Show Notes
Back in January, OpenClaw made AI agents accessible using existing subscriptions. Then on Good Friday, Anthropic sent an email: your subscription won't work for third-party tools anymore. Starting tomorrow. People scrambled to cheaper models—and found out they weren't as good. Workflows broke. Social media exploded with complaints.
The revelation: we handed our entire businesses to some other company. In real estate, there's a rule—don't build on rented land. That's exactly what we did. We built workflows on subscriptions, on models we don't own. And we can be deplatformed at any time.
The memory trap applies here, but worse. With a departing employee, you lose knowledge. With AI, you lose the tool AND the workflow. Skills locked in their software disappear when they shut you down.
Three questions to ask yourself:
- What happens if that model disappears tomorrow—what breaks?
- Can I take what this tool learned and move it somewhere else?
- What will it cost me to start over with a new tool?
The 7-9 month thesis: frontier models are that far ahead of local open source. Mac Studio with 256GB memory has a seven-month wait. By the time hardware arrives, local models may have caught up to where frontier models are today.
The IPO warning: OpenAI and Anthropic are going public. Shareholders expect profit. These companies are losing money. The only path: raise prices, cut services, or both. If you've built everything on their platform, you're exposed.
The bottom line: When something is easy and fast, you haven't seen the real cost yet. The landlord always shows up. Know your risk, or make the investment to own your infrastructure.
Got a business question? Ask Scott here: scotttodd.net/ask
📜 Full Transcript (Click to expand)
Scott Todd (00:00)
Back in January, our entire world's changed with this major release of an open source product that just made AI agents so much more of a reality. And in this episode, we're going to look at some of the lessons learned over the last three months, some things that I want you to take away from as we dig deeper into the world of AI agents. We're going to fix my business, the show that helps you grow and fix your business so that you can grow it to what you want it to be.
I'm your host Scott Todd. I built multiple seven-figure businesses and we're going to dig in today to the world of AI agents. Now, here's what happened. Back in January, this open source project just took off. It was called OpenClaw. And the whole concept of this project, if you're not familiar with it, is that ultimately we could use AI subscriptions that we already had. Primarily, a lot of people were using the Anthropic.
subscriptions, and this thing was crazy good. It still had problems, but remember, it was new to us and what made it so great was the ability to do automated tasks for us. We could just give it something through texting it and it would run off and do the work. It was like having just this insanely capable assistant. Now, when it was set up, one of the things that allowed people
to use is their subscriptions. They were able to use their subscriptions and that made it very affordable to use because you could be on a $200 a month plan or a $100 a month plan. You could get a lot of work done for that plan in that subscription that you are already paying for. However, Anthropic and other AI companies, they didn't like that idea because tokens are expensive. Okay.
What we pay for as a subscription is subsidized. And I don't think a lot of people realize that, but it is subsidized. And then fast forward a few more months and at the end of a couple of weeks ago on a Friday, right before Easter, Good Friday, late in the afternoon, Anthropics sends out this email to everybody saying, hey, beginning tomorrow at 3 p.m., your subscription will not work for this open-claw instance or any other third party.
bolt on. And as a result of that, if you do say here, we're going to charge you through the API key, which as I said is expensive. Okay. So we found out very quickly just how expensive AI is in helping you to run your businesses. mean, people were like, Hey, this is costing me a couple hundred dollars a day. It's costing me a hundred dollars a day. I mean, that's three to $6,000 a month.
back to VAs and back to your employees, right? It doesn't look so great anymore, does it? Not when it's that expensive. So in order to combat that, what a lot of people did, and look, I'm in that boat too. I did it too. What a lot of people did was they changed their models. They went to cheaper models. They went to open source models. They went to models that still supported their subscriptions. And what we all found out is that these other models were not as good as what we had built our business workflows on. They were not as good at the time.
as the anthropic models. And as a result of that, workflow suffered. All over social media, people were saying, hey, this thing is no good anymore. My guy is not as good anymore. My agent's not doing the work that they're doing. There's something wrong here. These models are terrible. What are we going to do?
