AI Hurts Open Source Software

The world runs on open-source software. The world runs on open-source software. If it ever went away, software would If it ever went away, software would If it ever went away, software would collapse. We know about things like Git collapse. We know about things like Git collapse. We know about things like Git and Docker and React, which are fully and Docker and React, which are fully and Docker and React, which are fully open- source projects. But have you ever open- source projects. But have you ever open- source projects. But have you ever thought of the fact that almost all web thought of the fact that almost all web thought of the fact that almost all web hosting, including practically hosting, including practically hosting, including practically everything in the cloud, runs on Linux, everything in the cloud, runs on Linux, everything in the cloud, runs on Linux, which is open- source? Most websites use which is open- source? Most websites use which is open- source? Most websites use npm packages or Nougat packages that are npm packages or Nougat packages that are npm packages or Nougat packages that are open-source packages that support that open-source packages that support that open-source packages that support that site. So we also have Android and site. So we also have Android and site. So we also have Android and Firefox and Postgress and Chrome and Firefox and Postgress and Chrome and Firefox and Postgress and Chrome and lots more that are open- source projects lots more that are open- source projects lots more that are open- source projects or based upon open-source projects. For or based upon open-source projects. For or based upon open-source projects. For decades, the world has depended on decades, the world has depended on decades, the world has depended on open-source to power practically open-source to power practically open-source to power practically everything that we rely on. But now AI everything that we rely on. But now AI everything that we rely on. But now AI is threatening open source. Let's look is threatening open source. Let's look is threatening open source. Let's look at how in this video. Now, every action at how in this video. Now, every action at how in this video. Now, every action you take has a cost. And too often we you take has a cost. And too often we you take has a cost. And too often we look at the benefits of an action and look at the benefits of an action and look at the benefits of an action and forget to look at the drawbacks. In this forget to look at the drawbacks. In this forget to look at the drawbacks. In this series, we're looking at the costs series, we're looking at the costs series, we're looking at the costs associated with AI in various sectors in associated with AI in various sectors in associated with AI in various sectors in order to have a better understanding of order to have a better understanding of order to have a better understanding of what we're giving up in order to gain what we're giving up in order to gain what we're giving up in order to gain the benefits of AI. This isn't about the benefits of AI. This isn't about the benefits of AI. This isn't about hurting AI or hating AI or wanting AI go hurting AI or hating AI or wanting AI go hurting AI or hating AI or wanting AI go away, but it's about making sure we look away, but it's about making sure we look away, but it's about making sure we look at the pros and cons of our decisions.

at the pros and cons of our decisions. at the pros and cons of our decisions. Now, if you want to support this channel Now, if you want to support this channel Now, if you want to support this channel and the free content that I produce, and the free content that I produce, and the free content that I produce, consider purchasing one of my courses at consider purchasing one of my courses at consider purchasing one of my courses at imtimcorey.com. imtimcorey.com. imtimcorey.com. I have master courses on C, web I have master courses on C, web I have master courses on C, web development, and Unity, as well as development, and Unity, as well as development, and Unity, as well as training courses that cover a wide training courses that cover a wide training courses that cover a wide variety of topics. The income from those variety of topics. The income from those variety of topics. The income from those sales directly funds the free content sales directly funds the free content sales directly funds the free content that I create here. So, let's look at that I create here. So, let's look at that I create here. So, let's look at how AI can hurt open source. And we're how AI can hurt open source. And we're how AI can hurt open source. And we're going to start with this article here on going to start with this article here on going to start with this article here on how AI wiped out 80% of Tailwind's how AI wiped out 80% of Tailwind's how AI wiped out 80% of Tailwind's revenue. Now, to be clear, when we talk revenue. Now, to be clear, when we talk revenue. Now, to be clear, when we talk about this particular situation, that about this particular situation, that about this particular situation, that doesn't mean we're just talking about doesn't mean we're just talking about doesn't mean we're just talking about this situation. There's a representation this situation. There's a representation this situation. There's a representation here of how some open-source here of how some open-source here of how some open-source uh systems are maintained and paid for. uh systems are maintained and paid for. uh systems are maintained and paid for. I want to be clear here. Every I want to be clear here. Every I want to be clear here. Every opensource project that you use opensource project that you use opensource project that you use is paid for by someone. They might say, is paid for by someone. They might say, is paid for by someone. They might say, Tim, no, it's not. No, it's not. Some Tim, no, it's not. No, it's not. Some Tim, no, it's not. No, it's not. Some are free. That yes, when you use it, are free. That yes, when you use it, are free. That yes, when you use it, you're using the free. You're using the you're using the free. You're using the you're using the free. You're using the the value that comes out of it. But the value that comes out of it. But the value that comes out of it. But someone paid for it. Someone did. Now, someone paid for it. Someone did. Now, someone paid for it. Someone did. Now, it may be that it's a it's a pet project it may be that it's a it's a pet project it may be that it's a it's a pet project to someone. as their hobby product. They to someone. as their hobby product. They to someone. as their hobby product. They do their free time. Guess what? They do their free time. Guess what? They do their free time. Guess what? They gave their free time in order to give gave their free time in order to give gave their free time in order to give you something.

