GLM 5.2 Is Free And Beats Claude On Most Work. So Why Can't Companies Switch?

I tried GLM 5.2 and it blew my mind. By I tried GLM 5.2 and it blew my mind. By the end of this video, you should know the end of this video, you should know the end of this video, you should know where GLM 5.2, an open-source model, can where GLM 5.2, an open-source model, can where GLM 5.2, an open-source model, can be clawed, where it can safely replace be clawed, where it can safely replace be clawed, where it can safely replace an expensive model, and where switching an expensive model, and where switching an expensive model, and where switching models is a bit of a trap because you're models is a bit of a trap because you're models is a bit of a trap because you're not replacing a model call. You're not replacing a model call. You're not replacing a model call. You're actually replacing a whole work system. actually replacing a whole work system. actually replacing a whole work system. And that's the thing I want to draw And that's the thing I want to draw And that's the thing I want to draw through in this video. So, let me start through in this video. So, let me start through in this video. So, let me start at the beginning here. GLM 5.2 did not at the beginning here. GLM 5.2 did not at the beginning here. GLM 5.2 did not fake impress me. It actually impressed fake impress me. It actually impressed fake impress me. It actually impressed me because it's not just cheap, and it's me because it's not just cheap, and it's me because it's not just cheap, and it's very cheap to run on the cloud, it's very cheap to run on the cloud, it's very cheap to run on the cloud, it's free if you set up your own servers, and free if you set up your own servers, and free if you set up your own servers, and for a lot of normal work, it's for a lot of normal work, it's for a lot of normal work, it's incredibly good. It's It's often better incredibly good. It's It's often better incredibly good. It's It's often better than Claude. And when I say normal work, than Claude. And when I say normal work, than Claude. And when I say normal work, I mean the fat middle of everyday AI I mean the fat middle of everyday AI I mean the fat middle of everyday AI tasks, right? So, if you're setting up a tasks, right? So, if you're setting up a tasks, right? So, if you're setting up a brochure site for a client, if you have brochure site for a client, if you have brochure site for a client, if you have a PowerPoint outline, it's a pretty a PowerPoint outline, it's a pretty a PowerPoint outline, it's a pretty standard deck. For a first pass copy, standard deck. For a first pass copy, standard deck. For a first pass copy, routine synthesis, for coding tasks that routine synthesis, for coding tasks that routine synthesis, for coding tasks that are tackling familiar problem types in are tackling familiar problem types in are tackling familiar problem types in coding, these are tasks with familiar coding, these are tasks with familiar coding, these are tasks with familiar shapes, with lots of examples, with shapes, with lots of examples, with shapes, with lots of examples, with outputs that a human can check quickly. outputs that a human can check quickly. outputs that a human can check quickly. The nerdier phrase for this is that this The nerdier phrase for this is that this The nerdier phrase for this is that this is the middle of the distribution work is the middle of the distribution work is the middle of the distribution work for AI. In other words, what you are for AI. In other words, what you are for AI. In other words, what you are getting is what someone has tried with getting is what someone has tried with getting is what someone has tried with models millions of times before, where models millions of times before, where models millions of times before, where the answer pattern is pretty normal, and the answer pattern is pretty normal, and the answer pattern is pretty normal, and the output is pretty easy to inspect.

the output is pretty easy to inspect. the output is pretty easy to inspect. How many different brochure sites have How many different brochure sites have How many different brochure sites have you seen, right? In that world, GLM 5.2 you seen, right? In that world, GLM 5.2 you seen, right? In that world, GLM 5.2 is incredible. It's fast, it's cheap, is incredible. It's fast, it's cheap, is incredible. It's fast, it's cheap, it's easy, and it's extremely high it's easy, and it's extremely high it's easy, and it's extremely high quality. It's higher quality than quality. It's higher quality than quality. It's higher quality than Claude. And a lot of those tasks, I Claude. And a lot of those tasks, I Claude. And a lot of those tasks, I don't think it's honest to say it's just don't think it's honest to say it's just don't think it's honest to say it's just good enough. I think it's It's more good enough. I think it's It's more good enough. I think it's It's more accurate to say this is the best model accurate to say this is the best model accurate to say this is the best model in the world at those center of in the world at those center of in the world at those center of distribution kinds of tasks, especially distribution kinds of tasks, especially distribution kinds of tasks, especially ones where front-end taste is important. ones where front-end taste is important. ones where front-end taste is important. And so, this is not a video about GLM And so, this is not a video about GLM And so, this is not a video about GLM 5.2 being bad, even though it's not my 5.2 being bad, even though it's not my 5.2 being bad, even though it's not my daily driver, and I'm going to explain daily driver, and I'm going to explain daily driver, and I'm going to explain why. And so, GLM 5.2 is incredible, but why. And so, GLM 5.2 is incredible, but why. And so, GLM 5.2 is incredible, but I'm still not using it every day. And in I'm still not using it every day. And in I'm still not using it every day. And in fact, a lot of companies I know are fact, a lot of companies I know are fact, a lot of companies I know are really struggling with the idea that really struggling with the idea that really struggling with the idea that they want to transition to more of a they want to transition to more of a they want to transition to more of a generic router where they can route to generic router where they can route to generic router where they can route to the cheapest model available, but it's the cheapest model available, but it's the cheapest model available, but it's not actually easy to do in practice. Why not actually easy to do in practice. Why not actually easy to do in practice. Why is that, right? We're going to talk is that, right? We're going to talk is that, right? We're going to talk about why that is, talk about where open about why that is, talk about where open about why that is, talk about where open source is going, talk about what the source is going, talk about what the source is going, talk about what the shape of work looks like in 2026, and shape of work looks like in 2026, and shape of work looks like in 2026, and we're going to tie it back into GLM 5.2 we're going to tie it back into GLM 5.2 we're going to tie it back into GLM 5.2 and the way we actually need to build to and the way we actually need to build to and the way we actually need to build to take advantage of models like this.

