How to Actually Choose the Right AI Agent

All right, so Marge, by the end of All right, so Marge, by the end of today's episode, what will everyone have today's episode, what will everyone have today's episode, what will everyone have learned from you? learned from you? learned from you? >> My goal is that by the end of this >> My goal is that by the end of this >> My goal is that by the end of this video, you understand that the harness video, you understand that the harness video, you understand that the harness of a model is [music] much more of a model is [music] much more of a model is [music] much more important than the model itself. important than the model itself. important than the model itself. >> It feels to me like Claude Code is like >> It feels to me like Claude Code is like >> It feels to me like Claude Code is like a wise old owl and then it feels like a wise old owl and then it feels like a wise old owl and then it feels like Codex is like the Rottweiler. It'll obey Codex is like the Rottweiler. It'll obey Codex is like the Rottweiler. It'll obey your commands and it will just keep your commands and it will just keep your commands and it will just keep going until [music] it's it's done. going until [music] it's it's done. going until [music] it's it's done. Because whenever we talk about, oh, Because whenever we talk about, oh, Because whenever we talk about, oh, Claude's better or Codex is better, you Claude's better or Codex is better, you Claude's better or Codex is better, you have this brain they're all fighting have this brain they're all fighting have this brain they're all fighting about, but everything around the brain about, but everything around the brain about, but everything around the brain is actually [music] what gets it to is actually [music] what gets it to is actually [music] what gets it to tick. This stuff isn't like magic at tick. This stuff isn't like magic at tick. This stuff isn't like magic at all. There is a whole factory of workers all. There is a whole factory of workers all. There is a whole factory of workers [music] that are making this model look [music] that are making this model look [music] that are making this model look way smarter than it is. My number one way smarter than it is. My number one way smarter than it is. My number one goal is to never be loyal to a provider goal is to never be loyal to a provider goal is to never be loyal to a provider only be loyal to my harness and my only be loyal to my harness and my only be loyal to my harness and my assets and I will switch the brain assets and I will switch the brain assets and I will switch the brain [music] interchangeably. [music] interchangeably. [music] interchangeably. >> So I think that is really important to >> So I think that is really important to >> So I think that is really important to be thinking you're building up your own be thinking you're building up your own be thinking you're building up your own IP and you need to make sure that you're IP and you need to make sure that you're IP and you need to make sure that you're protecting that. And I always think of protecting that. And I always think of protecting that. And I always think of the quote you can outsource the thinking the quote you can outsource the thinking the quote you can outsource the thinking [music] but you can never outsource the [music] but you can never outsource the [music] but you can never outsource the understanding. Skills and agents though understanding. Skills and agents though understanding. Skills and agents though decay incredibly fast to the point where decay incredibly fast to the point where decay incredibly fast to the point where Boris Churnney dropped a tweet saying Boris Churnney dropped a tweet saying Boris Churnney dropped a tweet saying you should [music] delete all of your you should [music] delete all of your you should [music] delete all of your skills every six months. All of them.

skills every six months. All of them. skills every six months. All of them. >> Did he? >> Did he? >> Did he? >> Do you know why? >> Do you know why? >> Do you know why? >> Why did he say that? >> All right. So, Mark, thank you so much >> All right. So, Mark, thank you so much for joining us today in person, which is for joining us today in person, which is for joining us today in person, which is awesome. We're in the beautiful country awesome. We're in the beautiful country awesome. We're in the beautiful country of Montenegro, which has been so much of Montenegro, which has been so much of Montenegro, which has been so much fun. We're here for an AI event and we fun. We're here for an AI event and we fun. We're here for an AI event and we figured why not sit down together and figured why not sit down together and figured why not sit down together and hopefully drop some sauce today. So, I'm hopefully drop some sauce today. So, I'm hopefully drop some sauce today. So, I'm super super pumped to be sitting down. super super pumped to be sitting down. super super pumped to be sitting down. If you guys don't know who Mark is, then If you guys don't know who Mark is, then If you guys don't know who Mark is, then hopefully after this video you start hopefully after this video you start hopefully after this video you start seeing his videos on your YouTube feed seeing his videos on your YouTube feed seeing his videos on your YouTube feed because his stuff is absolutely gold. because his stuff is absolutely gold. because his stuff is absolutely gold. I've been watching him for a while. I I've been watching him for a while. I I've been watching him for a while. I actually was watching him before I actually was watching him before I actually was watching him before I started making content. So, pretty cool started making content. So, pretty cool started making content. So, pretty cool moment for me to be able to sit down moment for me to be able to sit down moment for me to be able to sit down with Mr. Cashf here today. But, yeah, with Mr. Cashf here today. But, yeah, with Mr. Cashf here today. But, yeah, I'm super excited to dig in. been I'm super excited to dig in. been I'm super excited to dig in. been getting obviously tons of questions in getting obviously tons of questions in getting obviously tons of questions in the community and discussion around oh the community and discussion around oh the community and discussion around oh like we should be should we be using like we should be should we be using like we should be should we be using Hermes now or we've seen this local Hermes now or we've seen this local Hermes now or we've seen this local thing called PI and there's just so many thing called PI and there's just so many thing called PI and there's just so many tools going around and I think we want tools going around and I think we want tools going around and I think we want to make sure that what we're building at to make sure that what we're building at to make sure that what we're building at the end of the day is still relevant the end of the day is still relevant the end of the day is still relevant next year and the year after that next year and the year after that next year and the year after that because we don't know what might happen because we don't know what might happen because we don't know what might happen to cloud code next or codec so excited to cloud code next or codec so excited to cloud code next or codec so excited to dig in and um yeah thanks for kind of to dig in and um yeah thanks for kind of to dig in and um yeah thanks for kind of throwing together an excell let's let's throwing together an excell let's let's throwing together an excell let's let's start getting into it start getting into it start getting into it >> absolutely so so the first thing I want >> absolutely so so the first thing I want >> absolutely so so the first thing I want to do is just really lock in on this to do is just really lock in on this to do is just really lock in on this diagram specifically this little brain diagram specifically this little brain diagram specifically this little brain in a jar are here because whenever we in a jar are here because whenever we in a jar are here because whenever we talk about oh claude's better or codeex talk about oh claude's better or codeex talk about oh claude's better or codeex is better or gemini maybe one date might is better or gemini maybe one date might is better or gemini maybe one date might be better you have this brain they're be better you have this brain they're be better you have this brain they're all fighting about but everything around all fighting about but everything around all fighting about but everything around the brain is actually what gets it to the brain is actually what gets it to the brain is actually what gets it to tick so if we look at the different tick so if we look at the different tick so if we look at the different parts here so we have the ability to parts here so we have the ability to parts here so we have the ability to read we have the ability to write and read we have the ability to write and read we have the ability to write and edit files we have the ability to use edit files we have the ability to use edit files we have the ability to use what's called bash which basically takes

what's called bash which basically takes what's called bash which basically takes control of your computer makes folders control of your computer makes folders control of your computer makes folders moves folders all of these things aren't moves folders all of these things aren't moves folders all of these things aren't baked into the model and you see this if baked into the model and you see this if baked into the model and you see this if you ever use a local model and you say you ever use a local model and you say you ever use a local model and you say make me a website and spin it up on my make me a website and spin it up on my make me a website and spin it up on my local computer. It can do the first part local computer. It can do the first part local computer. It can do the first part but it can't do the second part. but it can't do the second part. but it can't do the second part. >> It is a brain that can tell you it can >> It is a brain that can tell you it can >> It is a brain that can tell you it can give you this output of the HTML but it give you this output of the HTML but it give you this output of the HTML but it can't go and spin up a local server on can't go and spin up a local server on can't go and spin up a local server on your computer. It doesn't have the limbs your computer. It doesn't have the limbs your computer. It doesn't have the limbs for that. So the more you start thinking for that. So the more you start thinking for that. So the more you start thinking about models in the sense that they are about models in the sense that they are about models in the sense that they are the brain but everything around them the brain but everything around them the brain but everything around them allows them to interact and have hands allows them to interact and have hands allows them to interact and have hands and legs then you start to really and legs then you start to really and legs then you start to really separate what is the importance of the separate what is the importance of the separate what is the importance of the brain versus everything around it and brain versus everything around it and brain versus everything around it and how can you enrich everything around it how can you enrich everything around it how can you enrich everything around it so you're less dependent on that brain so you're less dependent on that brain so you're less dependent on that brain >> because we are at a point where many >> because we are at a point where many >> because we are at a point where many local models whether it's a Kimmy or local models whether it's a Kimmy or local models whether it's a Kimmy or insert name of open source model here insert name of open source model here insert name of open source model here >> they can do like 80% of the day-to-day >> they can do like 80% of the day-to-day >> they can do like 80% of the day-to-day work work work >> you might not get the same firepower as >> you might not get the same firepower as >> you might not get the same firepower as you would with cloud code or codeex But you would with cloud code or codeex But you would with cloud code or codeex But for more and more roles increasingly for for more and more roles increasingly for for more and more roles increasingly for the next year especially I see a world the next year especially I see a world the next year especially I see a world where you run 70 to 80% on local where you run 70 to 80% on local where you run 70 to 80% on local assuming you have the hardware and you assuming you have the hardware and you assuming you have the hardware and you bring in the geniuses for genius level bring in the geniuses for genius level bring in the geniuses for genius level tasks are planning.

