’s okay to execute a thing or is it also then executing on it without asking?

Jason Glaspey: It depends on what it’s executing. So by default, clog code has certain privileges, but it’ll ask you like, “Hey, can I read this file?” And you’re like, “You can read all the files.” I was like, “Can I edit this file?” I’m like, “You can edit all the files. Can I run this command?” I’m like, “Only when I tell you.

Bjork Ostrom: ” Sure.

Jason Glaspey: You can’t just run that command anytime you want. You have to ask me every time you want to run it and I’ll hit approve and then that command can get run.

Bjork Ostrom: And then so you move

Jason Glaspey: This folder.

Bjork Ostrom: You

Jason Glaspey: Don’t want it to just be able to delete a folder because it got an idea.

Bjork Ostrom: So when you were setting up OpenClaw, you gave integrated cloud claude code into terminal, and then you essentially asked it to install OpenClaw.

Jason Glaspey: Correct.

Bjork Ostrom: And the 98% that it did on its own is just running through it. The 2% that you had to be involved with was what?

Jason Glaspey: Asking a couple questions, making a few decisions, probably moving a couple of things manually, but really it took all of … I didn’t have to copy and paste. It would do the copying and pasting for me and so it really was like, “Okay, I’m going to install it. Where do you want to install it in the normal spot or do you have a special spot?” And I’m like, “Normal spot. Great. It’s installed. Okay. Now, do you want to have this setup? Do you want to do this? Which LLM do you want to use? I want to use this one.

Bjork Ostrom: Great.

Jason Glaspey: We need to get an API key. Do you need help doing that? “ I’m like, ”Nope, I got one here. Great.” And it was walking me through the process step by step. And just to say, there are also companies out there who are selling one click installs of OpenClaw on a desktop or not on a remote server.

Bjork Ostrom: Virtual.

Jason Glaspey: I haven’t vetted any of them. Some of them I’m sure are very, very awesome and safe. There are also companies like DigitalOcean and Railway and Vercel that are allowing you to do one click installs. Those I think are very safe. And in every one of those things, you might lose a little bit of your own access to a setting by giving other people access, but it simplifies the maintenance and updates. So there’s benefits. Again, I would say for right now, let’s see, February 25th, 2026, if you don’t like the idea of opening terminal and playing, it might be too early for OpenClaw.

Bjork Ostrom: For people who are trying to decide, should I do this or should I not?

Jason Glaspey: Yeah. And I would say that it might be a week or two before a significant shift comes out. They are releasing new versions of OpenClaw every three days and they are not-

Bjork Ostrom: I would not be surprised if within a month there’s a version that you can install in a way that we would install Microsoft Word.

Jason Glaspey: Yeah. And when you get those, again, there will be trade-offs and it’s kind of like going to … I’m putting my website on WordPress.com. They’re going to limit which plugins you can use because they’re protecting a giant system that includes your website.

Bjork Ostrom: Right. WordPress.com versus the open source WordPress.

Jason Glaspey: Whereas if I download wordpress.org and I put it on my own server, let’s say Bluehost, Bluehost will let me do whatever I want, even if it blows me up. And so I think that cowork right now, Claude Cowork is the closest I have seen to that you can just install it and it can do awesome things. Some of the things that Claude Cowork can’t do by default is automation on a schedule, but you can have Cloud Cowork build you a script on your Mac or Windows that will still do something on a schedule. So you can kind of disconnect some of what OpenClaude gives you and build subsystems without it if there was something specific.

Bjork Ostrom: If you are kind of early stage and feel like you’re up for the challenge, I feel like the default install of OpenClaw gets you to that point where you get that message, like my friend Preston, where it’s like- And then what? “Hey, I just came online. Who am I? Who are you? ” And once you’re to that point, then you can start to have some of these conversations. And the thing that I would think about is, and the filter that I’ve been trying to use is anytime I come up against a thing that I’m trying to do, the first step that I try and think of is, is there a way that I don’t have to have this in my brain and in this case, open clock and have it in its brain specific example, and maybe we can trade back and forth some really basic ideas of how you could be using it. One of the instances that I’m going to set up is checking my calendar for any flights or travel that I’m doing from today until the next six months and looking to see if I also have flights booked for that.

There was one time last year I had a conference I was going to, it was like three days before I was like, “I didn’t book flights for this and now they’re $9 million.” And that’s a bummer because it was just an oversight. And that would be an example of something that once I have that problem, I should never have to have that problem again. So I can create through natural language, sending a message just like you would to your friend, except instead of messages on your iPhone or WhatsApp, you’re using Telegram. You could also use the Messages app or WhatsApp, but Telegram is default. I would create a voice message and I would probably use voice to text like Super Whisper and I’d say, “Here’s the problem I’m having. How can you help me ensure that that doesn’t happen again?” And so you’re not even having to come with the solution, you’re coming with the problem asking what the solution would be.