take advantage of models like this. take advantage of models like this. Because cheap AI, it's not a theory Because cheap AI, it's not a theory Because cheap AI, it's not a theory anymore. Cheap incredible AI is here. In anymore. Cheap incredible AI is here. In anymore. Cheap incredible AI is here. In fact, it's going to be here more and fact, it's going to be here more and fact, it's going to be here more and more and more and more because the US more and more and more because the US more and more and more because the US government is now slowing down frontier government is now slowing down frontier government is now slowing down frontier model releases. 5.6 is the latest model model releases. 5.6 is the latest model model releases. 5.6 is the latest model to be affected. It's apparently going to to be affected. It's apparently going to to be affected. It's apparently going to be released customer by customer, which be released customer by customer, which be released customer by customer, which is code for we don't know when we're is code for we don't know when we're is code for we don't know when we're going to get it. For the first time, going to get it. For the first time, going to get it. For the first time, there is no defined expected cadence for there is no defined expected cadence for there is no defined expected cadence for future model releases that are frontier, future model releases that are frontier, future model releases that are frontier, even though the labs are still doing a even though the labs are still doing a even though the labs are still doing a phenomenal job training and phenomenal job training and phenomenal job training and reinforcement learning their models. And reinforcement learning their models. And reinforcement learning their models. And so we're going to have more and more of so we're going to have more and more of so we're going to have more and more of this open source conversation. And a lot this open source conversation. And a lot this open source conversation. And a lot of the open source conversation is of the open source conversation is of the open source conversation is frankly about moving down the cost frankly about moving down the cost frankly about moving down the cost curve, right? Because these frontier curve, right? Because these frontier curve, right? Because these frontier model costs are expensive. If you're model costs are expensive. If you're model costs are expensive. If you're running a company, they get really running a company, they get really running a company, they get really expensive. There are stories going expensive. There are stories going expensive. There are stories going around where the numbers are absolutely around where the numbers are absolutely around where the numbers are absolutely eye-popping. Like one engineer spending eye-popping. Like one engineer spending eye-popping. Like one engineer spending $80,000 in token costs in a week. That's $80,000 in token costs in a week. That's $80,000 in token costs in a week. That's a lot. So if you have that kind of a lot. So if you have that kind of a lot. So if you have that kind of pricing power, if people are spending pricing power, if people are spending pricing power, if people are spending tens of thousands of dollars a week on tens of thousands of dollars a week on tens of thousands of dollars a week on tokens, there's an a tremendous amount tokens, there's an a tremendous amount tokens, there's an a tremendous amount of incentive to make these models work.

of incentive to make these models work. of incentive to make these models work. So why is it so hard? Why are why are we So why is it so hard? Why are why are we So why is it so hard? Why are why are we not seeing a tremendous tipping point not seeing a tremendous tipping point not seeing a tremendous tipping point away? Why are we still seeing Anthropic away? Why are we still seeing Anthropic away? Why are we still seeing Anthropic growing their revenue like crazy, OpenAI growing their revenue like crazy, OpenAI growing their revenue like crazy, OpenAI growing their revenue like crazy when growing their revenue like crazy when growing their revenue like crazy when these incredible good models exist? these incredible good models exist? these incredible good models exist? Well, there's a number of factors to Well, there's a number of factors to Well, there's a number of factors to that, and I want to list them for you so that, and I want to list them for you so that, and I want to list them for you so that you can actually understand the that you can actually understand the that you can actually understand the perspective. This is based on talking perspective. This is based on talking perspective. This is based on talking with engineers at companies as well as with engineers at companies as well as with engineers at companies as well as with leaders. The first one is the with leaders. The first one is the with leaders. The first one is the ergonomics of work. If you are just ergonomics of work. If you are just ergonomics of work. If you are just trying to get something you've heard trying to get something you've heard trying to get something you've heard about, seen about, you have a frontier about, seen about, you have a frontier about, seen about, you have a frontier model at you have a frontier model at model at you have a frontier model at model at you have a frontier model at home on your phone, you just want access home on your phone, you just want access home on your phone, you just want access to that. There's a lot of employee to that. There's a lot of employee to that. There's a lot of employee pressure around Claude and around OpenAI pressure around Claude and around OpenAI pressure around Claude and around OpenAI in a way that there just isn't for open in a way that there just isn't for open in a way that there just isn't for open source models. So, that's one piece. Uh, source models. So, that's one piece. Uh, source models. So, that's one piece. Uh, and it's not small. Like, when people and it's not small. Like, when people and it's not small. Like, when people are asking for it vocally saying this are asking for it vocally saying this are asking for it vocally saying this will help my work, overburdened IT will help my work, overburdened IT will help my work, overburdened IT departments tend to listen to that. departments tend to listen to that. departments tend to listen to that. Number two, it is actually very, very Number two, it is actually very, very Number two, it is actually very, very difficult to correctly figure out difficult to correctly figure out difficult to correctly figure out whether your