And right there, I had this revelation that, you know what? We just handed our entire businesses to some other company. And that's a problem. In real estate, there's this concept of don't build on rented land. And that's what we all did. We all built on rented land. We built our workflows on a subscription. We built our workflows on a model. We built workflows...
that were affordable at the time, but we did not own it. We do not own these AI models. And in fact, so many people will report on social media that they've been banned from this company or banned from that company. You can literally be deplatformed. And think about that for a minute. As we move forward in the world of AI agents, we're building on this model, this subscription model, or even if you're going to use the API tokens.
You're going to build your business workflows on something that could literally deplatform you. They can ban you. Okay? And ultimately what we're doing is we're building on rented land. We are handing over the keys to our entire business to a subscription, to a company. And when you think about it, and look, this isn't just small businesses. This is large businesses as well. Large businesses, all kinds of businesses are doing this.
Because those models, they're state of the art. They're fast. They're getting better and better with every single release. And what's available open source or what's available if you tried to run this locally, if you had the hardware to run it locally, is just not as good. The local models are getting better. But there is a lag. Because your state of the art models like, Claude or OpenAI,
what you can do locally or through open source at a cheaper rate is probably seven to nine months behind. So there's a trailing period of time in which this catches up to the state of the art models. But I want you to really sit with this for a minute and truly understand that, we've handed over the keys to some other company, and also there's other traps that we've created along the way.
I I've been studying business and entrepreneurship since I was a teenager. And in this entire time, the one thing that I've realized is that there's four traps that ultimately hold processes and companies back. One of them is what I call the memory trap. And the thing about the memory trap is that the memory trap is when a certain employee
knows how to do something, they become the linchpin of the business, they have all of the institutional knowledge, they walk out the door and guess what? They take with them that institutional knowledge. With AI, it's worse. Okay? With AI, it's worse because it's not just the knowledge, it's the actual tool, it's the actual workflow. We can't transfer these workflows easily. We can do it, we can replicate it.
in new models, we can replicate it with new services. But even with AI or with OpenClaw, it allowed us to create these skills, these repeatable skills that we could plug into other models. But we've also built these workflows on this model, on this particular company. And if something happens to that company, well, guess what? There goes our business too, right? Like we're on the struggle bus again. So what are we going to do about it?
look, I'm not telling you not to use AI. That would be crazy. That would be like, hey, don't use the internet because AI is literally not going anywhere. It really isn't going anywhere. Okay. But I want you to ask yourself these three questions because these are the questions I'm asking myself as we build out tools and we build out our automations. Number one, what happens if that model or that tool or that subscription
disappears tomorrow, what breaks? Okay. And if you're like, I can survive without it. Okay, you're fine. No problem. But if that thought of being banned by an AI company, D platform makes you just sick, and you're scared, guess what? You're building on rented land. Okay. As you look through this, the second question I want you asking or thinking to yourself is,
I take this tool, what this tool has learned and move it somewhere else? So, if you're building this correctly, you're building skills along the way. You're building repeatable skills that should become model agnostic. It should be able to be used with any software or any basically model of the future. But where is it? How do you have access to it? Do you know where it is or is it locked up somewhere? Okay. If it's locked up with, you know, their software,
The minute they shut you down, you're out of business. You can't take that with you. So you want to make sure those skills are local on your machine. You want to know where they are. You want to know how to redeploy them somewhere else. And then the third question I want you thinking about is, hey, what's it going to cost me to start over with a new tool? Because transitioning to a new tool, transitioning to a new model, transitioning to
wherever it is, is not going to be cheap and it's not going to be easy the more you build on it. Okay? So these are considerations, business considerations that have real risk that you're introducing into your business. Now for me, I'm beginning to wonder if this is the time in which it becomes the standard to make the investment in hardware that can run local AIs. Because
Look, whatever, if you tried to buy a Mac Studio, for example, of 256K, it's like eight grand. That's a business investment if that's what you're going to do. mean, theoretically, I know people that are not hiring people and they're using AI to not hire people. Okay, well, you're saving money in that case. So, is this the time to make the investment? I don't know. I I'm developing my own thesis from my own company and it's a thesis that you should
think about for your own company. Is this the right time? Because remember, even if you take the top of the line Mac Studio, for example, with today is available with 256 gig of memory, even if you wanted to buy it today, you can't get it. You're not going to get it now for seven months. Now, why is that seven months important? Because it goes back to what I was saying earlier, which is that the frontier models, the stuff that we're dealing with today is about seven to nine months ahead.