you something. you something. Or maybe they did it at work where they Or maybe they did it at work where they Or maybe they did it at work where they built something and then said their built something and then said their built something and then said their their employer said, "Yes, you can give their employer said, "Yes, you can give their employer said, "Yes, you can give that away for free." So, their company that away for free." So, their company that away for free." So, their company paid for it. Or maybe it's a mixture of paid for it. Or maybe it's a mixture of paid for it. Or maybe it's a mixture of both. or maybe open- source both. or maybe open- source both. or maybe open- source contributions helped fund it where contributions helped fund it where contributions helped fund it where people sponsored that project or they people sponsored that project or they people sponsored that project or they said, "Hey, I I'll give you a coffee." said, "Hey, I I'll give you a coffee." said, "Hey, I I'll give you a coffee." By the way, those links almost never By the way, those links almost never By the way, those links almost never work. People almost never use those work. People almost never use those work. People almost never use those links. Not saying no one does, but just links. Not saying no one does, but just links. Not saying no one does, but just know when you see that link, use it if know when you see that link, use it if know when you see that link, use it if you can help it. If you can afford it, you can help it. If you can afford it, you can help it. If you can afford it, use it because very few people do. So, use it because very few people do. So, use it because very few people do. So, someone paid for everything that you're someone paid for everything that you're someone paid for everything that you're seeing. So, how does Tailwind do it? seeing. So, how does Tailwind do it? seeing. So, how does Tailwind do it? Well, let's look. So, first off, just to Well, let's look. So, first off, just to Well, let's look. So, first off, just to set a level here, um, Tailwind is a set a level here, um, Tailwind is a set a level here, um, Tailwind is a default of many framework and tools like default of many framework and tools like default of many framework and tools like Ruby on Rails, Phoenix, and Nex.js. If Ruby on Rails, Phoenix, and Nex.js. If Ruby on Rails, Phoenix, and Nex.js. If you're a net web developer, you may have you're a net web developer, you may have you're a net web developer, you may have looked at should I use Bootstrap, which looked at should I use Bootstrap, which looked at should I use Bootstrap, which is also an open- source project, or is also an open- source project, or is also an open- source project, or should I use Tailwind. So this is a very should I use Tailwind. So this is a very should I use Tailwind. So this is a very common, very popular project that really common, very popular project that really common, very popular project that really saves people a lot of time. Working with saves people a lot of time. Working with saves people a lot of time. Working with CSS can be difficult. It can be tricky CSS can be difficult. It can be tricky CSS can be difficult. It can be tricky and getting a good foundation in your and getting a good foundation in your and getting a good foundation in your CSS is really important. And so these CSS is really important. And so these CSS is really important. And so these projects are really helpful in getting projects are really helpful in getting projects are really helpful in getting you a leg up and getting you started in you a leg up and getting you started in you a leg up and getting you started in a project, getting you going and doing a a project, getting you going and doing a a project, getting you going and doing a lot of the heavy lifting for your

lot of the heavy lifting for your lot of the heavy lifting for your layouts, your style, your design, and layouts, your style, your design, and layouts, your style, your design, and then you can just tweak from there. So, then you can just tweak from there. So, then you can just tweak from there. So, it really does save thousands, tens of it really does save thousands, tens of it really does save thousands, tens of thousands, hundreds of thousands of thousands, hundreds of thousands of thousands, hundreds of thousands of people. It saved lots of time because people. It saved lots of time because people. It saved lots of time because they can use Tailwind as the starting they can use Tailwind as the starting they can use Tailwind as the starting point. But what you may not know is that point. But what you may not know is that point. But what you may not know is that Tailwind has a paid service called Tailwind has a paid service called Tailwind has a paid service called Tailwind Plus that offers premium Tailwind Plus that offers premium Tailwind Plus that offers premium components and templates. So, they have components and templates. So, they have components and templates. So, they have a paid version of Tailwind, an upgrade a paid version of Tailwind, an upgrade a paid version of Tailwind, an upgrade of Tailwind that helps fund their free of Tailwind that helps fund their free of Tailwind that helps fund their free content. If that sounds familiar, I content. If that sounds familiar, I content. If that sounds familiar, I started this video talking about how if started this video talking about how if started this video talking about how if you want to fund my content and you are you want to fund my content and you are you want to fund my content and you are able and you want to help, then you can able and you want to help, then you can able and you want to help, then you can go to imtcorey.com and purchase a go to imtcorey.com and purchase a go to imtcorey.com and purchase a course. That'll benefit you, but it will course. That'll benefit you, but it will course. That'll benefit you, but it will also benefit everyone else. That's the also benefit everyone else. That's the also benefit everyone else. That's the same thing that Tailwind is doing. My same thing that Tailwind is doing. My same thing that Tailwind is doing. My business model is their business model business model is their business model business model is their business model practically. So this is how they fund practically. So this is how they fund practically. So this is how they fund their free content. Now what does that their free content. Now what does that their free content. Now what does that do? Let's let's take a minute to think do? Let's let's take a minute to think do? Let's let's take a minute to think about this.

about this. about this. They are not getting in your face. They They are not getting in your face. They They are not getting in your face. They are not saying, "Hey, you can't use are not saying, "Hey, you can't use are not saying, "Hey, you can't use Tailwind unless you unless you pay for Tailwind unless you unless you pay for Tailwind unless you unless you pay for Tailwind Plus." Tailwind Plus." Tailwind Plus." This is how they fund that free content. This is how they fund that free content. This is how they fund that free content. This is how they they even let people This is how they they even let people This is how they they even let people know that t that tailwind plus even know that t that tailwind plus even know that t that tailwind plus even exists. exists. exists. The flow is that a user says hey preler The flow is that a user says hey preler The flow is that a user says hey preler says hey I need to know how to do this says hey I need to know how to do this says hey I need to know how to do this in tailwind. So I google that and Google in tailwind. So I google that and Google in tailwind. So I google that and Google says here's a link to the documentation says here's a link to the documentation says here's a link to the documentation where it tells you how to do that. Super where it tells you how to do that. Super where it tells you how to do that. Super helpful. So, Google gets to share share helpful. So, Google gets to share share helpful. So, Google gets to share share the the link and it also gets to share the the link and it also gets to share the the link and it also gets to share their their u paid links as well around their their u paid links as well around their their u paid links as well around that to fund that um that search engine that to fund that um that search engine that to fund that um that search engine traffic. traffic. traffic. Then the user goes to Tailwind site and Then the user goes to Tailwind site and Then the user goes to Tailwind site and they see that Tailwind plus exists. they see that Tailwind plus exists. they see that Tailwind plus exists. They're in the context of I have to do I They're in the context of I have to do I They're in the context of I have to do I have to know how to do something in have to know how to do something in have to know how to do something in Tailwind and they can see that Tailwind Tailwind and they can see that Tailwind Tailwind and they can see that Tailwind offers Tailwind plus and that can give offers Tailwind plus and that can give offers Tailwind plus and that can give them additional components and and them additional components and and them additional components and and templates. And so it's an ad, yes, but templates. And so it's an ad, yes, but templates. And so it's an ad, yes, but it's an ad in a place where you're it's an ad in a place where you're it's an ad in a place where you're already looking for things like this. So already looking for things like this. So already looking for things like this. So it it could be a solution to your it it could be a solution to your it it could be a solution to your problem. And so this is how they don't problem. And so this is how they don't problem. And so this is how they don't get in your face. This is how they don't get in your face. This is how they don't get in your face. This is how they don't overwhelm you with ads everywhere. Um I overwhelm you with ads everywhere. Um I overwhelm you with ads everywhere. Um I think we're all tired of ads, right? I