task load is center of whether your task load is center of whether your task load is center of distribution or edge of distribution distribution or edge of distribution distribution or edge of distribution weighted. If it's edge distribution weighted. If it's edge distribution weighted. If it's edge distribution weighted, you actually do want the weighted, you actually do want the weighted, you actually do want the frontier models. If it's center of frontier models. If it's center of frontier models. If it's center of distribution, the open source models are distribution, the open source models are distribution, the open source models are going to be really, really good because going to be really, really good because going to be really, really good because they're common patterns. But, people they're common patterns. But, people they're common patterns. But, people don't They're not used to measuring don't They're not used to measuring don't They're not used to measuring their work that way. Individuals aren't, their work that way. Individuals aren't, their work that way. Individuals aren't, teams aren't. if you're a company trying teams aren't. if you're a company trying teams aren't. if you're a company trying to figure out what is your model to figure out what is your model to figure out what is your model strategy, you kind of got to tackle strategy, you kind of got to tackle strategy, you kind of got to tackle what is your distribution of tasks? And what is your distribution of tasks? And what is your distribution of tasks? And almost no one has asked that question almost no one has asked that question almost no one has asked that question properly yet. And people are trying to properly yet. And people are trying to properly yet. And people are trying to figure out how to measure that. The figure out how to measure that. The figure out how to measure that. The folks that have gone the farthest, folks that have gone the farthest, folks that have gone the farthest, actually, are folks like Flo Crivello,

actually, are folks like Flo Crivello, actually, are folks like Flo Crivello, who is, uh, leading the Lindy team, and who is, uh, leading the Lindy team, and who is, uh, leading the Lindy team, and who very publicly wrote up his journey who very publicly wrote up his journey who very publicly wrote up his journey to a deep seek architecture away from to a deep seek architecture away from to a deep seek architecture away from Claude. And, you know, he saved a lot, Claude. And, you know, he saved a lot, Claude. And, you know, he saved a lot, etc., etc. But, he was also very honest etc., etc. But, he was also very honest etc., etc. But, he was also very honest about the fact that the Lindy team had about the fact that the Lindy team had about the fact that the Lindy team had to essentially rewrite their harness to essentially rewrite their harness to essentially rewrite their harness from scratch around deep seek, and they from scratch around deep seek, and they from scratch around deep seek, and they could not just take all of their systems could not just take all of their systems could not just take all of their systems for working with Claude, all of their for working with Claude, all of their for working with Claude, all of their prompts, all of the way they handle prompts, all of the way they handle prompts, all of the way they handle memory, all of their tool calls, and memory, all of their tool calls, and memory, all of their tool calls, and just automatically lift and shift. It just automatically lift and shift. It just automatically lift and shift. It doesn't work that way. These models need doesn't work that way. These models need doesn't work that way. These models need their own harnesses. their own harnesses. their own harnesses. He was incentivized to do that because He was incentivized to do that because He was incentivized to do that because he is literally serving AI as a service, he is literally serving AI as a service, he is literally serving AI as a service, and if he can deliver a cheaper and more and if he can deliver a cheaper and more and if he can deliver a cheaper and more effective service that hits his margin, effective service that hits his margin, effective service that hits his margin, and it's it's tremendously impactful. and it's it's tremendously impactful. and it's it's tremendously impactful. For folks who are using AI internally For folks who are using AI internally For folks who are using AI internally for coding or for back office for coding or for back office for coding or for back office automation, that ROI is not as clear, automation, that ROI is not as clear, automation, that ROI is not as clear, and the incentive to move is not as and the incentive to move is not as and the incentive to move is not as clear, either. And so, what I have seen, clear, either. And so, what I have seen, clear, either. And so, what I have seen, and I have seen anecdotes from this, not and I have seen anecdotes from this, not and I have seen anecdotes from this, not just from Flo, but from other folks that just from Flo, but from other folks that just from Flo, but from other folks that I know personally. I know entrepreneurs I know personally. I know entrepreneurs I know personally. I know entrepreneurs who are wrestling with this today. The who are wrestling with this today. The who are wrestling with this today. The ones who are actually making the jump to ones who are actually making the jump to ones who are actually making the jump to open source and dealing with the open source and dealing with the open source and dealing with the different system prob dealing with the different system prob dealing with the different system prob dealing with the different tool called dealing with a different tool called dealing with a different tool called dealing with a different memory architecture, etc. That different memory architecture, etc. That different memory architecture, etc. That is tuned around the fact that these are is tuned around the fact that these are is tuned around the fact that these are center of distribution models. Those center of distribution models. Those center of distribution models. Those guys or those gals are focused on ROI guys or those gals are focused on ROI guys or those gals are focused on ROI for a particular AI tool they have in for a particular AI tool they have in for a particular AI tool they have in market. Just like Lindy, they see value market. Just like Lindy, they see value market. Just like Lindy, they see value back in their pockets when they can cut back in their pockets when they can cut back in their pockets when they can cut their token costs. And for everyone their token costs. And for everyone their token costs. And for everyone else, because the incentive is not as else, because the incentive is not as else, because the incentive is not as strong, you don't have the same