tasks are planning. tasks are planning. >> Mhm. >> Mhm. >> Mhm. >> I love that because I think when you >> I love that because I think when you >> I love that because I think when you start to really talk about cloud chat in start to really talk about cloud chat in start to really talk about cloud chat in the web the web the web >> versus a cloud code >> versus a cloud code >> versus a cloud code >> there is a gap there which honestly I >> there is a gap there which honestly I >> there is a gap there which honestly I wish they didn't call it cloud code. wish they didn't call it cloud code. wish they didn't call it cloud code. >> Yeah. Because the code part is all of >> Yeah. Because the code part is all of >> Yeah. Because the code part is all of these these tools that you mentioned and these these tools that you mentioned and these these tools that you mentioned and so like just to walk us through a really so like just to walk us through a really so like just to walk us through a really practical example. What is the practical example. What is the practical example. What is the difference between you asking claude difference between you asking claude difference between you asking claude chat to let's just say research chat to let's just say research chat to let's just say research something for you and create a PDF something for you and create a PDF something for you and create a PDF versus when you might ask cla code to do versus when you might ask cla code to do versus when you might ask cla code to do that exact same task? that exact same task? that exact same task? >> Yeah. So claude on the web versus cloud >> Yeah. So claude on the web versus cloud >> Yeah. So claude on the web versus cloud code on Yeah. So cloud on the web will code on Yeah. So cloud on the web will code on Yeah. So cloud on the web will have slightly different tools. So let's have slightly different tools. So let's have slightly different tools. So let's say you're using cloud chat and all it say you're using cloud chat and all it say you're using cloud chat and all it can really do is do the researching can really do is do the researching can really do is do the researching part, create the PDF part, but some part, create the PDF part, but some part, create the PDF part, but some parts in between maybe calling parts in between maybe calling parts in between maybe calling additional uh platforms or moving files additional uh platforms or moving files additional uh platforms or moving files on your computer might it might not have on your computer might it might not have on your computer might it might not have access to local files on your computer. access to local files on your computer. access to local files on your computer. It would have to do a lot of work It would have to do a lot of work It would have to do a lot of work assuming things in the background that assuming things in the background that assuming things in the background that it can't actually touch and feel and it can't actually touch and feel and it can't actually touch and feel and see. see. see. >> With the cloud code as a harness, it can >> With the cloud code as a harness, it can >> With the cloud code as a harness, it can not only interact with your local not only interact with your local not only interact with your local computer, it can interact with the computer, it can interact with the computer, it can interact with the cloud. So you have the best of both cloud. So you have the best of both cloud. So you have the best of both worlds. So you have one that is telling worlds. So you have one that is telling worlds. So you have one that is telling you hypothetically here's what we could you hypothetically here's what we could you hypothetically here's what we could do and it could can do some stuff do and it could can do some stuff do and it could can do some stuff increasingly it's getting better. I see increasingly it's getting better. I see increasingly it's getting better. I see a world where claude chat evaporates a world where claude chat evaporates a world where claude chat evaporates completely and all we have is co-work completely and all we have is co-work completely and all we have is co-work which is a light version of the harness which is a light version of the harness which is a light version of the harness of cloud code.

of cloud code. of cloud code. >> So everything we will interact with will >> So everything we will interact with will >> So everything we will interact with will have a harness. It's just to what extent have a harness. It's just to what extent have a harness. It's just to what extent is it highly capable to do the task is it highly capable to do the task is it highly capable to do the task you're looking for? you're looking for? you're looking for? >> 100%. And what I think is so cool about >> 100%. And what I think is so cool about >> 100%. And what I think is so cool about cloud code and you know when I started cloud code and you know when I started cloud code and you know when I started learning about it I remember how learning about it I remember how learning about it I remember how intimidated I was to start learning intimidated I was to start learning intimidated I was to start learning about it back in maybe January of like about it back in maybe January of like about it back in maybe January of like this year this year this year >> but when I just started asking it >> but when I just started asking it >> but when I just started asking it questions it feels like magic because questions it feels like magic because questions it feels like magic because yes you're interacting with the same yes you're interacting with the same yes you're interacting with the same model you might be used to but all of model you might be used to but all of model you might be used to but all of the harness stuff really just happens the harness stuff really just happens the harness stuff really just happens automatically because the harness is automatically because the harness is automatically because the harness is essentially built to understand here are essentially built to understand here are essentially built to understand here are the tools that I have as you can see in the tools that I have as you can see in the tools that I have as you can see in this diagram which well done on this this diagram which well done on this this diagram which well done on this diagram by the way it knows what's in diagram by the way it knows what's in diagram by the way it knows what's in there same way like you think there same way like you think there same way like you think >> you need you pick up a glass of water, >> you need you pick up a glass of water, >> you need you pick up a glass of water, your hands and your shoulder and it just your hands and your shoulder and it just your hands and your shoulder and it just works together to do it for you. So, I works together to do it for you. So, I works together to do it for you. So, I think that it's really cool to see think that it's really cool to see think that it's really cool to see something like this and even though it something like this and even though it something like this and even though it might at a glance look like you might might at a glance look like you might might at a glance look like you might have to know how to do the bash or the have to know how to do the bash or the have to know how to do the bash or the read or whatever. read or whatever. read or whatever. >> Yeah. >> Yeah. >> Yeah. >> But the model just takes care of it. >> But the model just takes care of it. >> But the model just takes care of it. >> Yeah. And what I want to focus on is >> Yeah. And what I want to focus on is >> Yeah. And what I want to focus on is although these come out of the box, although these come out of the box, although these come out of the box, right, that's the whole point of the right, that's the whole point of the right, that's the whole point of the cloth code harness. That's what made it cloth code harness. That's what made it cloth code harness. That's what made it amazing is you have the read, you have amazing is you have the read, you have amazing is you have the read, you have the edit, all the stuff is done for you.

the edit, all the stuff is done for you. the edit, all the stuff is done for you. And even with things like Pi, which is And even with things like Pi, which is And even with things like Pi, which is kind of like a very vanilla open- source kind of like a very vanilla open- source kind of like a very vanilla open- source version where you can build your own version where you can build your own version where you can build your own harness, that's all cool, but where you harness, that's all cool, but where you harness, that's all cool, but where you come in and where your channel has been come in and where your channel has been come in and where your channel has been really adding tons of value is what else really adding tons of value is what else really adding tons of value is what else can you add to this factory? So now you can you add to this factory? So now you can you add to this factory? So now you have things like Lego blocks that are have things like Lego blocks that are have things like Lego blocks that are modular and these are those skills, modular and these are those skills, modular and these are those skills, these plugins, um all of these these plugins, um all of these these plugins, um all of these additional things you can layer on. So additional things you can layer on. So additional things you can layer on. So then you have this ecosystem, we have then you have this ecosystem, we have then you have this ecosystem, we have this orchestra where you have the brain this orchestra where you have the brain this orchestra where you have the brain in the middle telling everything else in the middle telling everything else in the middle telling everything else exactly what the goal is. And depending exactly what the goal is. And depending exactly what the goal is. And depending on the intelligence of the model, it on the intelligence of the model, it on the intelligence of the model, it might become better at knowing ah for might become better at knowing ah for might become better at knowing ah for this I need some bash with a write and this I need some bash with a write and this I need some bash with a write and I'll need to go through what's called I'll need to go through what's called I'll need to go through what's called the agentic loop which is purely you ask the agentic loop which is purely you ask the agentic loop which is purely you ask thing thing gets executed result of thing thing gets executed result of thing thing gets executed result of thing happens the result could be an thing happens the result could be an thing happens the result could be an error it could be a success it takes error it could be a success it takes error it could be a success it takes that stimuli and it keeps going in that that stimuli and it keeps going in that that stimuli and it keeps going in that loop and its ability to keep going in loop and its ability to keep going in loop and its ability to keep going in that loop comma well is fully based on that loop comma well is fully based on that loop comma well is fully based on how sophisticated that harness is a how sophisticated that harness is a how sophisticated that harness is a better model will know how to use tools better model will know how to use tools better model will know how to use tools better. It's kind of like bringing an better. It's kind of like bringing an better. It's kind of like bringing an expert handy person who's worked for a expert handy person who's worked for a expert handy person who's worked for a year and studied for many ages versus year and studied for many ages versus year and studied for many ages versus someone who has all the battle scars, someone who has all the battle scars, someone who has all the battle scars, the really powerful types of tools and the really powerful types of tools and the really powerful types of tools and has a tasset knowledge of when and where has a tasset knowledge of when and where has a tasset knowledge of when and where to use them a little bit better.

to use them a little bit better. to use them a little bit better. >> They'll perform infinitely more >> They'll perform infinitely more >> They'll perform infinitely more powerfully than the first one. powerfully than the first one. powerfully than the first one. >> So the same concept here, both are going >> So the same concept here, both are going >> So the same concept here, both are going to be smart. So the model itself is to be smart. So the model itself is to be smart. So the model itself is great, but if I showed you right now an great, but if I showed you right now an great, but if I showed you right now an example, example, example, >> can we pop over to let's say an LM >> can we pop over to let's say an LM >> can we pop over to let's say an LM studio? So let's pop over here. studio? So let's pop over here. studio? So let's pop over here. >> Yeah. So what is LM Studio to anyone >> Yeah. So what is LM Studio to anyone >> Yeah. So what is LM Studio to anyone that's never used it before? that's never used it before? that's never used it before? >> Yeah. So think of LM Studio as your >> Yeah. So think of LM Studio as your >> Yeah. So think of LM Studio as your ability to run chat GPT with an open- ability to run chat GPT with an open- ability to run chat GPT with an open- source model. It doesn't have a harness source model. It doesn't have a harness source model. It doesn't have a harness by default. So you can just interact by default. So you can just interact by default. So you can just interact with the brain, which is beautiful with the brain, which is beautiful with the brain, which is beautiful because it'll show you exactly why we because it'll show you exactly why we because it'll show you exactly why we have an issue here. have an issue here. have an issue here. >> Mhm. >> Mhm. >> Mhm. >> So if I go and I'm just using Quen 27B >> So if I go and I'm just using Quen 27B >> So if I go and I'm just using Quen 27B here. I have some more powerful models, here. I have some more powerful models, here. I have some more powerful models, but I don't want this computer to but I don't want this computer to but I don't want this computer to explode while I'm recording it. explode while I'm recording it. explode while I'm recording it. >> So I'm just going to say using the >> So I'm just going to say using the >> So I'm just going to say using the beautiful Glido. beautiful Glido. beautiful Glido. >> Okay. So, I want you to make a very >> Okay. So, I want you to make a very >> Okay. So, I want you to make a very basic landing page for my AI basic landing page for my AI basic landing page for my AI consultancy. I want you to call it consultancy. I want you to call it consultancy. I want you to call it prompt advisors and I want you to spin prompt advisors and I want you to spin prompt advisors and I want you to spin it up locally on my computer so I can it up locally on my computer so I can it up locally on my computer so I can host it and show the entire audience. host it and show the entire audience. host it and show the entire audience. Now, if I run this over, no matter how Now, if I run this over, no matter how Now, if I run this over, no matter how much time this takes, it will be able to much time this takes, it will be able to much time this takes, it will be able to tell me that it can't spin it up. The tell me that it can't spin it up. The tell me that it can't spin it up. The reason why is it does not have the limb reason why is it does not have the limb reason why is it does not have the limb of being able to interact with my of being able to interact with my of being able to interact with my computer to even create that server.