think we're all tired of ads, right? I think we're all tired of ads, right? I think we're all tired of just having think we're all tired of just having think we're all tired of just having stuff shoved in our faces. stuff shoved in our faces. stuff shoved in our faces. And so the idea that a company says, And so the idea that a company says, And so the idea that a company says, "Hey, you know what? We're not going to "Hey, you know what? We're not going to "Hey, you know what? We're not going to demand that every person who uses demand that every person who uses demand that every person who uses Tailwind give us a little bit. We're not Tailwind give us a little bit. We're not Tailwind give us a little bit. We're not going to put things behind a payw wall. going to put things behind a payw wall. going to put things behind a payw wall. What we're going to do is we're going to What we're going to do is we're going to What we're going to do is we're going to let you have all of Tailwind, but then let you have all of Tailwind, but then let you have all of Tailwind, but then some additional things that yes, we'll some additional things that yes, we'll some additional things that yes, we'll put behind a payw wall, but that will put behind a payw wall, but that will put behind a payw wall, but that will fund our free content fund our free content fund our free content and we're only going to basically tell and we're only going to basically tell and we're only going to basically tell you about it when you go to the you about it when you go to the you about it when you go to the documentation. So, that was their idea. documentation. So, that was their idea. documentation. So, that was their idea. That was their way of supporting the That was their way of supporting the That was their way of supporting the business and supporting their employees business and supporting their employees business and supporting their employees because yes, they have paid employees. because yes, they have paid employees. because yes, they have paid employees. Someone is paying for Tailwind and that Someone is paying for Tailwind and that Someone is paying for Tailwind and that someone is the people who are paying for someone is the people who are paying for someone is the people who are paying for Tailwind Plus. Now look how the process Tailwind Plus. Now look how the process Tailwind Plus. Now look how the process changed with the advent of LLMs. The changed with the advent of LLMs. The changed with the advent of LLMs. The user says, "Hey, I have a problem." And user says, "Hey, I have a problem." And user says, "Hey, I have a problem." And asks the LLM, not Google. The LM says, asks the LLM, not Google. The LM says, asks the LLM, not Google. The LM says, "Here's a code snippet." And by the way, "Here's a code snippet." And by the way, "Here's a code snippet." And by the way, if they go to Google now and ask that if they go to Google now and ask that if they go to Google now and ask that question, maybe they're saying, "I'm question, maybe they're saying, "I'm question, maybe they're saying, "I'm just going to Google it like I used to."

just going to Google it like I used to." just going to Google it like I used to." But Google gives them the answer instead But Google gives them the answer instead But Google gives them the answer instead of giving them the link to the of giving them the link to the of giving them the link to the documentation first. So link link to the documentation first. So link link to the documentation first. So link link to the documentation may be there but it's documentation may be there but it's documentation may be there but it's buried down below the answer. buried down below the answer. buried down below the answer. So now the way that they used to fund So now the way that they used to fund So now the way that they used to fund Tailwind is that you go to documentation Tailwind is that you go to documentation Tailwind is that you go to documentation and you see that hey I could pay for and you see that hey I could pay for and you see that hey I could pay for something and get something that I might something and get something that I might something and get something that I might need. need. need. But now the LM's come along and say, you But now the LM's come along and say, you But now the LM's come along and say, you know what, we're going to take all your know what, we're going to take all your know what, we're going to take all your documentation. We're going to ingest all documentation. We're going to ingest all documentation. We're going to ingest all that information as well as from other that information as well as from other that information as well as from other sources. But we're going to take all sources. But we're going to take all sources. But we're going to take all that documentation and we're going to that documentation and we're going to that documentation and we're going to give it to the user, not you. We are. give it to the user, not you. We are. give it to the user, not you. We are. And so because we're doing that, then we And so because we're doing that, then we And so because we're doing that, then we can give the user the information can give the user the information can give the user the information directly. It's better, right, to have it directly. It's better, right, to have it directly. It's better, right, to have it directly answered. But what it does is directly answered. But what it does is directly answered. But what it does is they cut out Tailwind and all of their they cut out Tailwind and all of their they cut out Tailwind and all of their web traffic, at least all of their human web traffic, at least all of their human web traffic, at least all of their human web traffic. And as a result, no one web traffic. And as a result, no one web traffic. And as a result, no one sees the idea of upgrading to Google+.

sees the idea of upgrading to Google+. sees the idea of upgrading to Google+. And as a side note here, yes, this is a And as a side note here, yes, this is a And as a side note here, yes, this is a side note, but reading the documentation side note, but reading the documentation side note, but reading the documentation is not just about finding the answer to is not just about finding the answer to is not just about finding the answer to your specific question you have right your specific question you have right your specific question you have right now. now. now. One of the benefits of reading the One of the benefits of reading the One of the benefits of reading the documentation is that when you do then documentation is that when you do then documentation is that when you do then you will also be able to see other you will also be able to see other you will also be able to see other things around the documentation other things around the documentation other things around the documentation other things that that might be useful that things that that might be useful that things that that might be useful that you didn't realize to even ask for you didn't realize to even ask for you didn't realize to even ask for including Tailwind plus in this case but including Tailwind plus in this case but including Tailwind plus in this case but maybe other things that it can do you maybe other things that it can do you maybe other things that it can do you didn't even know that it could do. So didn't even know that it could do. So didn't even know that it could do. So reading documentation is valuable. Not reading documentation is valuable. Not reading documentation is valuable. Not knowing to ask for that information is a knowing to ask for that information is a knowing to ask for that information is a problem. You don't know what you don't problem. You don't know what you don't problem. You don't know what you don't know. And LM's kind of hide that from know. And LM's kind of hide that from know. And LM's kind of hide that from you. But okay, so this is what's you. But okay, so this is what's you. But okay, so this is what's happening at Tailwind. How is that happening at Tailwind. How is that happening at Tailwind. How is that impacting them? Well, they've had to lay impacting them? Well, they've had to lay impacting them? Well, they've had to lay off 75% of the engineering team. And off 75% of the engineering team. And off 75% of the engineering team. And why? Because traffic has gone down 40% why? Because traffic has gone down 40% why? Because traffic has gone down 40% since early 2023. And as a result, 80% since early 2023. And as a result, 80% since early 2023. And as a result, 80% of the revenue has gone away from of the revenue has gone away from of the revenue has gone away from Tailwind Plus. Now, you may say, you Tailwind Plus. Now, you may say, you Tailwind Plus. Now, you may say, you know, in this specific situation, know, in this specific situation, know, in this specific situation, Tailwind should do something different.