strong, you don't have the same strong, you don't have the same commitment to wade through the challenge commitment to wade through the challenge commitment to wade through the challenge of building a harness. And that is not a of building a harness. And that is not a of building a harness. And that is not a small thing. And one of the things I small thing. And one of the things I small thing. And one of the things I want you to take away from this video is want you to take away from this video is want you to take away from this video is that a model can be an incredible that a model can be an incredible that a model can be an incredible [snorts] brain in a jar. And it it just [snorts] brain in a jar. And it it just [snorts] brain in a jar. And it it just isn't useful to you without a harness. isn't useful to you without a harness. isn't useful to you without a harness. And so this is why I pay a ton of And so this is why I pay a ton of And so this is why I pay a ton of attention to harness innovations. And I attention to harness innovations. And I attention to harness innovations. And I want to name a couple that are top of want to name a couple that are top of want to name a couple that are top of mind as we look at GLM 5.2 in context. mind as we look at GLM 5.2 in context. mind as we look at GLM 5.2 in context. First, I notice that GLM 5.2 was First, I notice that GLM 5.2 was First, I notice that GLM 5.2 was released with its own Codex clone released with its own Codex clone released with its own Codex clone harness. That's one piece that I pay harness. That's one piece that I pay harness. That's one piece that I pay attention to. It looks like the open attention to. It looks like the open attention to. It looks like the open source model makers are realizing they source model makers are realizing they source model makers are realizing they need to deliver harnesses as well. And need to deliver harnesses as well. And need to deliver harnesses as well. And so I would expect more innovation in so I would expect more innovation in so I would expect more innovation in that direction. I notice that Codex is that direction. I notice that Codex is that direction. I notice that Codex is starting to call out publicly that you starting to call out publicly that you starting to call out publicly that you can use Codex the harness without using can use Codex the harness without using can use Codex the harness without using any OpenAI model. That's notable because any OpenAI model. That's notable because any OpenAI model. That's notable because there's a different path to value for there's a different path to value for there's a different path to value for OpenAI there. Maybe OpenAI's models are OpenAI there. Maybe OpenAI's models are OpenAI there. Maybe OpenAI's models are the default, but if they're calling out the default, but if they're calling out the default, but if they're calling out that they are actually the harness for that they are actually the harness for that they are actually the harness for all of work, it gives them a way to be all of work, it gives them a way to be all of work, it gives them a way to be stickier long term. Three, the Anthropic stickier long term. Three, the Anthropic stickier long term. Three, the Anthropic team is not just sitting there as all of team is not just sitting there as all of team is not just sitting there as all of these developments happen. They launched these developments happen. They launched these developments happen. They launched Claude Tag this week, and Claude Tag is Claude Tag this week, and Claude Tag is Claude Tag this week, and Claude Tag is an incredibly sticky product. It is a an incredibly sticky product. It is a an incredibly sticky product. It is a team level harness, and team level team level harness, and team level team level harness, and team level harnesses are where the energy is going harnesses are where the energy is going harnesses are where the energy is going because so much of the work we've got is because so much of the work we've got is because so much of the work we've got is individually productive work in AI. It's individually productive work in AI. It's individually productive work in AI. It's not team productive work. And we're not team productive work. And we're not team productive work. And we're trying to figure out, how do we align trying to figure out, how do we align trying to figure out, how do we align our efforts that are individually our efforts that are individually our efforts that are individually productive into something that is team productive into something that is team productive into something that is team productive? And Claude tag, which is productive? And Claude tag, which is productive? And Claude tag, which is just tag Claude, anyone can tag Claude just tag Claude, anyone can tag Claude just tag Claude, anyone can tag Claude and get work done in Slack, is one of

and get work done in Slack, is one of and get work done in Slack, is one of the first examples of a sticky viral the first examples of a sticky viral the first examples of a sticky viral consumer team harness. Where like if consumer team harness. Where like if consumer team harness. Where like if you're an ordinary knowledge worker at a you're an ordinary knowledge worker at a you're an ordinary knowledge worker at a particular company, you can envision particular company, you can envision particular company, you can envision using that as as a team harness. And you using that as as a team harness. And you using that as as a team harness. And you don't have to know the word team don't have to know the word team don't have to know the word team harness, it's just going to work. You harness, it's just going to work. You harness, it's just going to work. You tag Claude and it works. But look at it tag Claude and it works. But look at it tag Claude and it works. But look at it strategically from Anthropic's strategically from Anthropic's strategically from Anthropic's perspective. Now they're not just perspective. Now they're not just perspective. Now they're not just getting the engineers. Now they're getting the engineers. Now they're getting the engineers. Now they're getting everybody who's a knowledge getting everybody who's a knowledge getting everybody who's a knowledge worker in Slack and they're reading all worker in Slack and they're reading all worker in Slack and they're reading all of the messy