And from there, then you would partner on saying like, “Okay, you can get read access to the calendar. You can get, here’s when I would want you to check,” and you kind of build out the skill. So that would be one example. What’s another example in the first week maybe that you set up or you could think of a creator setting up?

Jason Glaspey: I think skills in general, to make sure your audience understands, a skill is a concept that Anthropic is using an AI agent land. So it’s not a generic word in this situation. It’s not

Bjork Ostrom: Like juggling.

Jason Glaspey: Yeah, it’s a terminology. And so a skill allows you to bring context and execution information to a task. And we don’t need to get into the technical defaults, but what skills allow you to do is have a lot of them and it doesn’t flood your context what an AI can remember at one given time with a bunch of information about skills that it doesn’t need. So it has progressive disclosure as it understands it needs this skill. It learns more about it without having to carry everything that skill provides. An example is you might have a skill for creating a PowerPoint. You might have a skill for creating an invoice. You might have a skill for publishing a blog post. One thing that I’ve seen people do, and that I think would be a pretty powerful tool is to like, if you are a publisher and let’s say you’re using a content management system like WordPress, it’s really easy now to automate the publishing of a document.

So you could have a Google Doc or a markdown file that has your blog post and you could work with OpenClaw to build your custom WordPress publishing skill. And that skill might include multiple steps. So first step might be write alt text for all of the images, like do the grammar check and whatnot. Then you might have a thing like identify five related blog posts and create cross-linking.

Bjork Ostrom: Using my site map.

Jason Glaspey: Yeah. Update my site map, look for cross-linking opportunities, so make sure I’m linking out and also identify which existing blog posts should be linking in. And that’s-

Bjork Ostrom: And with all of these, I think it’s important to point out, it’s not creating the content for you. There are some people who have a blog post, it’s created by AI. We’re not there yet. We’re not interested in doing that at that point. We want to have very human feeling content. And I think that’s a strategic advantage for us is like having high touch words that are actually fingers on a keyboard, just personal preference. But there are a lot of things that then go into that like you’re talking about that don’t matter, like generating the alt text or looking for highly related blog posts and then figuring out, okay, where am I going to link those or grammar check? All of those are examples of things that will take some cognitive lift off of your shoulders. So I think that’s a great example. Anything you’d add to that?

Jason Glaspey: There can be some cleanups that you might do. And in this case, let’s take a food blog site. You might choose that from now on you want to organize all of your ingredients. One of the things that I do whenever I go shopping is now I give my shopping list to AI and I ask it to organize by section of the grocery store.

So all of my dairy is in one spot. All of my meat is in one spot and it makes shopping so much easier to have my shopping list organized. So you could even like, let’s just say that’s something that you want to add. So it’s like when you’re processing my blog, make sure you look at the recipes and identify the ingredient list as probably in your cupboard, dairy section, dried goods section, break it down so that it transfers more easily to my own shopping list. You could be adding things like that that would take, God, that takes like two to 20 minutes every blog, every recipe. And now it happens in a fraction of a second. So what are some things that you could do to enhance what you’re already doing?

Bjork Ostrom: That’s great.

Jason Glaspey: One of the things that I like to talk about is we take the long cut with content now. We don’t take shortcuts. We take extra steps because they’re almost immediate and they get us from zero to B minus.

Bjork Ostrom: What do you mean by that?

Jason Glaspey: An example is, let’s say I spend all my time writing a really good article and I’m emotionally tired now. One thing that an ideal situation be is like double check the header, the title. Double check the entry. Is the first two sentences interesting? Does the article actually provide the promise that was in the title and in the opening? Can I leave the final paragraph, point them somewhere more interesting? Can I hand off them? If they just finished, where do I want to send them? What is the right place to hand off to? And what does that look like? These custom little things that … So now you might have a system like when I finish a blog post, I hit blah, and then it runs it and it comes back and it’s like, check, check, check, but your subject line is still really uninspiring. Or you could add this call to action at the bottom and then it’s like, yeah, do it.

It didn’t take me but a half a second to process it and then to decide if I want to follow its advice. So looking for those places to make a longer process more efficient.

Bjork Ostrom: Yeah, that’s great.

Jason Glaspey: Also, again, trying not to build a whole system at once, what is the next step that I would want to do and can I get that started? So for instance, maybe it’s, okay, now I want to go and … Man, I’m trying to think of something on the fly and I can’t. The next step after research is flight checks. “Hey, do I want to go to this country? Here’s what the prices would be. So now I don’t have to go look at Kayak. I have a general idea of what the next step would be so that that next step actually feels like it’s a downhill step, not an uphill step from this.