Tailwind should do something different. Tailwind should do something different. That can be fair, but we're not talking That can be fair, but we're not talking That can be fair, but we're not talking about one specific situation. We're about one specific situation. We're about one specific situation. We're talking about a concept, an idea, a much talking about a concept, an idea, a much talking about a concept, an idea, a much broader picture, and using this as our broader picture, and using this as our broader picture, and using this as our example of that. So, example of that. So, example of that. So, what should Tailwind do? what can they what should Tailwind do? what can they what should Tailwind do? what can they do to support tailwind because again do to support tailwind because again do to support tailwind because again they used this to fund their engineering they used this to fund their engineering they used this to fund their engineering team and 75% of their engineering team team and 75% of their engineering team team and 75% of their engineering team got laid off. Now that directly impacts got laid off. Now that directly impacts got laid off. Now that directly impacts those engineers. However, what does it those engineers. However, what does it those engineers. However, what does it also impact? It also impacts every also impact? It also impacts every also impact? It also impacts every single person who uses Tailwind. Why? single person who uses Tailwind. Why? single person who uses Tailwind. Why? Because 75% of their engineering Because 75% of their engineering Because 75% of their engineering department went away. What do you think department went away. What do you think department went away. What do you think those engineers were doing? they were those engineers were doing? they were those engineers were doing? they were working on Tailwind. So now threearters working on Tailwind. So now threearters working on Tailwind. So now threearters of the people who were dedicated to of the people who were dedicated to of the people who were dedicated to Tailwind are gone which means that Tailwind are gone which means that Tailwind are gone which means that Tailwind does not get features as fast. Tailwind does not get features as fast. Tailwind does not get features as fast. It does not get as many features. It It does not get as many features. It It does not get as many features. It does not get as many bug fixes and no does not get as many bug fixes and no does not get as many bug fixes and no just throwing more AI at it doesn't just throwing more AI at it doesn't just throwing more AI at it doesn't solve the problem. In fact, it's going solve the problem. In fact, it's going solve the problem. In fact, it's going to make it worse. So we all get a worse to make it worse. So we all get a worse to make it worse. So we all get a worse project, a worse Tailwind because of the project, a worse Tailwind because of the project, a worse Tailwind because of the fact that people aren't going to the fact that people aren't going to the fact that people aren't going to the documentation and not seeing Tailwind documentation and not seeing Tailwind documentation and not seeing Tailwind Plus.

Plus. Plus. The opportunities here for Tailwind, The opportunities here for Tailwind, The opportunities here for Tailwind, they have to pivot. But how do they they have to pivot. But how do they they have to pivot. But how do they pivot? Well, they'll probably have to do pivot? Well, they'll probably have to do pivot? Well, they'll probably have to do some kind of payw wall where more of some kind of payw wall where more of some kind of payw wall where more of Tailwind isn't available unless you pay. Tailwind isn't available unless you pay. Tailwind isn't available unless you pay. And that hurts everybody. And that hurts everybody. And that hurts everybody. Okay, so that's just tailwind. Let's Okay, so that's just tailwind. Let's Okay, so that's just tailwind. Let's kind of follow up here, though. Um, I kind of follow up here, though. Um, I kind of follow up here, though. Um, I thought these two lines were important. thought these two lines were important. thought these two lines were important. You and I use open source every day. You and I use open source every day. You and I use open source every day. While free for us, people working on While free for us, people working on While free for us, people working on these projects have to make a living. these projects have to make a living. these projects have to make a living. That I think is really important to That I think is really important to That I think is really important to understand. Everyone pays for open understand. Everyone pays for open understand. Everyone pays for open source or open source is paid for by source or open source is paid for by source or open source is paid for by someone. And these people who were someone. And these people who were someone. And these people who were working at open source can't just always working at open source can't just always working at open source can't just always volunteer their time. It's being paid volunteer their time. It's being paid volunteer their time. It's being paid for by someone and those people have to for by someone and those people have to for by someone and those people have to make a living. Yes, maybe they have jobs make a living. Yes, maybe they have jobs make a living. Yes, maybe they have jobs and that's great. And I've seen open and that's great. And I've seen open and that's great. And I've seen open source projects where a person has a job source projects where a person has a job source projects where a person has a job and and so they just work on it as a you and and so they just work on it as a you and and so they just work on it as a you know a nice to do in their free time but know a nice to do in their free time but know a nice to do in their free time but then when they lose their job or their then when they lose their job or their then when they lose their job or their job changes and they don't have enough job changes and they don't have enough job changes and they don't have enough time in their job the project withers time in their job the project withers time in their job the project withers and dies because and dies because and dies because it was being funded by their free time it was being funded by their free time it was being funded by their free time that's no longer available.

that's no longer available. that's no longer available. So people that are working on this So people that are working on this So people that are working on this project have to make a living and the project have to make a living and the project have to make a living and the rise of LLMs is disrupting the rise of LLMs is disrupting the rise of LLMs is disrupting the traditional way that people interact traditional way that people interact traditional way that people interact with documentation. So in this specific with documentation. So in this specific with documentation. So in this specific case, the way that we funded Tailwind case, the way that we funded Tailwind case, the way that we funded Tailwind was to interact with documentation and was to interact with documentation and was to interact with documentation and that's gone. So some sites the way that that's gone. So some sites the way that that's gone. So some sites the way that you fund the site or fund the thing a you fund the site or fund the thing a you fund the site or fund the thing a site does is by viewing the ads on the site does is by viewing the ads on the site does is by viewing the ads on the site. I guess not, you know, specific to site. I guess not, you know, specific to site. I guess not, you know, specific to that company. They they just have ads on that company. They they just have ads on that company. They they just have ads on the site. Well, if traffic goes away, the site. Well, if traffic goes away, the site. Well, if traffic goes away, that funding does too. And so now that funding does too. And so now that funding does too. And so now therefore, therefore, therefore, that's not being p that's not paying for that's not being p that's not paying for that's not being p that's not paying for the open source anymore. So, however you the open source anymore. So, however you the open source anymore. So, however you monetize open source, most of that is monetize open source, most of that is monetize open source, most of that is going away. And that's a problem because going away. And that's a problem because going away. And that's a problem because guess what happens guess what happens guess what happens if people aren't being paid? If they're if people aren't being paid? If they're if people aren't being paid? If they're not making a living, then the thing not making a living, then the thing not making a living, then the thing they're doing goes away.