of these local source models that can run locally. So theoretically, as long as that continues to grow seven to nine months away from now, whatever we are able to do locally in an open source environment where the stuff, the skills, the data stays on our local machines, ultimately it should time out to be where when the hardware arrives in seven to nine months, we're there.
we're back to where we are today. So you're not necessarily losing anything except for further progress of the state of the art models tomorrow. So is it the right time? I'm not telling you that it is. I'm not telling you that it's not. I want to put this out there for you because the other consideration is as AI providers go public, mean, OpenAI is supposed to go public this year. Anthropic is supposed to go public this year. As they go public, what do you think publicly traded companies want?
The shareholders expect a profit. Well, how do you make a profit? Well, you either have to raise your prices when you're losing money, by the way. Okay, like because they're losing money. These AI companies are not profitable today. So the only way for them to make money is to get more customers, which they have a lot of customers to still go get, honestly. There's a big opportunity there. So you can go get more customers, that's for sure.
But then at some point, you're going to have to raise your rates and or cut your services and or both. That's the only way to profitability. So what does that mean for you and your business if you've built everything on top of their platform? So as you think about this, as you think about AI agents, as you think about putting these things into your business, what I really want you thinking about is, hey, wait a minute. This skill that I'm putting out there today, this thing that we're building, is it on rented land?
And if so, you have to know the risk, right? Because, I've been in business long enough to know that when something is easy, like it's getting easier to build stuff and processes,
when it's getting easier.
and it's happening faster than you haven't seen the real cost yet because eventually that landlord is going to show up and eventually that landlord is going to change the terms and you may not like those terms. So you have to keep asking yourself, are you okay with the risk or is now the time to invest in your business? That's up for you. I know which way I'm leaning.
And that is I'm looking at the investment because I'm going to be here for a while. And I hope you will be too. All right. I hope that you found this episode interesting. I will see you in our next episode. And if you have a business question you want answered, if you want us to address something, head over to scotttodd.net forward slash ask, and I'll see you in our next episode.
Back in January, our entire world's changed with this major release of an open source product that just made AI agents so much more of a reality. And in this episode, we're going to look at some of the lessons learned over the last three months, some things that I want you to take away from as we dig deeper into the world of AI agents. We're going to fix my business, the show that helps you grow and fix your business so that you can grow it to what you want it to be.
I'm your host Scott Todd. I built multiple seven-figure businesses and we're going to dig in today to the world of AI agents. Now, here's what happened. Back in January, this open source project just took off. It was called OpenClaw. And the whole concept of this project, if you're not familiar with it, is that ultimately we could use AI subscriptions that we already had. Primarily, a lot of people were using the Anthropic.
subscriptions, and this thing was crazy good. It still had problems, but remember, it was new to us and what made it so great was the ability to do automated tasks for us. We could just give it something through texting it and it would run off and do the work. It was like having just this insanely capable assistant. Now, when it was set up, one of the things that allowed people
to use is their subscriptions. They were able to use their subscriptions and that made it very affordable to use because you could be on a $200 a month plan or a $100 a month plan. You could get a lot of work done for that plan in that subscription that you are already paying for. However, Anthropic and other AI companies, they didn't like that idea because tokens are expensive. Okay.