computer to even create that server. computer to even create that server. >> Mhm. >> Mhm. >> Mhm. >> You get this exact same request to >> You get this exact same request to >> You get this exact same request to codeex or cloud code. It's not going to codeex or cloud code. It's not going to codeex or cloud code. It's not going to sweat even twice because it comes with sweat even twice because it comes with sweat even twice because it comes with that out of the box. So now it's going that out of the box. So now it's going that out of the box. So now it's going to create what is the HTML itself to create what is the HTML itself to create what is the HTML itself because all it can do is take input and because all it can do is take input and because all it can do is take input and get output. Imagine if your brain was in get output. Imagine if your brain was in get output. Imagine if your brain was in a jar. All you can get is some form of a jar. All you can get is some form of a jar. All you can get is some form of stimuli and release the signal of like stimuli and release the signal of like stimuli and release the signal of like what you think the answer would be. Same what you think the answer would be. Same what you think the answer would be. Same concept. So one can theoretically do the concept. So one can theoretically do the concept. So one can theoretically do the thing, thing, thing, >> but it can't take it from a all the way >> but it can't take it from a all the way >> but it can't take it from a all the way to the touchdown. to the touchdown. to the touchdown. >> Yeah. The whole point of the harness is >> Yeah. The whole point of the harness is >> Yeah. The whole point of the harness is how do we go from the output of this how do we go from the output of this how do we go from the output of this very intelligent model to some form of very intelligent model to some form of very intelligent model to some form of tangible output in your hands. tangible output in your hands. tangible output in your hands. >> Totally. Yeah. And I I actually remember >> Totally. Yeah. And I I actually remember >> Totally. Yeah. And I I actually remember hearing some stories when you know hearing some stories when you know hearing some stories when you know Claude first started to come around Claude first started to come around Claude first started to come around first started to come around before we first started to come around before we first started to come around before we had the harnesses and you would hear had the harnesses and you would hear had the harnesses and you would hear these stories of these companies that these stories of these companies that these stories of these companies that were building products because they were were building products because they were were building products because they were having Claude write code and then just having Claude write code and then just having Claude write code and then just copying and pasting it into whatever copying and pasting it into whatever copying and pasting it into whatever they needed to actually build the code they needed to actually build the code they needed to actually build the code and host it and all that kind of thing. and host it and all that kind of thing. and host it and all that kind of thing. And that just it's another one of the And that just it's another one of the And that just it's another one of the examples that goes right along with like examples that goes right along with like examples that goes right along with like this is still inherently very powerful this is still inherently very powerful this is still inherently very powerful but not as powerful as when you kind of but not as powerful as when you kind of but not as powerful as when you kind of give it the whole agentic loop that you give it the whole agentic loop that you give it the whole agentic loop that you talked about earlier. All right guys, talked about earlier. All right guys, talked about earlier. All right guys, real quick. Huge thanks to Clay for real quick. Huge thanks to Clay for real quick. Huge thanks to Clay for sponsoring this part of the video. Now, sponsoring this part of the video. Now, sponsoring this part of the video. Now, one of the most common questions I get one of the most common questions I get one of the most common questions I get is where to actually find leads for cold is where to actually find leads for cold is where to actually find leads for cold outreach and how to learn enough about outreach and how to learn enough about outreach and how to learn enough about these people to send them something that these people to send them something that these people to send them something that they'll actually open because a raw list they'll actually open because a raw list they'll actually open because a raw list of names doesn't tell you things like if of names doesn't tell you things like if of names doesn't tell you things like if they have decision-making authority and they have decision-making authority and they have decision-making authority and how to reach them or what they even care how to reach them or what they even care how to reach them or what they even care about. So, Clay is a data enrichment and about. So, Clay is a data enrichment and about. So, Clay is a data enrichment and orchestration platform. And it gives you orchestration platform. And it gives you orchestration platform. And it gives you access to over 250 data providers and AI access to over 250 data providers and AI access to over 250 data providers and AI research tools all in one spot. So,

research tools all in one spot. So, research tools all in one spot. So, instead of asking your agent to dig up instead of asking your agent to dig up instead of asking your agent to dig up whatever it can find online, you can whatever it can find online, you can whatever it can find online, you can research a whole list at once and pull research a whole list at once and pull research a whole list at once and pull the right person to contact their work the right person to contact their work the right person to contact their work email and the size of their company. And email and the size of their company. And email and the size of their company. And if one provider comes back empty, Clay if one provider comes back empty, Clay if one provider comes back empty, Clay just moves down the list to the next one just moves down the list to the next one just moves down the list to the next one until it gets you a result. And what's until it gets you a result. And what's until it gets you a result. And what's really cool is the logic you build stays really cool is the logic you build stays really cool is the logic you build stays attached to your data. So the same steps attached to your data. So the same steps attached to your data. So the same steps can be run again and again on every new can be run again and again on every new can be run again and again on every new lead that you add. And you can see each lead that you add. And you can see each lead that you add. And you can see each step that it took, like which provider step that it took, like which provider step that it took, like which provider every value came from and what it cost every value came from and what it cost every value came from and what it cost you. You can build all of this from you. You can build all of this from you. You can build all of this from Cloud Code with the Clay CLI, which is Cloud Code with the Clay CLI, which is Cloud Code with the Clay CLI, which is what I'm doing right here. So try Clay what I'm doing right here. So try Clay what I'm doing right here. So try Clay using the link in the description, and using the link in the description, and using the link in the description, and you'll get 2,000 free credits. Now, you'll get 2,000 free credits. Now, you'll get 2,000 free credits. Now, let's get back to the video. Okay, so we let's get back to the video. Okay, so we let's get back to the video. Okay, so we talked about how important these pieces talked about how important these pieces talked about how important these pieces are because this is where you can really are because this is where you can really are because this is where you can really start to add in your own like subject start to add in your own like subject start to add in your own like subject matter expertise, matter expertise, matter expertise, >> which is really what helps make the >> which is really what helps make the >> which is really what helps make the system feel more like it's yours. And system feel more like it's yours. And system feel more like it's yours. And what's cool about this stuff is that as what's cool about this stuff is that as what's cool about this stuff is that as you build on like maybe these skills, you build on like maybe these skills, you build on like maybe these skills, context files, plugins, whatever it may context files, plugins, whatever it may context files, plugins, whatever it may be, you're not locking yourself in to be, you're not locking yourself in to be, you're not locking yourself in to that harness because these can be used that harness because these can be used that harness because these can be used across other harnesses and other models across other harnesses and other models across other harnesses and other models as well. as well. as well. >> So the question I wanted to ask you is >> So the question I wanted to ask you is >> So the question I wanted to ask you is as you switch through these things and as you switch through these things and as you switch through these things and you know your codecs can touch your you know your codecs can touch your you know your codecs can touch your AIOS, your second brain, whatever AIOS, your second brain, whatever AIOS, your second brain, whatever everyone's calling it these days, everyone's calling it these days, everyone's calling it these days, Hermes, openclaw, whatever comes next.

Hermes, openclaw, whatever comes next. Hermes, openclaw, whatever comes next. How do you personally, Mark Cashup, how How do you personally, Mark Cashup, how How do you personally, Mark Cashup, how do you think about the way that you do you think about the way that you do you think about the way that you switch between those harnesses and like switch between those harnesses and like switch between those harnesses and like you know if you like codecs for certain you know if you like codecs for certain you know if you like codecs for certain tasks, Hermes for certain tasks, what tasks, Hermes for certain tasks, what tasks, Hermes for certain tasks, what does that look like for you? does that look like for you? does that look like for you? >> For sure. So my number one goal is to >> For sure. So my number one goal is to >> For sure. So my number one goal is to never be loyal to a provider to only be never be loyal to a provider to only be never be loyal to a provider to only be loyal to my harness and my assets and I loyal to my harness and my assets and I loyal to my harness and my assets and I will switch the brain will switch the brain will switch the brain >> interchangeably. I have zero loyalty to >> interchangeably. I have zero loyalty to >> interchangeably. I have zero loyalty to that. that. that. >> So although I use cloud code a lot, it's >> So although I use cloud code a lot, it's >> So although I use cloud code a lot, it's more so I'm used to it. I know the more so I'm used to it. I know the more so I'm used to it. I know the rhythm of it. I know what to expect. rhythm of it. I know what to expect. rhythm of it. I know what to expect. Mhm. Mhm. Mhm. >> But I will create all of my skills so >> But I will create all of my skills so >> But I will create all of my skills so they can work with any language model they can work with any language model they can work with any language model open to closed source open to closed source open to closed source >> and I will prioritize really battle >> and I will prioritize really battle >> and I will prioritize really battle testing it with every single thing that testing it with every single thing that testing it with every single thing that I can. So I'll try with cloud code. Then I can. So I'll try with cloud code. Then I can. So I'll try with cloud code. Then I have the skill I call / poly skill. I have the skill I call / poly skill. I have the skill I call / poly skill. It'll convert any skill for cloud code It'll convert any skill for cloud code It'll convert any skill for cloud code and optimize it for codeex. and optimize it for codeex. and optimize it for codeex. >> Okay. >> Okay. >> Okay. >> So I'll make sure that every skill is >> So I'll make sure that every skill is >> So I'll make sure that every skill is eligible for both. eligible for both. eligible for both. >> That both all of the different models >> That both all of the different models >> That both all of the different models know exactly where to find the same know exactly where to find the same know exactly where to find the same assets. So I have one link that has all assets. So I have one link that has all assets. So I have one link that has all of my core assets. They're agnostic of of my core assets. They're agnostic of of my core assets. They're agnostic of working with any of those models. So I working with any of those models. So I working with any of those models. So I can move around as needed.