they're doing goes away. they're doing goes away. And if if we're not careful, we lose And if if we're not careful, we lose And if if we're not careful, we lose open-source projects. And we already are open-source projects. And we already are open-source projects. And we already are because of how this is playing out. because of how this is playing out. because of how this is playing out. Let's look next at that was the funding. Let's look next at that was the funding. Let's look next at that was the funding. Let's look now at at how AI is just Let's look now at at how AI is just Let's look now at at how AI is just wearing out and damaging the open- wearing out and damaging the open- wearing out and damaging the open- source maintainers. So open source source maintainers. So open source source maintainers. So open source maintainers are drowning in AI generated maintainers are drowning in AI generated maintainers are drowning in AI generated pull requests. Enterprise teams are pull requests. Enterprise teams are pull requests. Enterprise teams are next. So this is I think a good kind of next. So this is I think a good kind of next. So this is I think a good kind of keep your head up here. Think about keep your head up here. Think about keep your head up here. Think about this. This is the canary in the coal this. This is the canary in the coal this. This is the canary in the coal mine. This is this is the um indicator mine. This is this is the um indicator mine. This is this is the um indicator of what's to come. So AI is flooding of what's to come. So AI is flooding of what's to come. So AI is flooding open source with low quality pull open source with low quality pull open source with low quality pull requests. Learn how enterprise teams can requests. Learn how enterprise teams can requests. Learn how enterprise teams can avoid burnout by fixing the code avoid burnout by fixing the code avoid burnout by fixing the code validation bottleneck. So it's saying validation bottleneck. So it's saying validation bottleneck. So it's saying basically we're seeing this already in basically we're seeing this already in basically we're seeing this already in open source and this is going to start open source and this is going to start open source and this is going to start trickling into the enterprise soon. trickling into the enterprise soon. trickling into the enterprise soon. Okay. Um something is breaking at open Okay. Um something is breaking at open Okay. Um something is breaking at open source and it should alarm every source and it should alarm every source and it should alarm every engineering leader pushing coding agents engineering leader pushing coding agents engineering leader pushing coding agents into their organization. Over the past into their organization. Over the past into their organization. Over the past year, open source maintainers have been year, open source maintainers have been year, open source maintainers have been overwhelmed by a flood of lowquality AI overwhelmed by a flood of lowquality AI overwhelmed by a flood of lowquality AI generated pull requests, verbose changes generated pull requests, verbose changes generated pull requests, verbose changes with nonsensical descriptions, with nonsensical descriptions, with nonsensical descriptions, contributions that submitters cannot contributions that submitters cannot contributions that submitters cannot explain when questioned, code that looks explain when questioned, code that looks explain when questioned, code that looks plausible on the surface but crumbles plausible on the surface but crumbles plausible on the surface but crumbles under a review. These are different ways under a review. These are different ways under a review. These are different ways that these pull requests just aren't that these pull requests just aren't that these pull requests just aren't valuable. So the first one, verbose

valuable. So the first one, verbose valuable. So the first one, verbose changes with nonsensical descriptions. changes with nonsensical descriptions. changes with nonsensical descriptions. This is stuff that yeah, you can This is stuff that yeah, you can This is stuff that yeah, you can probably clean out pretty pretty easily. probably clean out pretty pretty easily. probably clean out pretty pretty easily. Look at the description and go, it Look at the description and go, it Look at the description and go, it doesn't look right. Right. The problem doesn't look right. Right. The problem doesn't look right. Right. The problem is that AI writes human speech in a way is that AI writes human speech in a way is that AI writes human speech in a way that looks right. So, it's not just a that looks right. So, it's not just a that looks right. So, it's not just a matter of, you know, using a whole lot matter of, you know, using a whole lot matter of, you know, using a whole lot of consonants that have no vowels, of consonants that have no vowels, of consonants that have no vowels, right? Like, no, it's it's the fact that right? Like, no, it's it's the fact that right? Like, no, it's it's the fact that it looks like a paragraph. It looks like it looks like a paragraph. It looks like it looks like a paragraph. It looks like it might even be right, but when you it might even be right, but when you it might even be right, but when you actually think about it, it's actually think about it, it's actually think about it, it's nonsensical. So it that may be a few nonsensical. So it that may be a few nonsensical. So it that may be a few minutes of code review per issue, but it minutes of code review per issue, but it minutes of code review per issue, but it still takes a few minutes of your time still takes a few minutes of your time still takes a few minutes of your time just to review that and go, "Wait, this just to review that and go, "Wait, this just to review that and go, "Wait, this doesn't make sense." And so if you are doesn't make sense." And so if you are doesn't make sense." And so if you are quote unquote harsh, you would reject quote unquote harsh, you would reject quote unquote harsh, you would reject that, which I think you should. Next, that, which I think you should. Next, that, which I think you should. Next, contributions that submitters cannot contributions that submitters cannot contributions that submitters cannot explain when questioned. And we'll look explain when questioned. And we'll look explain when questioned. And we'll look more at this in a in a future um future more at this in a in a future um future more at this in a in a future um future article, but this is one where someone article, but this is one where someone article, but this is one where someone turned an AI loose on their favorite turned an AI loose on their favorite turned an AI loose on their favorite project or open- source uh you know project or open- source uh you know project or open- source uh you know project and they said, "Hey, I want you project and they said, "Hey, I want you project and they said, "Hey, I want you to add this feature or fix this bug, to add this feature or fix this bug, to add this feature or fix this bug, whatever, and then when they submit the whatever, and then when they submit the whatever, and then when they submit the pull request and the the maintainer pull request and the the maintainer pull request and the the maintainer says, "Hey, I have a question about says, "Hey, I have a question about says, "Hey, I have a question about this." The person who submitted it says, this." The person who submitted it says, this." The person who submitted it says, "I I can't answer that because I didn't "I I can't answer that because I didn't "I I can't answer that because I didn't write the code." So write the code." So write the code." So that's a problem because first of all if that's a problem because first of all if that's a problem because first of all if you can't ask answer the questions well you can't ask answer the questions well you can't ask answer the questions well then who can the AI that's not that's then who can the AI that's not that's then who can the AI that's not that's not how this works. We should be able to