What we pay for as a subscription is subsidized. And I don't think a lot of people realize that, but it is subsidized. And then fast forward a few more months and at the end of a couple of weeks ago on a Friday, right before Easter, Good Friday, late in the afternoon, Anthropics sends out this email to everybody saying, hey, beginning tomorrow at 3 p.m., your subscription will not work for this open-claw instance or any other third party.
bolt on. And as a result of that, if you do say here, we're going to charge you through the API key, which as I said is expensive. Okay. So we found out very quickly just how expensive AI is in helping you to run your businesses. mean, people were like, Hey, this is costing me a couple hundred dollars a day. It's costing me a hundred dollars a day. I mean, that's three to $6,000 a month.
back to VAs and back to your employees, right? It doesn't look so great anymore, does it? Not when it's that expensive. So in order to combat that, what a lot of people did, and look, I'm in that boat too. I did it too. What a lot of people did was they changed their models. They went to cheaper models. They went to open source models. They went to models that still supported their subscriptions. And what we all found out is that these other models were not as good as what we had built our business workflows on. They were not as good at the time.
as the anthropic models. And as a result of that, workflow suffered. All over social media, people were saying, hey, this thing is no good anymore. My guy is not as good anymore. My agent's not doing the work that they're doing. There's something wrong here. These models are terrible. What are we going to do?
And right there, I had this revelation that, you know what? We just handed our entire businesses to some other company. And that's a problem. In real estate, there's this concept of don't build on rented land. And that's what we all did. We all built on rented land. We built our workflows on a subscription. We built our workflows on a model. We built workflows...
that were affordable at the time, but we did not own it. We do not own these AI models. And in fact, so many people will report on social media that they've been banned from this company or banned from that company. You can literally be deplatformed. And think about that for a minute. As we move forward in the world of AI agents, we're building on this model, this subscription model, or even if you're going to use the API tokens.
You're going to build your business workflows on something that could literally deplatform you. They can ban you. Okay? And ultimately what we're doing is we're building on rented land. We are handing over the keys to our entire business to a subscription, to a company. And when you think about it, and look, this isn't just small businesses. This is large businesses as well. Large businesses, all kinds of businesses are doing this.
Because those models, they're state of the art. They're fast. They're getting better and better with every single release. And what's available open source or what's available if you tried to run this locally, if you had the hardware to run it locally, is just not as good. The local models are getting better. But there is a lag. Because your state of the art models like, Claude or OpenAI,
what you can do locally or through open source at a cheaper rate is probably seven to nine months behind. So there's a trailing period of time in which this catches up to the state of the art models. But I want you to really sit with this for a minute and truly understand that, we've handed over the keys to some other company, and also there's other traps that we've created along the way.
I I've been studying business and entrepreneurship since I was a teenager. And in this entire time, the one thing that I've realized is that there's four traps that ultimately hold processes and companies back. One of them is what I call the memory trap. And the thing about the memory trap is that the memory trap is when a certain employee
knows how to do something, they become the linchpin of the business, they have all of the institutional knowledge, they walk out the door and guess what? They take with them that institutional knowledge. With AI, it's worse. Okay? With AI, it's worse because it's not just the knowledge, it's the actual tool, it's the actual workflow. We can't transfer these workflows easily. We can do it, we can replicate it.
in new models, we can replicate it with new services. But even with AI or with OpenClaw, it allowed us to create these skills, these repeatable skills that we could plug into other models. But we've also built these workflows on this model, on this particular company. And if something happens to that company, well, guess what? There goes our business too, right? Like we're on the struggle bus again. So what are we going to do about it?