can move around as needed. can move around as needed. >> Yeah. >> Yeah. >> Yeah. >> Now with your question, codeex recently >> Now with your question, codeex recently >> Now with your question, codeex recently I've been running 60% of the time. I've been running 60% of the time. I've been running 60% of the time. Although I've been very loyal loyal to Although I've been very loyal loyal to Although I've been very loyal loyal to Claude Code for the vast amount of time. Claude Code for the vast amount of time. Claude Code for the vast amount of time. Claude Code is amazing at ideiation and Claude Code is amazing at ideiation and Claude Code is amazing at ideiation and planning to an extent. It's a visionary. planning to an extent. It's a visionary. planning to an extent. It's a visionary. >> It likes to go back and forth. It likes >> It likes to go back and forth. It likes >> It likes to go back and forth. It likes to judge you. But the one thing it to judge you. But the one thing it to judge you. But the one thing it doesn't like to do sometimes is follow doesn't like to do sometimes is follow doesn't like to do sometimes is follow the exact instruction in the exact way the exact instruction in the exact way the exact instruction in the exact way you gave it. you gave it. you gave it. >> So I see Codeex as a surgeon and Claude >> So I see Codeex as a surgeon and Claude >> So I see Codeex as a surgeon and Claude Code as a gifted artist. Code as a gifted artist. Code as a gifted artist. >> Yeah. >> Yeah. >> Yeah. >> And many times I have to bring in Codeex >> And many times I have to bring in Codeex >> And many times I have to bring in Codeex to look over the plan of Claude Code and to look over the plan of Claude Code and to look over the plan of Claude Code and I make them fight in a loop for 10 I make them fight in a loop for 10 I make them fight in a loop for 10 different rounds until Claude Code different rounds until Claude Code different rounds until Claude Code finally has all the missing parts that finally has all the missing parts that finally has all the missing parts that Codex could see. All the things it Codex could see. All the things it Codex could see. All the things it wasn't anticipating. wasn't anticipating. wasn't anticipating. >> Totally. Yeah. I heard this tweet that >> Totally. Yeah. I heard this tweet that >> Totally. Yeah. I heard this tweet that or I saw this tweet that I thought was or I saw this tweet that I thought was or I saw this tweet that I thought was awesome and I want to see if you agree awesome and I want to see if you agree awesome and I want to see if you agree and I think you will because I have a and I think you will because I have a and I think you will because I have a very very similar philosophy to the way very very similar philosophy to the way very very similar philosophy to the way I I think about the two. But it I I think about the two. But it I I think about the two. But it basically said like it feels to me like basically said like it feels to me like basically said like it feels to me like cloud code is like a wise old owl. You cloud code is like a wise old owl. You cloud code is like a wise old owl. You can talk to it, you plan with it, it'll, can talk to it, you plan with it, it'll, can talk to it, you plan with it, it'll, you know, push back on you a little bit you know, push back on you a little bit you know, push back on you a little bit and then it feels like Codeex is like and then it feels like Codeex is like and then it feels like Codeex is like the Rottweiler that will grab onto the the Rottweiler that will grab onto the the Rottweiler that will grab onto the task and it will just it'll obey your task and it will just it'll obey your task and it will just it'll obey your commands and it will just keep going commands and it will just keep going commands and it will just keep going until it's it's done essentially because until it's it's done essentially because until it's it's done essentially because I think that the the verification loops I think that the the verification loops I think that the the verification loops inside Codex feel really really sharp to inside Codex feel really really sharp to inside Codex feel really really sharp to me. But um I think it's important that me. But um I think it's important that me. But um I think it's important that we kind of have that acknowledgement of we kind of have that acknowledgement of we kind of have that acknowledgement of you know which harness is best, which you know which harness is best, which you know which harness is best, which model is best. And it's it's more so model is best. And it's it's more so model is best. And it's it's more so which one is best for this specific which one is best for this specific which one is best for this specific task. Yeah, it might be, you know, a task. Yeah, it might be, you know, a task. Yeah, it might be, you know, a fivestep process, but for step one and

fivestep process, but for step one and fivestep process, but for step one and two, maybe that's where you go for the two, maybe that's where you go for the two, maybe that's where you go for the codeex and then, you know, or vice codeex and then, you know, or vice codeex and then, you know, or vice versa. So, I think that's that's good to versa. So, I think that's that's good to versa. So, I think that's that's good to hear you say as well. Now, where do you hear you say as well. Now, where do you hear you say as well. Now, where do you see see see >> um you know, I think that Hermes and >> um you know, I think that Hermes and >> um you know, I think that Hermes and Open Cloud kind of get bucketed in Open Cloud kind of get bucketed in Open Cloud kind of get bucketed in together as well. Where do you see the together as well. Where do you see the together as well. Where do you see the differentiation there and you know with differentiation there and you know with differentiation there and you know with COD and cloud code as well? COD and cloud code as well? COD and cloud code as well? >> Well, with Hermes, what are you doing? >> Well, with Hermes, what are you doing? >> Well, with Hermes, what are you doing? You are bringing in the Hermes harness You are bringing in the Hermes harness You are bringing in the Hermes harness >> and you're just looping in whatever >> and you're just looping in whatever >> and you're just looping in whatever whatever model you want. So the reason whatever model you want. So the reason whatever model you want. So the reason why people have to really make their why people have to really make their why people have to really make their Hermes agent tailored to them, a special Hermes agent tailored to them, a special Hermes agent tailored to them, a special snowflake, is depending on how they want snowflake, is depending on how they want snowflake, is depending on how they want to use these models with Hermes, they to use these models with Hermes, they to use these models with Hermes, they have to keep hacking Hermes harness. Not have to keep hacking Hermes harness. Not have to keep hacking Hermes harness. Not this all these skill MD files, etc. this all these skill MD files, etc. this all these skill MD files, etc. They're all great, but the thing that They're all great, but the thing that They're all great, but the thing that made Hermes better than Open Claw is its made Hermes better than Open Claw is its made Hermes better than Open Claw is its harness. So with many tasks if you go harness. So with many tasks if you go harness. So with many tasks if you go one to one Hermes agent versus codeex one to one Hermes agent versus codeex one to one Hermes agent versus codeex you'll have a different result vanilla you'll have a different result vanilla you'll have a different result vanilla but you can eventually massage Hermes but you can eventually massage Hermes but you can eventually massage Hermes agents harness agents harness agents harness >> to do this verification loop that you >> to do this verification loop that you >> to do this verification loop that you mentioned earlier they really like about mentioned earlier they really like about mentioned earlier they really like about codeex is that it performs a very codeex is that it performs a very codeex is that it performs a very similar to it so you can do a level of similar to it so you can do a level of similar to it so you can do a level of monkey see monkey do.

monkey see monkey do. monkey see monkey do. >> Yeah. >> Yeah. >> Yeah. >> So one thing that I did and you know you >> So one thing that I did and you know you >> So one thing that I did and you know you can go obviously no affiliation here can go obviously no affiliation here can go obviously no affiliation here it's open source. If you go to pi.dev it's open source. If you go to pi.dev it's open source. If you go to pi.dev dev, you will have this harness that you dev, you will have this harness that you dev, you will have this harness that you can bring onto your computer. You can can bring onto your computer. You can can bring onto your computer. You can copy with one command or if you're copy with one command or if you're copy with one command or if you're feeling daunted, what I did is I take feeling daunted, what I did is I take feeling daunted, what I did is I take this link, I feed it to codeex or cloud this link, I feed it to codeex or cloud this link, I feed it to codeex or cloud code and say go read the documentation, code and say go read the documentation, code and say go read the documentation, fan out some agents, learn about this fan out some agents, learn about this fan out some agents, learn about this whole harness thing. Once I do that, I whole harness thing. Once I do that, I whole harness thing. Once I do that, I can then ask it to go through what are can then ask it to go through what are can then ask it to go through what are called JSON L files. Basically, every called JSON L files. Basically, every called JSON L files. Basically, every conversation you have on your codeex and conversation you have on your codeex and conversation you have on your codeex and cloud code exists on your computer. So, cloud code exists on your computer. So, cloud code exists on your computer. So, I'll have it go look through all the I'll have it go look through all the I'll have it go look through all the conversations because they have the conversations because they have the conversations because they have the metadata of what tools were called, what metadata of what tools were called, what metadata of what tools were called, what verifications were done, in what order, verifications were done, in what order, verifications were done, in what order, and then I could have it monkey see and then I could have it monkey see and then I could have it monkey see monkey do. How do I start to make my own monkey do. How do I start to make my own monkey do. How do I start to make my own version of the harness that would react version of the harness that would react version of the harness that would react and do the same things based on similar and do the same things based on similar and do the same things based on similar types of tasks? types of tasks? types of tasks? >> You can start to reverse engineer all of >> You can start to reverse engineer all of >> You can start to reverse engineer all of the things that you love about cloud the things that you love about cloud the things that you love about cloud code and codecs. bring it to your own code and codecs. bring it to your own code and codecs. bring it to your own harness and eventually you can have one harness and eventually you can have one harness and eventually you can have one harness for everything where you bring harness for everything where you bring harness for everything where you bring in all these models as a brain that's in all these models as a brain that's in all these models as a brain that's leased you swap out as you need. So leased you swap out as you need. So leased you swap out as you need. So that's where I think we will get to that's where I think we will get to that's where I think we will get to where right now everything's tribal where right now everything's tribal where right now everything's tribal right YouTube is tribal X is tribal I am right YouTube is tribal X is tribal I am right YouTube is tribal X is tribal I am team CEX I am team cloud code I am team team CEX I am team cloud code I am team team CEX I am team cloud code I am team open source you all are Neanderthalss open source you all are Neanderthalss open source you all are Neanderthalss right for me I'm anti- tribal I am how right for me I'm anti- tribal I am how right for me I'm anti- tribal I am how do I make a system where any brain that do I make a system where any brain that do I make a system where any brain that could serve me for the best speed for could serve me for the best speed for could serve me for the best speed for the best rate at the best time can be the best rate at the best time can be the best rate at the best time can be swapped in with little to no acclimation swapped in with little to no acclimation swapped in with little to no acclimation needed needed needed >> so it's not a pain for me if Gemini