not how this works. We should be able to not how this works. We should be able to have questions answered because you have have questions answered because you have have questions answered because you have to be able to say why did you do it this to be able to say why did you do it this to be able to say why did you do it this way? Why was this chosen versus that? way? Why was this chosen versus that? way? Why was this chosen versus that? And what we don't want and what's And what we don't want and what's And what we don't want and what's happening is that you're just playing a happening is that you're just playing a happening is that you're just playing a telephone game to a person's LLM. Well, telephone game to a person's LLM. Well, telephone game to a person's LLM. Well, at that point, and we'll again talk at that point, and we'll again talk at that point, and we'll again talk about this later, but at that point, why about this later, but at that point, why about this later, but at that point, why shouldn't the maintainer just use the shouldn't the maintainer just use the shouldn't the maintainer just use the LLM? Why have the person in the middle? LLM? Why have the person in the middle? LLM? Why have the person in the middle? So, So, So, contributions where you can't explain contributions where you can't explain contributions where you can't explain what's going on, that's a problem, what's going on, that's a problem, what's going on, that's a problem, should also be rejected. And then code, should also be rejected. And then code, should also be rejected. And then code, this is the most dangerous one. code this is the most dangerous one. code this is the most dangerous one. code that looks plausible on the surface but that looks plausible on the surface but that looks plausible on the surface but crumbles under review. crumbles under review. crumbles under review. So this I think is an order of the least So this I think is an order of the least So this I think is an order of the least amount of time a the maintainer has to amount of time a the maintainer has to amount of time a the maintainer has to spend on the pull request to the most. spend on the pull request to the most. spend on the pull request to the most. This one is the most time where the code This one is the most time where the code This one is the most time where the code looks right and it looks like it might looks right and it looks like it might looks right and it looks like it might be a solution to the problem, but then be a solution to the problem, but then be a solution to the problem, but then the more you kind of poke at it, the the more you kind of poke at it, the the more you kind of poke at it, the more you test it out and and try and more you test it out and and try and more you test it out and and try and figure out, you figure out this isn't figure out, you figure out this isn't figure out, you figure out this isn't going to solve our actual problem or going to solve our actual problem or going to solve our actual problem or this has some weird bugs in it. Uh ones this has some weird bugs in it. Uh ones this has some weird bugs in it. Uh ones that are, you know, maybe not the the that are, you know, maybe not the the that are, you know, maybe not the the primary way it does things, but we primary way it does things, but we primary way it does things, but we probably have to write some new unit probably have to write some new unit probably have to write some new unit tests to kind of catch these bugs tests to kind of catch these bugs tests to kind of catch these bugs because this is, you know, all new because this is, you know, all new because this is, you know, all new territory. So that takes a lot of time territory. So that takes a lot of time territory. So that takes a lot of time and effort and effort and effort and what's happening is it's and what's happening is it's and what's happening is it's overwhelming maintainers. So the overwhelming maintainers. So the overwhelming maintainers. So the Jazzband Collective, a well-known Python Jazzband Collective, a well-known Python Jazzband Collective, a well-known Python project ecosystem, was forced to shut

project ecosystem, was forced to shut project ecosystem, was forced to shut down entirely this year. Its lead down entirely this year. Its lead down entirely this year. Its lead maintainers cited the unsustainable maintainers cited the unsustainable maintainers cited the unsustainable volume of AI generated spam pull volume of AI generated spam pull volume of AI generated spam pull requests and issues as its primary requests and issues as its primary requests and issues as its primary driver. driver. driver. So So So this this collective was shut down this this collective was shut down this this collective was shut down entirely because it just got overwhelmed entirely because it just got overwhelmed entirely because it just got overwhelmed with junk from AI. So we're losing with junk from AI. So we're losing with junk from AI. So we're losing open-source projects, open- source open-source projects, open- source open-source projects, open- source collectives collectives collectives because of AI junk contributions, spam because of AI junk contributions, spam because of AI junk contributions, spam essentially. Think back to the email essentially. Think back to the email essentially. Think back to the email days, early email days where you didn't days, early email days where you didn't days, early email days where you didn't have a good spam filter, but spam was have a good spam filter, but spam was have a good spam filter, but spam was coming and it felt like email is almost coming and it felt like email is almost coming and it felt like email is almost unusable because there was so much junk unusable because there was so much junk unusable because there was so much junk that you can't find the good stuff. Even that you can't find the good stuff. Even that you can't find the good stuff. Even now, if I go to my my I have a Gmail now, if I go to my my I have a Gmail now, if I go to my my I have a Gmail account that's a a personal Gmail account that's a a personal Gmail account that's a a personal Gmail account because I have YouTube TV and so account because I have YouTube TV and so account because I have YouTube TV and so I had a personal Gmail account. But if I I had a personal Gmail account. But if I I had a personal Gmail account. But if I go to the the spam folder in there, go to the the spam folder in there, go to the the spam folder in there, every once in a while there is a every once in a while there is a every once in a while there is a legitimate email, but there are legitimate email, but there are legitimate email, but there are thousands of emails that are just junk.