context that lives in Slack of the messy context that lives in Slack of the messy context that lives in Slack that no one knows how to codify and that that no one knows how to codify and that that no one knows how to codify and that is now getting fed into Claude is now getting fed into Claude is now getting fed into Claude automatically and it can be something automatically and it can be something automatically and it can be something that the Anthropic team learns from that the Anthropic team learns from that the Anthropic team learns from within privacy policies long term for within privacy policies long term for within privacy policies long term for Claude in the context of that company to Claude in the context of that company to Claude in the context of that company to start to own the harness itself in a way start to own the harness itself in a way start to own the harness itself in a way that no company can get away from. It's that no company can get away from. It's that no company can get away from. It's an incredibly sticky experience because an incredibly sticky experience because an incredibly sticky experience because you think about it. Let's say you you you think about it. Let's say you you you think about it. Let's say you you know that GLM 5.2 is a lot cheaper, know that GLM 5.2 is a lot cheaper, know that GLM 5.2 is a lot cheaper, which it is. It's like 98% cheaper or which it is. It's like 98% cheaper or which it is. It's like 98% cheaper or something like that. If it's that much something like that. If it's that much something like that. If it's that much cheaper than Claude and it's just about cheaper than Claude and it's just about cheaper than Claude and it's just about as good on most tasks, it is rational to as good on most tasks, it is rational to as good on most tasks, it is rational to build a routing system and assign most build a routing system and assign most build a routing system and assign most tasks to GLM 5.2. Except that hey, are tasks to GLM 5.2. Except that hey, are tasks to GLM 5.2. Except that hey, are you going to have Claude tag, right? Are you going to have Claude tag, right? Are you going to have Claude tag, right? Are you going to go to tag in Claude on that you going to go to tag in Claude on that you going to go to tag in Claude on that stuff? Is that convenience going to be stuff? Is that convenience going to be stuff? Is that convenience going to be there? Are are you going to have to there? Are are you going to have to there? Are are you going to have to restart the job of giving this AI restart the job of giving this AI restart the job of giving this AI context from your company because Claude context from your company because Claude context from your company because Claude magically acquired it in Slack and you magically acquired it in Slack and you magically acquired it in Slack and you didn't have to think about it? We have didn't have to think about it? We have didn't have to think about it? We have taught companies for decades that data taught companies for decades that data taught companies for decades that data is alpha. Data is something you have an is alpha. Data is something you have an is alpha. Data is something you have an edge with if you're serious. If data is

edge with if you're serious. If data is edge with if you're serious. If data is alpha, what do we think about giving all alpha, what do we think about giving all alpha, what do we think about giving all of that data to a frontier model of that data to a frontier model of that data to a frontier model provider as context? Even if they don't provider as context? Even if they don't provider as context? Even if they don't release it into training data, even if release it into training data, even if release it into training data, even if they if the privacy policy is really they if the privacy policy is really they if the privacy policy is really good and they're behaving really good and they're behaving really good and they're behaving really ethically, which I have no reason to ethically, which I have no reason to ethically, which I have no reason to think they're not, you still are think they're not, you still are think they're not, you still are effectively renting your own context effectively renting your own context effectively renting your own context back to yourself because Claude is going back to yourself because Claude is going back to yourself because Claude is going to be in your slack as a team level to be in your slack as a team level to be in your slack as a team level harness and is going to be incredibly harness and is going to be incredibly harness and is going to be incredibly close to all the work your team does and close to all the work your team does and close to all the work your team does and it's going to be impossible to rip out. it's going to be impossible to rip out. it's going to be impossible to rip out. No matter how cheap the GLM 5.2 class No matter how cheap the GLM 5.2 class No matter how cheap the GLM 5.2 class models are, models are, models are, how can you rip out the model that's how can you rip out the model that's how can you rip out the model that's that close to context? And I think that that close to context? And I think that that close to context? And I think that the GLM 5.2 team knows this. That's why the GLM 5.2 team knows this. That's why the GLM 5.2 team knows this. That's why they released a harness, a Codex-like they released a harness, a Codex-like they released a harness, a Codex-like interface with their AI. It's a first interface with their AI. It's a first interface with their AI. It's a first stab at it. But we got to get much stab at it. But we got to get much stab at it. But we got to get much farther there in tech, where the farther there in tech, where the farther there in tech, where the companies that know they need harnesses companies that know they need harnesses companies that know they need harnesses generally cannot afford to hire the AI generally cannot afford to hire the AI generally cannot afford to hire the AI talent to build those harnesses unless talent to build those harnesses unless talent to build those harnesses unless they're extraordinary companies because they're extraordinary companies because they're extraordinary companies because that AI talent is so in demand right now that AI talent is so in demand right now that AI talent is so in demand right now that it can charge anything it wants and that it can charge anything it wants and that it can charge anything it wants and it usually goes to one of the it usually goes to one of the it usually goes to one of the hyperscalers or another large company.