>> so it's not a pain for me if Gemini >> so it's not a pain for me if Gemini wakes up tomorrow wakes up tomorrow wakes up tomorrow >> truly wakes up and now it becomes >> truly wakes up and now it becomes >> truly wakes up and now it becomes amazing saying, "Okay, it could probably amazing saying, "Okay, it could probably amazing saying, "Okay, it could probably take me 24 hours to swap everything to take me 24 hours to swap everything to take me 24 hours to swap everything to Gemini." Gemini." Gemini." >> Mhm. >> Mhm. >> Mhm. >> And [clears throat] I love that >> And [clears throat] I love that >> And [clears throat] I love that nimleness. nimleness. nimleness. >> Yeah. >> Yeah. >> Yeah. >> Because while Codeex is amazing today, >> Because while Codeex is amazing today, >> Because while Codeex is amazing today, Cloud Code might come cuz they're Cloud Code might come cuz they're Cloud Code might come cuz they're probably going to IPO sooner. Make a big probably going to IPO sooner. Make a big probably going to IPO sooner. Make a big push. And if we finally get a mythos push. And if we finally get a mythos push. And if we finally get a mythos that's not nerfed or neutered to that's not nerfed or neutered to that's not nerfed or neutered to infinity, infinity, infinity, >> you might want to move all your stuff >> you might want to move all your stuff >> you might want to move all your stuff there. there. there. >> Yeah. Yeah. And who knows what could >> Yeah. Yeah. And who knows what could >> Yeah. Yeah. And who knows what could happen from a price perspective as well happen from a price perspective as well happen from a price perspective as well for us as consumers. So I think that is for us as consumers. So I think that is for us as consumers. So I think that is really important to be thinking you're really important to be thinking you're really important to be thinking you're building out your own IP essentially and building out your own IP essentially and building out your own IP essentially and you need to make sure that you're you need to make sure that you're you need to make sure that you're protecting that and it's it's nimble. So protecting that and it's it's nimble. So protecting that and it's it's nimble. So I love that point there. I love that point there. I love that point there. >> Now you mentioned something earlier >> Now you mentioned something earlier >> Now you mentioned something earlier about making sure that your skills and about making sure that your skills and about making sure that your skills and your whole ecosystem is model agnostic your whole ecosystem is model agnostic your whole ecosystem is model agnostic and you have a special skill that you and you have a special skill that you and you have a special skill that you use to make sure that they can work and use to make sure that they can work and use to make sure that they can work and are optimized for different models and are optimized for different models and are optimized for different models and harnesses as well. harnesses as well. harnesses as well. >> What does that actual process look like? >> What does that actual process look like? >> What does that actual process look like? Because typically when we see like our Because typically when we see like our Because typically when we see like our skill files or folders, it's usually a skill files or folders, it's usually a skill files or folders, it's usually a markdown file and that's sometimes assoc markdown file and that's sometimes assoc markdown file and that's sometimes assoc or you know kind of like also has in or you know kind of like also has in or you know kind of like also has in there maybe a few Python scripts or there maybe a few Python scripts or there maybe a few Python scripts or whatever the skill does. Maybe there's whatever the skill does. Maybe there's whatever the skill does. Maybe there's some assets that go along with it but some assets that go along with it but some assets that go along with it but ultimately you've kind of got just like ultimately you've kind of got just like ultimately you've kind of got just like a master markdown file. So how do you a master markdown file. So how do you a master markdown file. So how do you actually make sure that codecs can pick actually make sure that codecs can pick actually make sure that codecs can pick it up and use it as well or other agents it up and use it as well or other agents it up and use it as well or other agents could pick it up?

could pick it up? could pick it up? >> Absolutely. So the main thing to >> Absolutely. So the main thing to >> Absolutely. So the main thing to remember is that like you said all these remember is that like you said all these remember is that like you said all these skill files structurally look similar. skill files structurally look similar. skill files structurally look similar. They have this thing at the top called They have this thing at the top called They have this thing at the top called YAML where it's the name of the skill YAML where it's the name of the skill YAML where it's the name of the skill and what's called kebab case. You have and what's called kebab case. You have and what's called kebab case. You have the description and in the description the description and in the description the description and in the description you have a series of trigger words where you have a series of trigger words where you have a series of trigger words where when user does X I want you to invoke when user does X I want you to invoke when user does X I want you to invoke the skill to do Y. the skill to do Y. the skill to do Y. >> So what I did is I offload this. I asked >> So what I did is I offload this. I asked >> So what I did is I offload this. I asked codeex go and take a look at all the codeex go and take a look at all the codeex go and take a look at all the documentation from cloud code. Look at documentation from cloud code. Look at documentation from cloud code. Look at your own documentation. look at the your own documentation. look at the your own documentation. look at the documentation from let's say this documentation from let's say this documentation from let's say this specific other provider and go see how specific other provider and go see how specific other provider and go see how to build a versatile Swiss knife skill to build a versatile Swiss knife skill to build a versatile Swiss knife skill that will work for all of them as that will work for all of them as that will work for all of them as optimized as possible. optimized as possible. optimized as possible. >> Mhm. >> Mhm. >> Mhm. >> So Codeex prioritizes some things about >> So Codeex prioritizes some things about >> So Codeex prioritizes some things about skills that cloud code doesn't and vice skills that cloud code doesn't and vice skills that cloud code doesn't and vice versa. So how do we make sure that both versa. So how do we make sure that both versa. So how do we make sure that both are included? Now when it comes to are included? Now when it comes to are included? Now when it comes to scripts, Python is Python luckily. So scripts, Python is Python luckily. So scripts, Python is Python luckily. So that is already generic on its own. No that is already generic on its own. No that is already generic on its own. No need to worry about that. how you invoke need to worry about that. how you invoke need to worry about that. how you invoke that Python, how it knows when to use it that Python, how it knows when to use it that Python, how it knows when to use it and how to use it. That's where you and how to use it. That's where you and how to use it. That's where you might need to massage it a little bit. might need to massage it a little bit. might need to massage it a little bit. >> So even when I use this / poly skill, it >> So even when I use this / poly skill, it >> So even when I use this / poly skill, it will always look for these slight will always look for these slight will always look for these slight differences knowing ah codeex might miss differences knowing ah codeex might miss differences knowing ah codeex might miss this in the way that you're triggering this in the way that you're triggering this in the way that you're triggering it using cloud code. So let's make the it using cloud code. So let's make the it using cloud code. So let's make the description that much more beefier.

description that much more beefier. description that much more beefier. >> So the likelihood that it picks it up >> So the likelihood that it picks it up >> So the likelihood that it picks it up across the board is much higher. across the board is much higher. across the board is much higher. >> Totally. So I just have AI do the dirty >> Totally. So I just have AI do the dirty >> Totally. So I just have AI do the dirty work to go see AI documentation and I work to go see AI documentation and I work to go see AI documentation and I update these monthly on a cron job. So update these monthly on a cron job. So update these monthly on a cron job. So every month every month every month >> I will auto audit audit my entire >> I will auto audit audit my entire >> I will auto audit audit my entire ecosystem ecosystem ecosystem >> refine all my skills see where I'm not >> refine all my skills see where I'm not >> refine all my skills see where I'm not using skills that I've added because a using skills that I've added because a using skills that I've added because a lot of people your audience and mine lot of people your audience and mine lot of people your audience and mine have bloated skill repositories where have bloated skill repositories where have bloated skill repositories where they downloaded some awesome skills they downloaded some awesome skills they downloaded some awesome skills thing they have 300 of them they load thing they have 300 of them they load thing they have 300 of them they load every single time and they use five. every single time and they use five. every single time and they use five. >> So it reduces the number of skills that >> So it reduces the number of skills that >> So it reduces the number of skills that I have. It combines the ones where I have. It combines the ones where I have. It combines the ones where there's an opportunity for a compound there's an opportunity for a compound there's an opportunity for a compound skill and then it makes sure that skill and then it makes sure that skill and then it makes sure that they're all model agnostic. I love that. they're all model agnostic. I love that. they're all model agnostic. I love that. Yeah. I think what's really important Yeah. I think what's really important Yeah. I think what's really important there is, you know, what you said, you there is, you know, what you said, you there is, you know, what you said, you have AI do the dirty work, but you are have AI do the dirty work, but you are have AI do the dirty work, but you are still very much in control and you still still very much in control and you still still very much in control and you still understand. And I always think of the understand. And I always think of the understand. And I always think of the quote, you can outsource the thinking, quote, you can outsource the thinking, quote, you can outsource the thinking, but you can never outsource the but you can never outsource the but you can never outsource the understanding. Mhm. understanding. Mhm. understanding. Mhm. >> And I think it's a great mindset shift >> And I think it's a great mindset shift >> And I think it's a great mindset shift to realize that to realize that to realize that even us as creators, a lot of the things even us as creators, a lot of the things even us as creators, a lot of the things that we don't know or that we need to that we don't know or that we need to that we don't know or that we need to learn, we have AI help us with it, but learn, we have AI help us with it, but learn, we have AI help us with it, but we still understand we still understand we still understand how to feed in like the documentation how to feed in like the documentation how to feed in like the documentation for it to look through and we understand for it to look through and we understand for it to look through and we understand now that it has this knowledge what to now that it has this knowledge what to now that it has this knowledge what to do with it. And I always kind of say do with it. And I always kind of say do with it. And I always kind of say this in my videos even though it might this in my videos even though it might this in my videos even though it might like hurt the views. Sure.