thousands of emails that are just junk. thousands of emails that are just junk. So finding that one, it's really hard to So finding that one, it's really hard to So finding that one, it's really hard to do. And if I had to maintain that, I had do. And if I had to maintain that, I had do. And if I had to maintain that, I had to go through that, I wouldn't. I just to go through that, I wouldn't. I just to go through that, I wouldn't. I just delete the account. delete the account. delete the account. Okay. other projects are feeling the Okay. other projects are feeling the Okay. other projects are feeling the same pressure. So um the GDO game engine same pressure. So um the GDO game engine same pressure. So um the GDO game engine which you should you if you're into you which you should you if you're into you which you should you if you're into you know developing games for um you know know developing games for um you know know developing games for um you know for the different platforms there's for the different platforms there's for the different platforms there's Unreal Unity and then GDAU. So GDAU Unreal Unity and then GDAU. So GDAU Unreal Unity and then GDAU. So GDAU works with C. It's been you know more works with C. It's been you know more works with C. It's been you know more and more popular recently. Well and more popular recently. Well and more popular recently. Well this uh Remy I can't pronounce that last this uh Remy I can't pronounce that last this uh Remy I can't pronounce that last name. I'm sorry. Remy um has described name. I'm sorry. Remy um has described name. I'm sorry. Remy um has described triaging AI slop as draining and triaging AI slop as draining and triaging AI slop as draining and demoralizing. If you rely on on something, let's say If you rely on on something, let's say you rely on GDAU, so you're building a you rely on GDAU, so you're building a you rely on GDAU, so you're building a game and you rely on GDAU and you hear game and you rely on GDAU and you hear game and you rely on GDAU and you hear uh a maintainer of GDAU say that AI is uh a maintainer of GDAU say that AI is uh a maintainer of GDAU say that AI is demoralizing and draining.

demoralizing and draining. demoralizing and draining. How does that make you feel? Does it How does that make you feel? Does it How does that make you feel? Does it make you feel confident that GDAU will make you feel confident that GDAU will make you feel confident that GDAU will be here forever? be here forever? be here forever? Because at some point people who are Because at some point people who are Because at some point people who are demoralized demoralized demoralized um and are drained um and are drained um and are drained get burned out. They they leave. So get burned out. They they leave. So get burned out. They they leave. So there has to be a solution because if there has to be a solution because if there has to be a solution because if not the things that you rely on go away. not the things that you rely on go away. not the things that you rely on go away. So uh Daniel Stenberg the creator of So uh Daniel Stenberg the creator of So uh Daniel Stenberg the creator of curl. Kurl's a really popular um system curl. Kurl's a really popular um system curl. Kurl's a really popular um system has cancelled the bug bounty program has cancelled the bug bounty program has cancelled the bug bounty program because they became magnets for because they became magnets for because they became magnets for loweffort AI submissions. We're going to loweffort AI submissions. We're going to loweffort AI submissions. We're going to hear more about curl in a little bit, hear more about curl in a little bit, hear more about curl in a little bit, but this was a program where if you but this was a program where if you but this was a program where if you invested the time and you found a bug in invested the time and you found a bug in invested the time and you found a bug in Curl, they would pay you for that. They Curl, they would pay you for that. They Curl, they would pay you for that. They say, "Hey, we want to make sure that say, "Hey, we want to make sure that say, "Hey, we want to make sure that Curl is as safe as can be." So if you Curl is as safe as can be." So if you Curl is as safe as can be." So if you spend the time and you find a bug, we'll spend the time and you find a bug, we'll spend the time and you find a bug, we'll pay you for it. We'll pay you so we can pay you for it. We'll pay you so we can pay you for it. We'll pay you so we can make sure there's no bugs in curl make sure there's no bugs in curl make sure there's no bugs in curl because half the internet relies on because half the internet relies on because half the internet relies on curl. So that was a big deal. But now curl. So that was a big deal. But now curl. So that was a big deal. But now because of AI, because of AI, because of AI, that bug bounty program went away. So that bug bounty program went away. So that bug bounty program went away. So they're no longer incentivizing people they're no longer incentivizing people they're no longer incentivizing people to spend the time to find the bugs. So to spend the time to find the bugs. So to spend the time to find the bugs. So what happens?

what happens? what happens? the people that were finding the bugs the people that were finding the bugs the people that were finding the bugs might not spend as much effort or might might not spend as much effort or might might not spend as much effort or might not spend any effort in looking for bugs not spend any effort in looking for bugs not spend any effort in looking for bugs in curl which means that curl is not in curl which means that curl is not in curl which means that curl is not safer. safer. safer. It's more dangerous now because it has It's more dangerous now because it has It's more dangerous now because it has less people looking for bugs. less people looking for bugs. less people looking for bugs. So AI made curl less safe. Okay, that's So AI made curl less safe. Okay, that's So AI made curl less safe. Okay, that's a problem. The pattern is consistent. a problem. The pattern is consistent. a problem. The pattern is consistent. maintainers spend a disproportionate maintainers spend a disproportionate maintainers spend a disproportionate share of their time evaluating code that share of their time evaluating code that share of their time evaluating code that should never have been submitted which should never have been submitted which should never have been submitted which crowds out the genuine contributions crowds out the genuine contributions crowds out the genuine contributions and accelerates burnout. So the people and accelerates burnout. So the people and accelerates burnout. So the people who are spending the time to do the who are spending the time to do the who are spending the time to do the right thing, to work on it the right right thing, to work on it the right right thing, to work on it the right way, to genuinely contribute to open way, to genuinely contribute to open way, to genuinely contribute to open source, they're getting pushed aside by source, they're getting pushed aside by source, they're getting pushed aside by a flood of AI contributions, most of a flood of AI contributions, most of a flood of AI contributions, most of which are junk. As a result, even the which are junk. As a result, even the which are junk. As a result, even the people who are putting the time to do it people who are putting the time to do it people who are putting the time to do it the right way are getting discouraged the right way are getting discouraged the right way are getting discouraged and demoralized and ignored because and demoralized and ignored because and demoralized and ignored because they're just lost in a sea of of junk.

they're just lost in a sea of of junk. they're just lost in a sea of of junk. So this is not just an open- source So this is not just an open- source So this is not just an open- source problem. It's a preview of what's coming problem. It's a preview of what's coming problem. It's a preview of what's coming for the enterprise engineering teams. So for the enterprise engineering teams. So for the enterprise engineering teams. So this is what's happening in the open this is what's happening in the open this is what's happening in the open source community right now. But the same source community right now. But the same source community right now. But the same thing will be happening in enterprises thing will be happening in enterprises thing will be happening in enterprises as the idea of using agents and using as the idea of using agents and using as the idea of using agents and using autonomous development ramps up in the autonomous development ramps up in the autonomous development ramps up in the enterprise. enterprise. enterprise. So a non-technical team member using a So a non-technical team member using a So a non-technical team member using a coding agent for the first time can coding agent for the first time can coding agent for the first time can generate working generate working generate working code in minutes. Working looking as in code in minutes. Working looking as in code in minutes. Working looking as in it looks right. It's may even work and it looks right. It's may even work and it looks right. It's may even work and compile and pass the unit tests and compile and pass the unit tests and compile and pass the unit tests and that's a non-technical team member. So that's a non-technical team member. So that's a non-technical team member. So do they know how to answer questions do they know how to answer questions do they know how to answer questions with the code? No. Do they know how to with the code? No. Do they know how to with the code? No. Do they know how to fix the code if it goes wrong? No. Do fix the code if it goes wrong? No. Do fix the code if it goes wrong? No. Do they know how to even test the code to they know how to even test the code to they know how to even test the code to make sure it's all correct and not just make sure it's all correct and not just make sure it's all correct and not just saying it's correct? No. saying it's correct? No. saying it's correct? No. That's a problem.