hyperscalers or another large company. hyperscalers or another large company. And so we're in the dynamic where the And so we're in the dynamic where the And so we're in the dynamic where the only companies that can build their own only companies that can build their own only companies that can build their own last-mile harnesses, their own auto last-mile harnesses, their own auto last-mile harnesses, their own auto routers, are companies that can afford routers, are companies that can afford routers, are companies that can afford that, that can afford the AI talent to that, that can afford the AI talent to that, that can afford the AI talent to do that, which is very scarce. And so if do that, which is very scarce. And so if do that, which is very scarce. And so if you actually think through this dynamic you actually think through this dynamic you actually think through this dynamic with GLM 5.2 and how it's possible but with GLM 5.2 and how it's possible but with GLM 5.2 and how it's possible but at the same time we can have an at the same time we can have an at the same time we can have an incredible open-source model that we're incredible open-source model that we're incredible open-source model that we're excited about and also that Anthropic excited about and also that Anthropic excited about and also that Anthropic still has pricing power to charge a lot still has pricing power to charge a lot still has pricing power to charge a lot for their tokens even though their for their tokens even though their for their tokens even though their tokens are just marginally better, it's tokens are just marginally better, it's tokens are just marginally better, it's actually not a story of intelligence. actually not a story of intelligence. actually not a story of intelligence. It's a story of the last-mile in AI and It's a story of the last-mile in AI and It's a story of the last-mile in AI and the fact that the talent to build the the fact that the talent to build the the fact that the talent to build the last-mile in AI is incredibly scarce. last-mile in AI is incredibly scarce. last-mile in AI is incredibly scarce. Which should, honestly, for a lot of you Which should, honestly, for a lot of you Which should, honestly, for a lot of you watching, be a source for optimism. If watching, be a source for optimism. If watching, be a source for optimism. If we have that scarce a talent, where we have that scarce a talent, where we have that scarce a talent, where people are ending up locked into people are ending up locked into people are ending up locked into contracts with a frontier model provider contracts with a frontier model provider contracts with a frontier model provider because they don't know how to build a because they don't know how to build a because they don't know how to build a harness for themselves, wow is there a harness for themselves, wow is there a harness for themselves, wow is there a lot of opportunity in knowing how to lot of opportunity in knowing how to lot of opportunity in knowing how to build an AI. Like it's an incredible build an AI. Like it's an incredible build an AI. Like it's an incredible opportunity right now. It is not easy to opportunity right now. It is not easy to opportunity right now. It is not easy to do this work. It's not easy to know this do this work. It's not easy to know this do this work. It's not easy to know this is how you handle a tool call in GLM 5.2 is how you handle a tool call in GLM 5.2 is how you handle a tool call in GLM 5.2 and how you should do it differently and how you should do it differently and how you should do it differently from Claude. So does figuring out how from Claude. So does figuring out how from Claude. So does figuring out how memory will work for that system. So memory will work for that system. So memory will work for that system. So does figuring out how the system prompt does figuring out how the system prompt does figuring out how the system prompt needs to change because it's a center of needs to change because it's a center of needs to change because it's a center of distribution model. It's a lot of distribution model. It's a lot of distribution model. It's a lot of technical work. And if you know how to technical work. And if you know how to technical work. And if you know how to do that work or know how to do parts of do that work or know how to do parts of do that work or know how to do parts of that work to essentially refactor that work to essentially refactor that work to essentially refactor agentic pipelines so they work with an agentic pipelines so they work with an agentic pipelines so they work with an open-source model, you are going to be open-source model, you are going to be open-source model, you are going to be incredibly in demand. Especially if you incredibly in demand. Especially if you incredibly in demand. Especially if you compare that with the ability to route

compare that with the ability to route compare that with the ability to route tasks where you can take a task and tasks where you can take a task and tasks where you can take a task and recognize on the fly that it's a recognize on the fly that it's a recognize on the fly that it's a frontier model task and it should go to frontier model task and it should go to frontier model task and it should go to a frontier model versus everything else a frontier model versus everything else a frontier model versus everything else going to a cheaper open-source model. going to a cheaper open-source model. going to a cheaper open-source model. That is going to be a huge investment That is going to be a huge investment That is going to be a huge investment theme for companies in 2026, 2027 and theme for companies in 2026, 2027 and theme for companies in 2026, 2027 and they're going to keep innovating. Claude they're going to keep innovating. Claude they're going to keep innovating. Claude tag is a fantastic example of how of how tag is a fantastic example of how of how tag is a fantastic example of how of how incentives in frontier close-source incentives in frontier close-source incentives in frontier close-source models are giving us incredible models are giving us incredible models are giving us incredible experiences. If you have pricing power, experiences. If you have pricing power, experiences. If you have pricing power, you are heavily incentivized to make you are heavily incentivized to make you are heavily incentivized to make sure that your experience is as sure that your experience is as sure that your experience is as convenient and ergonomic as possible. convenient and ergonomic as possible. convenient and ergonomic as possible. And so features like Claude Claude tag And so features like Claude Claude tag And so features like Claude Claude tag are going to appear really, really fast, are going to appear really, really fast, are going to appear really, really fast, really rapidly, really completely from really rapidly, really completely from really rapidly, really completely from teams at Anthropic, also from OpenAI teams at Anthropic, also from OpenAI teams at Anthropic, also from OpenAI because they're incentivized to keep because they're incentivized to keep because they're incentivized to keep those those prices high and to go after those those prices high and to go after those those prices high and to go after that business. And with open-source that business. And with open-source that business. And with open-source models, you don't have the same margin models, you don't have the same margin models, you don't have the same margin to work