Bjork Ostrom: “That you have to power up the cognitive effort again. An example might be in the world of publishing, maybe you write a blog post and you have a process that is to generate from the content you’ve written, short snippets that would work well across social. That could be an example of a next step within the context of a blog post. I’ll share one more example of an immediate use that we had. For a long time, I’ve had this idea of I want to create essentially a database of people I know. So when I have a conversation with them, I can make a mental note. So it’s kind of like a second brain for relationships. And then if I’m seeing them again two years later, my dentist does this really well. He walks in, he’s like, ” Bjork, how’s it going? Your kids are six and eight now, right?

“Th’re actually five and seven, but if I saw him later, this is what he would say. I’m like, ” Man. “And then eventually I was like, ” Well, I think he probably just has notes and he just refers to him every time before he goes in and sees a patient. “But it still feels good, even though I kind of know that’s what’s happening and I want to be that. And so I’ve long thought like, ” I should do this. “And I just never did it. And because it took time, I experimented with some software, but then I didn’t update it. So then I was like, ” Here’s what I’m trying to do in Telegram, natural language, send it over. Hey, that makes sense. Here’s what I need in Notion if you’re going to put it in there. Okay, give it access. “And then I created a people page, it pulls it up and it’s like, ” Hey, it doesn’t look like there’s actually anything here.

“And I was like, ” Yeah, I haven’t done anything. “It’s like, ” Oh, what do you want me to do? “I was like, ” Maybe create categories, let’s see, work, family, friends, and then it’s like, okay, “and it creates them, and then it’s like, ” Who do you want me to add? Here’s some suggestions. “I was like, ” Okay, yeah, add those and add these two people just to see how it works, and then it does it. “And so now for me, what that’s done is it allows me to have a frictionless way to accomplish something that I’ve really wanted to accomplish, which is taking notes on conversations that I’ve had. So when I talk to people, I can have a reference point and know better like, ” Hey, what are your kids’ names or what were you up to last? “Within the context of business, I think that could apply because you could do that with vendors, you could do it with brand context if you do sponsored content and then here’s where it gets really cool and you can start to build the picture.

You could also say,“ And I want you to every three months pre-draft an email to this contact checking in on the holiday that is six months from now and provide some context around the value that we could add if we have a sponsored relationship. ”And you don’t just press send on that, you revise it, you kind of get it to the point where you feel best about it before sending it off, but that’s an example of the cognitive lift that can result in creating some systems around this stuff that allows us to do the things that are more human. Again, it doesn’t matter, nobody in the world is going to, their life isn’t going to be made better if you’re sending out 20 sponsored content emails in a day, that’s going to help your business, that’s going to help you, but it’s not going to help your business within the context of awareness, of eyeballs.

That’s just the grunt work of the business. And so as much as you can remove that grunt work, you’re going to benefit from it. So any others that you can think of just to inspire some ideas around how people could be thinking about this?

Jason Glaspey: Yeah. We have a client that we … So I have a company Fleet of Geniuses. We do AI consulting and help people basically figure out AI. And one of our clients, what we’re doing is they have a product that they are a reseller of things. I’m not going to get into details, but they get something and they’re like, “ Oh, we need to reach out to the right people to see if those people want to buy this. ”We created a system that was like, ” Okay, which company is this company in and what are their genres and details about the industry? Where do those companies live? Where would we find information about them? Are there lists? Are there organizations that track that? Are there conventions that these companies would be going to? We could look at lists that go to the conventions, attendee lists. And so this AI goes out and just accumulates raw data about every company that might be relevant.

Then it goes to a second phase where it analyzes, are these companies still active? Are they still open for business? Have they been acquired? Is there any news on their acquisition properties To vet, are these still relevant as we thought? “And then it looks at the recent news and our recent news and where is Their overlap and what’s been happening over the past year. And then it crafts a draft email that outlines everything that needs to be shared with the details of the sale, the opportunity, why we think it’s a good fit for them and how we understand it fits into their thesis.

And then it drafts that and puts it in and it doesn’t send it, but then the right person can go in and just look through their drafts and be like, send, send, send, send, send. But it’s a really repeatable system. And I’m actually helping my wife set it up for one of her friends to do podcast outreach. Like, “Hey, you have this podcast. I’m a speaker. This is what I’m talking about right now. These are people you’ve had on your podcast recently that are in the same field as me. This is why I think what I have to bring to your podcast would be really beneficial. Here’s how to follow up.” And so the steps would be all the same. Who are the right podcasters? Are those podcasts still active? Have they had anyone on like me lately? And then what does a custom introduction look like with this template and what’s relevant to them?

And that process is not a difficult one to make. Those are each steps that an AI might run for two to 20 minutes and then come back with a list. And then you do a vetting and you would have to describe how you want it to vet.