like hurt the views. Sure. like hurt the views. Sure. >> Is that ultimately like just use it as >> Is that ultimately like just use it as >> Is that ultimately like just use it as your your thought partner as long as your your thought partner as long as your your thought partner as long as you're not outsourcing everything. I you're not outsourcing everything. I you're not outsourcing everything. I think that's a really important way to think that's a really important way to think that's a really important way to think about how you think about how you think about how you >> as a person continue to learn more too >> as a person continue to learn more too >> as a person continue to learn more too as well. as well. as well. >> Absolutely. And one thing you can do is >> Absolutely. And one thing you can do is >> Absolutely. And one thing you can do is again if we move into this world of you again if we move into this world of you again if we move into this world of you owning your own harness, you can have owning your own harness, you can have owning your own harness, you can have the same task be executed and then cla the same task be executed and then cla the same task be executed and then cla can watch it and it can run it within can watch it and it can run it within can watch it and it can run it within the terminal on your PI harness, run it the terminal on your PI harness, run it the terminal on your PI harness, run it on the other model providers, observe on the other model providers, observe on the other model providers, observe exactly what happened and what was the exactly what happened and what was the exactly what happened and what was the end result and try to continually end result and try to continually end result and try to continually understand and reverse engineer what understand and reverse engineer what understand and reverse engineer what happened with the others that's not happened with the others that's not happened with the others that's not happening with yours. happening with yours. happening with yours. >> Mhm. Back in March of 26, we had the >> Mhm. Back in March of 26, we had the >> Mhm. Back in March of 26, we had the cloud code harness leak. If you remember cloud code harness leak. If you remember cloud code harness leak. If you remember that, it was a map file. It was leaked that, it was a map file. It was leaked that, it was a map file. It was leaked to the whole world. to the whole world. to the whole world. >> I spent four to five days, and I I'm >> I spent four to five days, and I I'm >> I spent four to five days, and I I'm pretty sure you went and made a couple pretty sure you went and made a couple pretty sure you went and made a couple videos as well, videos as well, videos as well, >> looking through every single piece of >> looking through every single piece of >> looking through every single piece of it. Mhm. it. Mhm. it. Mhm. >> And the coolest part was the majority of >> And the coolest part was the majority of >> And the coolest part was the majority of the harness was full of all these the harness was full of all these the harness was full of all these crutches they gave to the brain, the crutches they gave to the brain, the crutches they gave to the brain, the model to not swear at the user to detect model to not swear at the user to detect model to not swear at the user to detect when you had swear words when you had swear words when you had swear words >> through a list of regex of all the swear >> through a list of regex of all the swear >> through a list of regex of all the swear words you would have that would tell the words you would have that would tell the words you would have that would tell the brain how to react. So once I saw there brain how to react. So once I saw there brain how to react. So once I saw there were so many crutches for this were so many crutches for this were so many crutches for this supposedly AGI level model, it made me supposedly AGI level model, it made me supposedly AGI level model, it made me wake up. Ah, this stuff isn't like ma wake up. Ah, this stuff isn't like ma wake up. Ah, this stuff isn't like ma troop magic at all. there is a whole troop magic at all. there is a whole troop magic at all. there is a whole factory of workers that are making this factory of workers that are making this factory of workers that are making this model look way smarter than it is.

model look way smarter than it is. model look way smarter than it is. >> And as soon as you understand that one >> And as soon as you understand that one >> And as soon as you understand that one concept, that's when you snap out of it concept, that's when you snap out of it concept, that's when you snap out of it and you start looking at model and you start looking at model and you start looking at model benchmarks very differently, you are not benchmarks very differently, you are not benchmarks very differently, you are not as wowed by what bench it crushed. as wowed by what bench it crushed. as wowed by what bench it crushed. You're more wowed by how well the model You're more wowed by how well the model You're more wowed by how well the model provider do in now improving their provider do in now improving their provider do in now improving their harness to have a symbiotic relationship harness to have a symbiotic relationship harness to have a symbiotic relationship with this brand new model's brain. with this brand new model's brain. with this brand new model's brain. That's how that's why I care about That's how that's why I care about That's how that's why I care about empirical. How well does this do when I empirical. How well does this do when I empirical. How well does this do when I push it versus how well are they telling push it versus how well are they telling push it versus how well are they telling me it should perform based on how smart me it should perform based on how smart me it should perform based on how smart it is? it is? it is? >> Absolutely. There there's a question I >> Absolutely. There there's a question I >> Absolutely. There there's a question I wanted to ask you that I get a ton and wanted to ask you that I get a ton and wanted to ask you that I get a ton and um it it it revolves around this whole um it it it revolves around this whole um it it it revolves around this whole harness idea. It revolves specifically harness idea. It revolves specifically harness idea. It revolves specifically around this idea of around this idea of around this idea of >> building out your own second brain or >> building out your own second brain or >> building out your own second brain or OS. Mhm. OS. Mhm. OS. Mhm. >> The question that I get a lot is about >> The question that I get a lot is about >> The question that I get a lot is about as you every month are adding in new as you every month are adding in new as you every month are adding in new things, whether that be skills or um you things, whether that be skills or um you things, whether that be skills or um you know an LM wiki, how do you yourself know an LM wiki, how do you yourself know an LM wiki, how do you yourself make sure that it's staying optimized? make sure that it's staying optimized? make sure that it's staying optimized? And I put that in air quotes because I And I put that in air quotes because I And I put that in air quotes because I don't I truly don't believe that there's don't I truly don't believe that there's don't I truly don't believe that there's only one optimal way to do it. I think only one optimal way to do it. I think only one optimal way to do it. I think it's just a matter of making sure that it's just a matter of making sure that it's just a matter of making sure that you can feel when it's maybe searching you can feel when it's maybe searching you can feel when it's maybe searching too long for something that it should too long for something that it should too long for something that it should find right away or it's hallucinating find right away or it's hallucinating find right away or it's hallucinating information because of the bloat in a information because of the bloat in a information because of the bloat in a certain folder. So I would love to hear certain folder. So I would love to hear certain folder. So I would love to hear just kind of like dive into your brain a just kind of like dive into your brain a just kind of like dive into your brain a little bit.

little bit. little bit. >> How do you think about keeping that >> How do you think about keeping that >> How do you think about keeping that organized and organized and organized and >> efficient? Sure. >> efficient? Sure. >> efficient? Sure. >> So the biggest thing that I've done is >> So the biggest thing that I've done is >> So the biggest thing that I've done is create what's called a rot.md create what's called a rot.md create what's called a rot.md file. file. file. >> Okay. >> Okay. >> Okay. >> Rot meaning decay. So if you have >> Rot meaning decay. So if you have >> Rot meaning decay. So if you have different layers of an AIOS and you have different layers of an AIOS and you have different layers of an AIOS and you have entire courses on this, you you covered entire courses on this, you you covered entire courses on this, you you covered this at length. You have five to six this at length. You have five to six this at length. You have five to six layers. One could be your identity, then layers. One could be your identity, then layers. One could be your identity, then your substrate, which is your core your substrate, which is your core your substrate, which is your core context that shouldn't change that much. context that shouldn't change that much. context that shouldn't change that much. Then you have your skills, your rules, Then you have your skills, your rules, Then you have your skills, your rules, your hooks, you have your agents, and your hooks, you have your agents, and your hooks, you have your agents, and then you have additional things you can then you have additional things you can then you have additional things you can add on. All of these different parts add on. All of these different parts add on. All of these different parts decay, become obsolete or rot at decay, become obsolete or rot at decay, become obsolete or rot at different rates. Mhm. different rates. Mhm. different rates. Mhm. >> So who you are, what you do, your goals, >> So who you are, what you do, your goals, >> So who you are, what you do, your goals, your aspirations unlikely to change very your aspirations unlikely to change very your aspirations unlikely to change very quickly. quickly. quickly. >> So you could have 1 to 3 months maybe >> So you could have 1 to 3 months maybe >> So you could have 1 to 3 months maybe even maybe longer if it's a company even maybe longer if it's a company even maybe longer if it's a company actually using this where it's relevant. actually using this where it's relevant. actually using this where it's relevant. >> Skills like I said I optimate up I >> Skills like I said I optimate up I >> Skills like I said I optimate up I update and optimize on a monthly basis. update and optimize on a monthly basis. update and optimize on a monthly basis. I just made a brand new one. [laughter] I just made a brand new one. [laughter] I just made a brand new one. [laughter] >> I like the word actually. >> I like the word actually. >> I like the word actually. >> And then you have let's say your rules. >> And then you have let's say your rules. >> And then you have let's say your rules. >> Your rules can change daily. I have >> Your rules can change daily. I have >> Your rules can change daily. I have rules that change daily because as I do rules that change daily because as I do rules that change daily because as I do new tasks with the same AIOS, I find new new tasks with the same AIOS, I find new new tasks with the same AIOS, I find new limitations and new edge cases.

limitations and new edge cases. limitations and new edge cases. >> So I find ways to refine and make better >> So I find ways to refine and make better >> So I find ways to refine and make better skills that are more compound. skills that are more compound. skills that are more compound. >> My rules probably decay every week. >> My rules probably decay every week. >> My rules probably decay every week. >> Yeah. >> Yeah. >> Yeah. >> My hooks though, so things that it >> My hooks though, so things that it >> My hooks though, so things that it always checks. So if I push a client always checks. So if I push a client always checks. So if I push a client project to GitHub, it wants to make sure project to GitHub, it wants to make sure project to GitHub, it wants to make sure that there's a hook that fires to remove that there's a hook that fires to remove that there's a hook that fires to remove any PII, any sensitive data of that any PII, any sensitive data of that any PII, any sensitive data of that client that I don't want living on my client that I don't want living on my client that I don't want living on my GitHub. GitHub. GitHub. >> Mhm. >> Mhm. >> Mhm. >> That that might not decay for 6 months. >> That that might not decay for 6 months. >> That that might not decay for 6 months. I might change it a little bit, but it I might change it a little bit, but it I might change it a little bit, but it won't really be viable to be decaying at won't really be viable to be decaying at won't really be viable to be decaying at a very high rate. Skills and agents a very high rate. Skills and agents a very high rate. Skills and agents though decay incredibly fast to the though decay incredibly fast to the though decay incredibly fast to the point where Boris Churnney dropped a point where Boris Churnney dropped a point where Boris Churnney dropped a tweet saying you should delete all of tweet saying you should delete all of tweet saying you should delete all of your skills every 6 months. All of them. your skills every 6 months. All of them. your skills every 6 months. All of them. >> Did he? >> Did he? >> Did he? >> Do you know why? >> Do you know why? >> Do you know why? >> Why did he say that? Because a skill is >> Why did he say that? Because a skill is >> Why did he say that? Because a skill is basically an e extra crutch that we're basically an e extra crutch that we're basically an e extra crutch that we're adding to the system to help the brain adding to the system to help the brain adding to the system to help the brain do things or the brain better understand do things or the brain better understand do things or the brain better understand how to use all these tools to do the how to use all these tools to do the how to use all these tools to do the thing. But if the model truly gets way thing. But if the model truly gets way thing. But if the model truly gets way smarter, it might not need the skill to smarter, it might not need the skill to smarter, it might not need the skill to shortcut what tools to use at disposal.