look, I'm not telling you not to use AI. That would be crazy. That would be like, hey, don't use the internet because AI is literally not going anywhere. It really isn't going anywhere. Okay. But I want you to ask yourself these three questions because these are the questions I'm asking myself as we build out tools and we build out our automations. Number one, what happens if that model or that tool or that subscription
disappears tomorrow, what breaks? Okay. And if you're like, I can survive without it. Okay, you're fine. No problem. But if that thought of being banned by an AI company, D platform makes you just sick, and you're scared, guess what? You're building on rented land. Okay. As you look through this, the second question I want you asking or thinking to yourself is,
I take this tool, what this tool has learned and move it somewhere else? So, if you're building this correctly, you're building skills along the way. You're building repeatable skills that should become model agnostic. It should be able to be used with any software or any basically model of the future. But where is it? How do you have access to it? Do you know where it is or is it locked up somewhere? Okay. If it's locked up with, you know, their software,
The minute they shut you down, you're out of business. You can't take that with you. So you want to make sure those skills are local on your machine. You want to know where they are. You want to know how to redeploy them somewhere else. And then the third question I want you thinking about is, hey, what's it going to cost me to start over with a new tool? Because transitioning to a new tool, transitioning to a new model, transitioning to
wherever it is, is not going to be cheap and it's not going to be easy the more you build on it. Okay? So these are considerations, business considerations that have real risk that you're introducing into your business. Now for me, I'm beginning to wonder if this is the time in which it becomes the standard to make the investment in hardware that can run local AIs. Because
Look, whatever, if you tried to buy a Mac Studio, for example, of 256K, it's like eight grand. That's a business investment if that's what you're going to do. mean, theoretically, I know people that are not hiring people and they're using AI to not hire people. Okay, well, you're saving money in that case. So, is this the time to make the investment? I don't know. I I'm developing my own thesis from my own company and it's a thesis that you should
think about for your own company. Is this the right time? Because remember, even if you take the top of the line Mac Studio, for example, with today is available with 256 gig of memory, even if you wanted to buy it today, you can't get it. You're not going to get it now for seven months. Now, why is that seven months important? Because it goes back to what I was saying earlier, which is that the frontier models, the stuff that we're dealing with today is about seven to nine months ahead.
of these local source models that can run locally. So theoretically, as long as that continues to grow seven to nine months away from now, whatever we are able to do locally in an open source environment where the stuff, the skills, the data stays on our local machines, ultimately it should time out to be where when the hardware arrives in seven to nine months, we're there.
we're back to where we are today. So you're not necessarily losing anything except for further progress of the state of the art models tomorrow. So is it the right time? I'm not telling you that it is. I'm not telling you that it's not. I want to put this out there for you because the other consideration is as AI providers go public, mean, OpenAI is supposed to go public this year. Anthropic is supposed to go public this year. As they go public, what do you think publicly traded companies want?
The shareholders expect a profit. Well, how do you make a profit? Well, you either have to raise your prices when you're losing money, by the way. Okay, like because they're losing money. These AI companies are not profitable today. So the only way for them to make money is to get more customers, which they have a lot of customers to still go get, honestly. There's a big opportunity there. So you can go get more customers, that's for sure.
But then at some point, you're going to have to raise your rates and or cut your services and or both. That's the only way to profitability. So what does that mean for you and your business if you've built everything on top of their platform? So as you think about this, as you think about AI agents, as you think about putting these things into your business, what I really want you thinking about is, hey, wait a minute. This skill that I'm putting out there today, this thing that we're building, is it on rented land?
And if so, you have to know the risk, right? Because, I've been in business long enough to know that when something is easy, like it's getting easier to build stuff and processes,
when it's getting easier.
and it's happening faster than you haven't seen the real cost yet because eventually that landlord is going to show up and eventually that landlord is going to change the terms and you may not like those terms. So you have to keep asking yourself, are you okay with the risk or is now the time to invest in your business? That's up for you. I know which way I'm leaning.
And that is I'm looking at the investment because I'm going to be here for a while. And I hope you will be too. All right. I hope that you found this episode interesting. I will see you in our next episode. And if you have a business question you want answered, if you want us to address something, head over to scotttodd.net forward slash ask, and I'll see you in our next episode.