But even if you were loose in those explanations in the first few attempts, it would be able to figure things out. And I mean, you could basically, if you took the last five minutes of this podcast and just gave that to OpenClaw and say, “Let’s build this.” It would be able to walk you through. “Okay, well, here’s three different ways we can do it. Which one sounds better to you? I’m going to need access to this and this and this. ”And I mean, it could be done in an hour or two, at least to the point where you could run it and start getting something like, “ Yeah, I like that. I’m going to go ahead and reach out to that person. I never would’ve thought of that otherwise. This was garbage, this one’s garbage, this one doesn’t match. ”But in two hours, you might have four emails sent off that otherwise you would’ve spent the whole time figuring out what do I do?

Bjork Ostrom: It’s a huge opportunity for us. It’s one of the instances I’ve thought for Food Blogger Pro is that building that podcast list. And right now it’s manual. We’re thinking of it, we’re sharing ideas, which still happens and is still important, but there’s 99.9% of the really high quality people that we could have on we haven’t thought of. I’m lucky enough to have a connection with you. I’m lucky enough to have awareness of a lot of other really smart people, but I don’t have awareness of the 99.9% of the other people who could come on and be massively helpful for this audience. And I could spend a lot of time roaming the internet, looking at other podcasts, interviews that have happened, sending out those emails. But that’s a great example of one where you could surface those and you could create it and it might not nail it right away, but over time you could say,” Hey, here’s why this one was really great.

I just want you to know that for context, and here’s why this one wasn’t great. “This was actually somebody who’s in fashion. It’s a little bit outside of what we do, so we want to stick to food. So there are so many opportunities here. We could have this conversation for three hours. Would love to have you back on because this world is changing quickly, but I want to hear a little bit about what you’re up to, how you’re working with companies right now because you have such a deep expertise and there’s so much value with this. You talked about your company a little bit. Talk about what that looks like, how you work with folks, how people could connect with you, Jason, follow along with what you’re up to. What does that look like?

Jason Glaspey: Yeah. I don’t post much on Twitter these days, but I’m Jason Vlaspy. Fog Fleet of Geniuses is our company. And right now what we’re doing is we’re helping companies figure out what it means to be AI enabled. And what a lot of that is, is we’re building systems that understand how to process bigger systems. What we’re finding is really it’s just, again, build systems that compound. And that’s what we’re trying to do both internally and externally. How does AI actually help the world? How do we not just leave this to the capitalists who are ultimately going to look for opportunities to exploit, but how do we actually use AI in ways that make the world better, that make our communities operate more smoothly? How do we remove the administrative burden of just being alive?

And so a lot of what I’m talking about right now is that work and I’m building open source software for our community. I love connecting with people who are trying to take AI beyond just how can I make a little bit more money, but into like, this is a way to rechange how we … It’s an opportunity to reshape the world. It’s an opportunity to change how we interact with people. I think there’s an opportunity to go from hierarchy towards networks and networks have a much better way to take care of the individual in that network than sometimes the top down. So we’re looking for those types of opportunities to impact the world and to shape what comes next because the next two to five years are going to be wild. And everybody is off balance. Everybody is a little disconnected and disoriented. Nobody has a magic ball.

And so how do we be patient and diligent and thoughtful in these moments, but also make sure we’re really driving towards a world change that we can all be excited about, not that just benefits a few people who have the best position when it all started.

Bjork Ostrom: The two things that I love about that, we talk about compounding a lot, the psychology of over time, how do you do little things that make a big difference? And what you wrapped that up in was not just compounding, but compounding towards good, compounding towards a better life. And I think it’s a great example of discovery mode as it relates to AI, which is what is the opportunity to make my life or other lives better because of this? There’s threats, there’s very real impact on businesses. A lot of it is negative in the world of content. We’ve seen that, but there’s also lots of really cool opportunities. And how do we take those opportunities and compound them towards good for us and others? I think that’s a great note to end on. If people want to follow up with you, Jason, what’s the best way to do that?

Jason Glaspey: You can reach out [email protected] or Twitter’s great, Jason Glaspey, or they can contact you, Bjork. You can be our intermediary.

Bjork Ostrom: I love it. I’ll have an AI agent that will make the introduction. Perfect. Actually won’t. I won’t, but maybe. I’ll

Jason Glaspey: Have an AI agent that won’t hear it.

Bjork Ostrom: Thanks so much for coming on, Jason. Appreciate it.

Jason Glaspey: My pleasure. Love you. You’re awesome.

Emily Walker: Hey there, this is Emily. Thank you so much for listening to that episode of the podcast. As always, if you enjoyed the episode, please leave a rating or review wherever you listen to podcasts and share the episode with your community. Next week on the podcast, Bjork is interviewing Carrie Forrest from Clean Eating Kitchen. Make sure to tune in next Tuesday for that episode, and in the meantime, have a wonderful week.