shortcut what tools to use at disposal. shortcut what tools to use at disposal. It might be able to accomplish the goal It might be able to accomplish the goal It might be able to accomplish the goal of the skill with purely a vague prompt of the skill with purely a vague prompt of the skill with purely a vague prompt versus the vague prompt plus the skill versus the vague prompt plus the skill versus the vague prompt plus the skill >> that's injected every single time. So >> that's injected every single time. So >> that's injected every single time. So you want to make sure that your skill is you want to make sure that your skill is you want to make sure that your skill is actually adding value and not holding actually adding value and not holding actually adding value and not holding back this dragon that gets only bigger back this dragon that gets only bigger back this dragon that gets only bigger with time and smarter with time and more with time and smarter with time and more with time and smarter with time and more powerful. So skills is one thing you powerful. So skills is one thing you powerful. So skills is one thing you want to audit very aggressively because want to audit very aggressively because want to audit very aggressively because you might not have a skill issue with you might not have a skill issue with you might not have a skill issue with that thing you built it for 3 months that thing you built it for 3 months that thing you built it for 3 months down the line, 6 months down the line down the line, 6 months down the line down the line, 6 months down the line with things like let's say agents. You with things like let's say agents. You with things like let's say agents. You hire agents like you hire employees. You hire agents like you hire employees. You hire agents like you hire employees. You might not always need a bookkeeping might not always need a bookkeeping might not always need a bookkeeping agent because if the model gets good agent because if the model gets good agent because if the model gets good enough, you might be able to give vague enough, you might be able to give vague enough, you might be able to give vague prompt. It would know exactly what are prompt. It would know exactly what are prompt. It would know exactly what are the 18 to 20 different subtasks based on the 18 to 20 different subtasks based on the 18 to 20 different subtasks based on all the memory assuming your memory gets all the memory assuming your memory gets all the memory assuming your memory gets better that it needs to execute. better that it needs to execute. better that it needs to execute. >> So we could live in a world where you >> So we could live in a world where you >> So we could live in a world where you have a handful of skills, a handful of have a handful of skills, a handful of have a handful of skills, a handful of rules, and those skills are actually rules, and those skills are actually rules, and those skills are actually very specific task at knowledge that it very specific task at knowledge that it very specific task at knowledge that it would never know no matter how smart it would never know no matter how smart it would never know no matter how smart it was versus step-by-step instructions.

was versus step-by-step instructions. was versus step-by-step instructions. Wow. Yeah, that is really interesting. Wow. Yeah, that is really interesting. Wow. Yeah, that is really interesting. You know, I always make it a point to You know, I always make it a point to You know, I always make it a point to every time a new model comes out and I every time a new model comes out and I every time a new model comes out and I have, you know, I've got the new Opus have, you know, I've got the new Opus have, you know, I've got the new Opus model or whatever plugged into Cloud model or whatever plugged into Cloud model or whatever plugged into Cloud Code, I always make it a point to have Code, I always make it a point to have Code, I always make it a point to have that model run through all my skills that model run through all my skills that model run through all my skills just to make sure that it can use them just to make sure that it can use them just to make sure that it can use them in the same way and and things like in the same way and and things like in the same way and and things like that. But I have never thought really that. But I have never thought really that. But I have never thought really to, you know, I hadn't seen that tweet. to, you know, I hadn't seen that tweet. to, you know, I hadn't seen that tweet. I had I hadn't thought to go back to I had I hadn't thought to go back to I had I hadn't thought to go back to some of the like through all of them and some of the like through all of them and some of the like through all of them and say like, do we even need this? And what say like, do we even need this? And what say like, do we even need this? And what happens if you try to run that process happens if you try to run that process happens if you try to run that process without the skill? without the skill? without the skill? >> Exactly. So >> Exactly. So >> Exactly. So >> given your meteoric rise on YouTube, if >> given your meteoric rise on YouTube, if >> given your meteoric rise on YouTube, if I told you, hey Nate, here's how you I told you, hey Nate, here's how you I told you, hey Nate, here's how you make a YouTube video, right? Two years make a YouTube video, right? Two years make a YouTube video, right? Two years ago, this might have been a helpful ago, this might have been a helpful ago, this might have been a helpful skill for you to learn from me before I skill for you to learn from me before I skill for you to learn from me before I I started before you. But now I gave you I started before you. But now I gave you I started before you. But now I gave you that same play. You'd look at me, tilt that same play. You'd look at me, tilt that same play. You'd look at me, tilt your head, and be like, have you looked your head, and be like, have you looked your head, and be like, have you looked at our subscriber [laughter] counts? at our subscriber [laughter] counts? at our subscriber [laughter] counts? >> Right? You don't need this skill >> Right? You don't need this skill >> Right? You don't need this skill anymore. You have outgrown this skill. anymore. You have outgrown this skill. anymore. You have outgrown this skill. >> That's really interesting. So as a model >> That's really interesting. So as a model >> That's really interesting. So as a model becomes very diverse and has training becomes very diverse and has training becomes very diverse and has training data, you might not need the skill data, you might not need the skill data, you might not need the skill anymore. It might not it might be anymore. It might not it might be anymore. It might not it might be obsolete for where the model's at.

obsolete for where the model's at. obsolete for where the model's at. >> Have you had an example yourself where >> Have you had an example yourself where >> Have you had an example yourself where you've been able to remove a skill you've been able to remove a skill you've been able to remove a skill because the model and the artist now because the model and the artist now because the model and the artist now just can crush it out of the water? just can crush it out of the water? just can crush it out of the water? >> Basically now if I tell it go and read >> Basically now if I tell it go and read >> Basically now if I tell it go and read all the conversations that we've had and all the conversations that we've had and all the conversations that we've had and tell me 15 different things we can tell me 15 different things we can tell me 15 different things we can optimize. Mhm. optimize. Mhm. optimize. Mhm. >> I used to have a skill that would tell >> I used to have a skill that would tell >> I used to have a skill that would tell exactly where to look for the files that exactly where to look for the files that exactly where to look for the files that are associated with our conversations, are associated with our conversations, are associated with our conversations, how to break it down so it doesn't blow how to break it down so it doesn't blow how to break it down so it doesn't blow its context window with 50,000 tokens at its context window with 50,000 tokens at its context window with 50,000 tokens at once times 100. once times 100. once times 100. >> I used to have to really explain that >> I used to have to really explain that >> I used to have to really explain that bit by bit. bit by bit. bit by bit. >> And now I went from skill to now we're >> And now I went from skill to now we're >> And now I went from skill to now we're moving to skill and now I have one basic moving to skill and now I have one basic moving to skill and now I have one basic sentence and it knows exactly to sentence and it knows exactly to sentence and it knows exactly to optimize based on my cloud. I'm always optimize based on my cloud. I'm always optimize based on my cloud. I'm always trying to optimize my uh context window trying to optimize my uh context window trying to optimize my uh context window and my context use. it on its own can do and my context use. it on its own can do and my context use. it on its own can do that without me telling it to. that without me telling it to. that without me telling it to. >> That, you know, that makes me think of >> That, you know, that makes me think of >> That, you know, that makes me think of something that's that's really something that's that's really something that's that's really interesting because of that ability for interesting because of that ability for interesting because of that ability for it to be more creative and arguably a it to be more creative and arguably a it to be more creative and arguably a better problem solver because it knows better problem solver because it knows better problem solver because it knows how to get to that end goal. Mhm. how to get to that end goal. Mhm. how to get to that end goal. Mhm. >> That might also be kind of a scary thing >> That might also be kind of a scary thing >> That might also be kind of a scary thing because in some cases maybe the skill because in some cases maybe the skill because in some cases maybe the skill keeps it on stricter guard rails keeps it on stricter guard rails keeps it on stricter guard rails >> potentially >> potentially >> potentially >> because sometimes now if it has to try >> because sometimes now if it has to try >> because sometimes now if it has to try to find its own way to from point A to to find its own way to from point A to to find its own way to from point A to point B, it might try to create some point B, it might try to create some point B, it might try to create some things that could potentially be harmful things that could potentially be harmful things that could potentially be harmful to the system or something like that as to the system or something like that as to the system or something like that as well.

well. well. >> But that's where the rest of the AIOS >> But that's where the rest of the AIOS >> But that's where the rest of the AIOS comes in because like the AIOS is not comes in because like the AIOS is not comes in because like the AIOS is not just a skill. You have the rules, you just a skill. You have the rules, you just a skill. You have the rules, you have different layers to babysit. have different layers to babysit. have different layers to babysit. >> So we might live in a world where we we >> So we might live in a world where we we >> So we might live in a world where we we need way less skills and much stricter need way less skills and much stricter need way less skills and much stricter rules. And your cloud MD now just rules. And your cloud MD now just rules. And your cloud MD now just basically says go and follow these rules basically says go and follow these rules basically says go and follow these rules for these kind of situations for these kind of situations for these kind of situations >> and that could be sufficient enough. >> and that could be sufficient enough. >> and that could be sufficient enough. >> Yeah. >> Yeah. >> Yeah. >> So some of these layers, even the >> So some of these layers, even the >> So some of these layers, even the concept of assigning an agent and concept of assigning an agent and concept of assigning an agent and creating your own agent that's all spun creating your own agent that's all spun creating your own agent that's all spun up the same way every time, that could up the same way every time, that could up the same way every time, that could become obsolete. So, I'm not saying it become obsolete. So, I'm not saying it become obsolete. So, I'm not saying it is, but I'm saying we should be is, but I'm saying we should be is, but I'm saying we should be open-minded that this stuff will evolve open-minded that this stuff will evolve open-minded that this stuff will evolve and as long as you keep your core assets and as long as you keep your core assets and as long as you keep your core assets nimble, then you should be good to go. I nimble, then you should be good to go. I nimble, then you should be good to go. I love that. You know, I did see which I love that. You know, I did see which I love that. You know, I did see which I thought was really interesting. I don't thought was really interesting. I don't thought was really interesting. I don't remember if it was Opus 5 or Fable 5, remember if it was Opus 5 or Fable 5, remember if it was Opus 5 or Fable 5, but one of those drops but one of those drops but one of those drops >> from Enthropic, you know, how they >> from Enthropic, you know, how they >> from Enthropic, you know, how they dropped the blog with benchmarks and dropped the blog with benchmarks and dropped the blog with benchmarks and just like how to use it, how to prompt just like how to use it, how to prompt just like how to use it, how to prompt it. it. it. One of the things I remember being at One of the things I remember being at One of the things I remember being at the top was the top was the top was give it more ambitious tasks. And you give it more ambitious tasks. And you give it more ambitious tasks. And you might just skim through that, right? But might just skim through that, right? But might just skim through that, right? But to me, that made me think, well, why do to me, that made me think, well, why do to me, that made me think, well, why do they feel the need to put that in there?