with, you don't have the same to work with, you don't have the same to work with, you don't have the same cash flow to work with and you don't cash flow to work with and you don't cash flow to work with and you don't have the same incentive to dig in and have the same incentive to dig in and have the same incentive to dig in and deploy thousands of forward-deployed deploy thousands of forward-deployed deploy thousands of forward-deployed engineers and really make these engineers and really make these engineers and really make these harnesses sing. And so one of the really harnesses sing. And so one of the really harnesses sing. And so one of the really interesting facts that we come to after interesting facts that we come to after interesting facts that we come to after all of this can simultaneously be an all of this can simultaneously be an all of this can simultaneously be an incredible model, a model that a lot of incredible model, a model that a lot of incredible model, a model that a lot of entrepreneurs switch to when the ROI is entrepreneurs switch to when the ROI is entrepreneurs switch to when the ROI is clear and they're technically savvy clear and they're technically savvy clear and they're technically savvy enough to do it, and also not a model enough to do it, and also not a model enough to do it, and also not a model that is easy for a given company that that is easy for a given company that that is easy for a given company that you turn up in phone book to actually you turn up in phone book to actually you turn up in phone book to actually use. It any given company is going to use. It any given company is going to use. It any given company is going to have to think about how they use GLM 5.2 have to think about how they use GLM 5.2 have to think about how they use GLM 5.2 to use it usefully, and they're going to to use it usefully, and they're going to to use it usefully, and they're going to have to think a lot less to sign up for have to think a lot less to sign up for have to think a lot less to sign up for a frontier model contract that's going a frontier model contract that's going a frontier model contract that's going to fit right into their existing

to fit right into their existing to fit right into their existing workflows. That last mile is literally a workflows. That last mile is literally a workflows. That last mile is literally a trillion-dollar last mile in AI. And one trillion-dollar last mile in AI. And one trillion-dollar last mile in AI. And one of the biggest open questions right now of the biggest open questions right now of the biggest open questions right now is whether we will scale our talent fast is whether we will scale our talent fast is whether we will scale our talent fast enough to enable businesses to tackle enough to enable businesses to tackle enough to enable businesses to tackle that problem set without paying so much that problem set without paying so much that problem set without paying so much that they can't afford it. I don't know that they can't afford it. I don't know that they can't afford it. I don't know what the answer's going to be, but what the answer's going to be, but what the answer's going to be, but that's a question we're going to have an that's a question we're going to have an that's a question we're going to have an answer to. We will all collectively answer to. We will all collectively answer to. We will all collectively answer together in the next 3 to 6 answer together in the next 3 to 6 answer together in the next 3 to 6 months. We are going to find out, months. We are going to find out, months. We are going to find out, especially as the US government has this especially as the US government has this especially as the US government has this effective pause in place on frontier effective pause in place on frontier effective pause in place on frontier model releases, model releases, model releases, and the open-source systems are going to and the open-source systems are going to and the open-source systems are going to continue to be available, we're going to continue to be available, we're going to continue to be available, we're going to find out whether companies can adjust to find out whether companies can adjust to find out whether companies can adjust to the fact that intelligence is 98% the fact that intelligence is 98% the fact that intelligence is 98% cheaper and takes a last mile to build. cheaper and takes a last mile to build. cheaper and takes a last mile to build. Can they actually build that last mile? Can they actually build that last mile? Can they actually build that last mile? Can they find teams to build that last Can they find teams to build that last Can they find teams to build that last mile? If you are in an agency or in a mile? If you are in an agency or in a mile? If you are in an agency or in a consulting space, this is a golden goose consulting space, this is a golden goose consulting space, this is a golden goose moment. Like you have a chance here. You moment. Like you have a chance here. You moment. Like you have a chance here. You can really go to town and basically can really go to town and basically can really go to town and basically promise to save people a ton of money on promise to save people a ton of money on promise to save people a ton of money on tokens as part of your ROI proposition, tokens as part of your ROI proposition, tokens as part of your ROI proposition, as long as you can deliver that refactor as long as you can deliver that refactor as long as you can deliver that refactor in a way that maintains quality, which in a way that maintains quality, which in a way that maintains quality, which is not a trivial task. If it was easy, is not a trivial task. If it was easy, is not a trivial task. If it was easy, we wouldn't be having this video. So, we wouldn't be having this video. So, we wouldn't be having this video. So, where does this leave us? GLM 5.2 is an where does this leave us? GLM 5.2 is an where does this leave us? GLM 5.2 is an incredible model. It is important not to incredible model. It is important not to incredible model. It is important not to shame a model or diss a model because shame a model or diss a model because shame a model or diss a model because it's good at center of distribution it's good at center of distribution it's good at center of distribution task, because by definition that is most task, because by definition that is most task, because by definition that is most of our work. Collectively as a species, of our work. Collectively as a species, of our work. Collectively as a species, most of our knowledge work is center of most of our knowledge work is center of most of our knowledge work is center of distribution, just by definition. And if

distribution, just by definition. And if distribution, just by definition. And if that's the case, a model that's really that's the case, a model that's really that's the case, a model that's really good at that is worth taking really good at that is worth taking really good at that is worth taking really seriously. And if we take it seriously, seriously. And if we take it seriously, seriously. And if we take it seriously, that means we have to take the last mile that means we have to take the last mile that means we have to take the last mile seriously. We have to take the idea that seriously. We have to take the idea that seriously. We have to take the idea that we need a harness for that last mile we need a harness for that last mile we need a harness for that last mile seriously. And that's a lot of what I seriously. And that's a lot of what I seriously. And that's a lot of what I have been doing in public is starting to have been doing in public is starting to have been doing in public is starting to articulate what it takes to build a articulate what it takes to build a articulate what it takes to build a harness, whether it's open skills or harness, whether it's open skills or harness, whether it's open skills or open brain or open engine, which I've open brain or open engine, which I've open brain or open engine, which I've all talked about on this channel. How do all talked about on this channel. How do all talked about on this channel. How do you start to take these pieces and put you start to take these pieces and put you start to take these pieces and put them together in a way that is agent them together in a way that is agent them together in a way that is agent agnostic, that is model agnostic, so you agnostic, that is model agnostic, so you agnostic, that is model agnostic, so you can start to install those pieces and can start to install those pieces and can start to install those pieces and actually take advantage of all the actually take advantage of all the actually take advantage of all the intelligence on tap. Whether it's intelligence on tap. Whether it's intelligence on tap. Whether it's Claude, whether it's it's Codex, whether Claude, whether it's it's Codex, whether Claude, whether it's it's Codex, whether it's Hermes, whether it's whatever it's Hermes, whether it's whatever it's Hermes, whether it's whatever whatever system you want, whether it's whatever system you want, whether it's whatever system you want, whether it's your own iPhone 2, you should be able to your own iPhone 2, you should be able to your own iPhone 2, you should be able to easily build to that last mile. And and easily build to that last mile. And and easily build to that last mile. And and I know that there's a lot of custom work I know that there's a lot of custom work I know that there's a lot of custom work for individual companies, and that's why for individual companies, and that's why for individual companies, and that's why I keep saying this is a time for I keep saying this is a time for I keep saying this is a time for builders. But if we don't start down builders. But if we don't start down builders. But if we don't start down that path, we're essentially going to be that path, we're essentially going to be that path, we're essentially going to be renting our company brain and company renting our company brain and company renting our company brain and company context back from the frontier model context back from the frontier model context back from the frontier model providers.

providers. providers. And they're going to have it. And And they're going to have it. And And they're going to have it. And they're going to be able to use it to they're going to be able to use it to they're going to be able to use it to continue to improve their systems and continue to improve their systems and continue to improve their systems and make them more useful, and they'll be make them more useful, and they'll be make them more useful, and they'll be incredibly convenient, incredibly sticky incredibly convenient, incredibly sticky incredibly convenient, incredibly sticky products. And what are we going to do? products. And what are we going to do? products. And what are we going to do? We're going to have to use them. So, We're going to have to use them. So, We're going to have to use them. So, this is a very pivotal moment for this is a very pivotal moment for this is a very pivotal moment for corporations. The firm has never faced a corporations. The firm has never faced a corporations. The firm has never faced a moment where the firm's brain has been moment where the firm's brain has been moment where the firm's brain has been on rent. And that is what we're on the on rent. And that is what we're on the on rent. And that is what we're on the verge of with tools like Claude Tag, verge of with tools like Claude Tag, verge of with tools like Claude Tag, which are incredibly useful. I'm not which are incredibly useful. I'm not which are incredibly useful. I'm not saying they're not useful, they're very saying they're not useful, they're very saying they're not useful, they're very useful. That's exactly the dangerous useful. That's exactly the dangerous useful. That's exactly the dangerous thing. So, I would encourage you thing. So, I would encourage you thing. So, I would encourage you if you are even if it's a tiny company, if you are even if it's a tiny company, if you are even if it's a tiny company, let's say you're building your own let's say you're building your own let's say you're building your own agency, you're an individual agency, you're an individual agency, you're an individual entrepreneur, think seriously, just as entrepreneur, think seriously, just as entrepreneur, think seriously, just as you would if you're a larger company you would if you're a larger company you would if you're a larger company leader, think seriously about whether leader, think seriously about whether leader, think seriously about whether you want to rent that context and you want to rent that context and you want to rent that context and intelligence or not. Think seriously intelligence or not. Think seriously intelligence or not. Think seriously about where you want to go with your about where you want to go with your about where you want to go with your context long term. Ask yourself, do you context long term. Ask yourself, do you context long term. Ask yourself, do you have an idea of the distribution of your have an idea of the distribution of your have an idea of the distribution of your tasks? Do you have access to technical tasks? Do you have access to technical tasks? Do you have access to technical talent that you can use to build out talent that you can use to build out talent that you can use to build out that last mile? What are the task sets that last mile? What are the task sets that last mile? What are the task sets that you would want to assign that would that you would want to assign that would that you would want to assign that would save you a ton in tokens? A lot of save you a ton in tokens? A lot of save you a ton in tokens? A lot of people don't sit down and get pencil and people don't sit down and get pencil and people don't sit down and get pencil and paper and actually ask themselves those paper and actually ask themselves those paper and actually ask themselves those kinds of questions. And I have a whole kinds of questions. And I have a whole kinds of questions. And I have a whole sort of question set that's in more sort of question set that's in more sort of question set that's in more detail that I've been going over with detail that I've been going over with detail that I've been going over with leaders. I put that on the Substack. Uh leaders. I put that on the Substack. Uh leaders. I put that on the Substack. Uh but this is a really serious thing. This but this is a really serious thing. This but this is a really serious thing. This is a moment for open source. GLON 5.2 is a moment for open source. GLON 5.2 is a moment for open source. GLON 5.2 opened that door for all of us, and it's opened that door for all of us, and it's opened that door for all of us, and it's going to be up to us to see how we take going to be up to us to see how we take going to be up to us to see how we take advantage of it. Good luck with that.

advantage of it. Good luck with that. advantage of it. Good luck with that. Cheers. Bye.