they feel the need to put that in there? they feel the need to put that in there? Maybe they feel like people aren't Maybe they feel like people aren't Maybe they feel like people aren't pushing the models to their true limits pushing the models to their true limits pushing the models to their true limits and getting out of the model's way. And and getting out of the model's way. And and getting out of the model's way. And that's when I started to run these just that's when I started to run these just that's when I started to run these just interesting experiments. I'd set a SL interesting experiments. I'd set a SL interesting experiments. I'd set a SL goal and I'd say like kind of something goal and I'd say like kind of something goal and I'd say like kind of something along the lines of like impress me like along the lines of like impress me like along the lines of like impress me like build me build me this but impress me build me build me this but impress me build me build me this but impress me and show me what you can really do. And and show me what you can really do. And and show me what you can really do. And I just thought that that was interesting I just thought that that was interesting I just thought that that was interesting that they put that in there and it just that they put that in there and it just that they put that in there and it just shows like that kind of makes me think shows like that kind of makes me think shows like that kind of makes me think that whole skill thing like the skill that whole skill thing like the skill that whole skill thing like the skill that we we maybe have baked in a long that we we maybe have baked in a long that we we maybe have baked in a long time ago and now we just kind of blindly time ago and now we just kind of blindly time ago and now we just kind of blindly use because it it works. we are almost use because it it works. we are almost use because it it works. we are almost yeah kind of yeah kind of yeah kind of baking in or or limiting what the model baking in or or limiting what the model baking in or or limiting what the model could truly do because it's it's kind of could truly do because it's it's kind of could truly do because it's it's kind of going down these these guardrails. So I going down these these guardrails. So I going down these these guardrails. So I think that's that's really interesting. think that's that's really interesting. think that's that's really interesting. You know, as we sort of start to wrap up You know, as we sort of start to wrap up You know, as we sort of start to wrap up here today, I'd be interested to hear here today, I'd be interested to hear here today, I'd be interested to hear from from you. What is something that from from you. What is something that from from you. What is something that you think that you do within your you think that you do within your you think that you do within your cloud code or your harness setup that cloud code or your harness setup that cloud code or your harness setup that most people don't do and that you think most people don't do and that you think most people don't do and that you think is important that everyone should be is important that everyone should be is important that everyone should be doing? doing? doing? >> So the main thing I think about is many >> So the main thing I think about is many >> So the main thing I think about is many times I have let's say 25 different times I have let's say 25 different times I have let's say 25 different operating systems.

operating systems. operating systems. >> So some people like to create one big >> So some people like to create one big >> So some people like to create one big one. one. one. >> I create very specific ones that live in >> I create very specific ones that live in >> I create very specific ones that live in an isolated world. an isolated world. an isolated world. >> Okay. So my tax and finance lives very >> Okay. So my tax and finance lives very >> Okay. So my tax and finance lives very differently from my consulting OS which differently from my consulting OS which differently from my consulting OS which is very living very differently from my is very living very differently from my is very living very differently from my school OS for content for there versus school OS for content for there versus school OS for content for there versus everything else. everything else. everything else. >> So me segmenting every single part of my >> So me segmenting every single part of my >> So me segmenting every single part of my business my education offers everything business my education offers everything business my education offers everything we do for enterprise clients each thing we do for enterprise clients each thing we do for enterprise clients each thing has its own set of operating systems. has its own set of operating systems. has its own set of operating systems. >> It's more to upkeep for sure but as you >> It's more to upkeep for sure but as you >> It's more to upkeep for sure but as you build more you start to build more build more you start to build more build more you start to build more leverage. So I have one mega system that leverage. So I have one mega system that leverage. So I have one mega system that depending on its audit of every single depending on its audit of every single depending on its audit of every single operating system I have will come up operating system I have will come up operating system I have will come up with a series of things that we need to with a series of things that we need to with a series of things that we need to make that specific thing better that make that specific thing better that make that specific thing better that specific operating system work better specific operating system work better specific operating system work better and one that generalizes across all of and one that generalizes across all of and one that generalizes across all of them. them. them. >> So as you create different folders >> So as you create different folders >> So as you create different folders you'll have some project level skills you'll have some project level skills you'll have some project level skills rules cloud and then global. Mhm. rules cloud and then global. Mhm. rules cloud and then global. Mhm. >> Some people make everything global, >> Some people make everything global, >> Some people make everything global, which is awful because as you add new which is awful because as you add new which is awful because as you add new things, you might notice under things, you might notice under things, you might notice under performance cuz you forgot that you have performance cuz you forgot that you have performance cuz you forgot that you have this global spectre of rules applying to this global spectre of rules applying to this global spectre of rules applying to every single thing.

every single thing. every single thing. >> So for me, I have a different concept >> So for me, I have a different concept >> So for me, I have a different concept where I think about promotion. where I think about promotion. where I think about promotion. >> Everything is project until deserves to >> Everything is project until deserves to >> Everything is project until deserves to be promoted to global so that I know at be promoted to global so that I know at be promoted to global so that I know at all times what is the running total of all times what is the running total of all times what is the running total of everything that's running globally. everything that's running globally. everything that's running globally. >> And I have a very few number of things >> And I have a very few number of things >> And I have a very few number of things that are global. Everything's project that are global. Everything's project that are global. Everything's project specific. But that also gives me this specific. But that also gives me this specific. But that also gives me this the flexibility to have a very small the flexibility to have a very small the flexibility to have a very small blast radius if I want to be blast radius if I want to be blast radius if I want to be experimental. If I want to audit and experimental. If I want to audit and experimental. If I want to audit and push a certain folder for a certain type push a certain folder for a certain type push a certain folder for a certain type of task and not have that bleed over to of task and not have that bleed over to of task and not have that bleed over to another project and not know why is this another project and not know why is this another project and not know why is this not working, is the model worse? Is the not working, is the model worse? Is the not working, is the model worse? Is the harness worse? Or is my setup worse? So harness worse? Or is my setup worse? So harness worse? Or is my setup worse? So because I'm so meticulous about is this because I'm so meticulous about is this because I'm so meticulous about is this a model problem, is this a harness a model problem, is this a harness a model problem, is this a harness problem? Or is this an organization problem? Or is this an organization problem? Or is this an organization problem? when I can isolate things, I problem? when I can isolate things, I problem? when I can isolate things, I can find the issue faster. can find the issue faster. can find the issue faster. >> So, Fable is amazing with my tax, but >> So, Fable is amazing with my tax, but >> So, Fable is amazing with my tax, but all of a sudden, it is horrific with my all of a sudden, it is horrific with my all of a sudden, it is horrific with my consulting operating system. consulting operating system. consulting operating system. >> It might not be a model issue.

>> It might not be a model issue. >> It might not be a model issue. >> It might not be a skill issue. It could >> It might not be a skill issue. It could >> It might not be a skill issue. It could be an organization issue for that model be an organization issue for that model be an organization issue for that model that I might have to adhere to for this that I might have to adhere to for this that I might have to adhere to for this specific project. specific project. specific project. >> And that helps me control bloat and find >> And that helps me control bloat and find >> And that helps me control bloat and find the area of resistance that I need to the area of resistance that I need to the area of resistance that I need to move to actually get the reforms I'm move to actually get the reforms I'm move to actually get the reforms I'm looking for. I think that's really looking for. I think that's really looking for. I think that's really smart. I think especially because these smart. I think especially because these smart. I think especially because these things are so so autonomous and agentic. things are so so autonomous and agentic. things are so so autonomous and agentic. >> You have to be able to >> You have to be able to >> You have to be able to find the actual variable that that find the actual variable that that find the actual variable that that killed the thing. Otherwise, there is no killed the thing. Otherwise, there is no killed the thing. Otherwise, there is no learning there and there's no learning there and there's no learning there and there's no improvement there. improvement there. improvement there. >> And it it it for the most part you can >> And it it it for the most part you can >> And it it it for the most part you can help like it can help you find out why help like it can help you find out why help like it can help you find out why it went wrong and where. But I think it went wrong and where. But I think it went wrong and where. But I think that level of isolation and I think the that level of isolation and I think the that level of isolation and I think the key is there that you know how your key is there that you know how your key is there that you know how your different folders are drilled down. Yes, different folders are drilled down. Yes, different folders are drilled down. Yes, >> you know, you're not just blindly >> you know, you're not just blindly >> you know, you're not just blindly trusting that it can find everything. trusting that it can find everything. trusting that it can find everything. You still intuitively You still intuitively You still intuitively >> I have a feeling that if you are looking >> I have a feeling that if you are looking >> I have a feeling that if you are looking for a specific deliverable, you could for a specific deliverable, you could for a specific deliverable, you could find it just by clicking through your find it just by clicking through your find it just by clicking through your files and folders because you kind of files and folders because you kind of files and folders because you kind of know where the things live and you know know where the things live and you know know where the things live and you know how to drill down. how to drill down. how to drill down. >> Yes. >> Yes. >> Yes. >> And I think that that's just highlights >> And I think that that's just highlights >> And I think that that's just highlights the point again that you cannot the point again that you cannot the point again that you cannot outsource your understanding. You still outsource your understanding. You still outsource your understanding. You still have to understand where everything is have to understand where everything is have to understand where everything is and how it works.

and how it works. and how it works. >> Yeah. The last thing I was going to say >> Yeah. The last thing I was going to say >> Yeah. The last thing I was going to say is that this becomes an acquired skill. is that this becomes an acquired skill. is that this becomes an acquired skill. And if you want to get to mastery, And if you want to get to mastery, And if you want to get to mastery, mastery is just understanding how the mastery is just understanding how the mastery is just understanding how the entire factory works end to end and entire factory works end to end and entire factory works end to end and where each piece lies. So if you want where each piece lies. So if you want where each piece lies. So if you want more leverage, if you're running into more leverage, if you're running into more leverage, if you're running into issues where you're kind of being issues where you're kind of being issues where you're kind of being intellectually lazy and saying, "Oh, intellectually lazy and saying, "Oh, intellectually lazy and saying, "Oh, this model sucks." Before you pass fully this model sucks." Before you pass fully this model sucks." Before you pass fully judgment on it, you want to make sure judgment on it, you want to make sure judgment on it, you want to make sure that maybe it sucks, for the setup that that maybe it sucks, for the setup that that maybe it sucks, for the setup that you currently have. So when in doubt, you currently have. So when in doubt, you currently have. So when in doubt, audit your setup, audit your skills, audit your setup, audit your skills, audit your setup, audit your skills, audit the bloat because different models