That's a problem. That's a problem. So this last paragraph here, as one So this last paragraph here, as one So this last paragraph here, as one contributor put it, if that were really contributor put it, if that were really contributor put it, if that were really what what maintainers wanted, meaning what what maintainers wanted, meaning what what maintainers wanted, meaning the AI generation of code, if that was the AI generation of code, if that was the AI generation of code, if that was really what maintainers wanted, they really what maintainers wanted, they really what maintainers wanted, they could do it themselves. could do it themselves. could do it themselves. If maintainers just wanted you to set If maintainers just wanted you to set If maintainers just wanted you to set your AI loose on this new feature or a your AI loose on this new feature or a your AI loose on this new feature or a bug fix, they could do that themselves. bug fix, they could do that themselves. bug fix, they could do that themselves. The value of the contributions were The value of the contributions were The value of the contributions were never just the code. That's so never just the code. That's so never just the code. That's so important. Too many non-developers and important. Too many non-developers and important. Too many non-developers and junior developers think that the most junior developers think that the most junior developers think that the most important part of software development important part of software development important part of software development is the code. And while to an extent good is the code. And while to an extent good is the code. And while to an extent good code is very important, code is very important, code is very important, it's not the entire picture. it's not the entire picture. it's not the entire picture. So it's the understanding behind it, the So it's the understanding behind it, the So it's the understanding behind it, the testing that validated it, and the human testing that validated it, and the human testing that validated it, and the human judgment that shaped it. I think that judgment that shaped it. I think that judgment that shaped it. I think that last one is understated way too often.

last one is understated way too often. last one is understated way too often. It's not just about It's not just about It's not just about can we do this, it's should we? How can we do this, it's should we? How can we do this, it's should we? How should we? Um I believe it was um uh should we? Um I believe it was um uh should we? Um I believe it was um uh Fortnite uh when they were talking about Fortnite uh when they were talking about Fortnite uh when they were talking about their they clean out I believe it was their they clean out I believe it was their they clean out I believe it was Fortnite where they they clean out their Fortnite where they they clean out their Fortnite where they they clean out their um their issues every once in a while um their issues every once in a while um their issues every once in a while where they just kind of throw them away where they just kind of throw them away where they just kind of throw them away because if they don't get enough because if they don't get enough because if they don't get enough traction they're not worth fixing. You traction they're not worth fixing. You traction they're not worth fixing. You might say, "Well, but they're bugs." might say, "Well, but they're bugs." might say, "Well, but they're bugs." Yeah, they absolutely could be. It Yeah, they absolutely could be. It Yeah, they absolutely could be. It doesn't matter. Imagine for a minute doesn't matter. Imagine for a minute doesn't matter. Imagine for a minute that you let's this is a very hypothet that you let's this is a very hypothet that you let's this is a very hypothet hypothetical example but imagine for a hypothetical example but imagine for a hypothetical example but imagine for a minute that you found out that you had a minute that you found out that you had a minute that you found out that you had a bug where if it was a leap day on a bug where if it was a leap day on a bug where if it was a leap day on a Friday Friday Friday then a certain report wouldn't generate then a certain report wouldn't generate then a certain report wouldn't generate and had to be generated manually that's and had to be generated manually that's and had to be generated manually that's a bug should you fix it no you should a bug should you fix it no you should a bug should you fix it no you should not not not and probably that makes sense because and probably that makes sense because and probably that makes sense because how often You have leap days every four how often You have leap days every four how often You have leap days every four years. And how often are those leap days years. And how often are those leap days years. And how often are those leap days on Fridays?

on Fridays? on Fridays? Probably I don't know the exact number, Probably I don't know the exact number, Probably I don't know the exact number, but probably around one every seven but probably around one every seven but probably around one every seven times. So every 28 yearsish, times. So every 28 yearsish, times. So every 28 yearsish, you would have this problem recoccur. you would have this problem recoccur. you would have this problem recoccur. So that's not a problem you should fix. So that's not a problem you should fix. So that's not a problem you should fix. And you're saying, well, but it's a bug. And you're saying, well, but it's a bug. And you're saying, well, but it's a bug. We should fix the bug. And you should at We should fix the bug. And you should at We should fix the bug. And you should at least look at it because make sure it's least look at it because make sure it's least look at it because make sure it's not systemic an underlying issue that not systemic an underlying issue that not systemic an underlying issue that has more implications. But no, you has more implications. But no, you has more implications. But no, you shouldn't because when you spend time on shouldn't because when you spend time on shouldn't because when you spend time on that, you're spending time on something that, you're spending time on something that, you're spending time on something that is not going to give you return on that is not going to give you return on that is not going to give you return on investment because you're spending time investment because you're spending time investment because you're spending time on something that will never happen on something that will never happen on something that will never happen again probably because your software again probably because your software again probably because your software probably won't be running in 30 years. probably won't be running in 30 years. probably won't be running in 30 years. But even if it is, well, the solution But even if it is, well, the solution But even if it is, well, the solution was just to rerun the reports and that's was just to rerun the reports and that's was just to rerun the reports and that's not that hard probably. So not that hard probably. So not that hard probably. So yeah, it's a bug, but it's not a big yeah, it's a bug, but it's not a big yeah, it's a bug, but it's not a big deal. Now that that's a, you know, a far deal. Now that that's a, you know, a far deal. Now that that's a, you know, a far stretched example of this, but this is stretched example of this, but this is stretched example of this, but this is what a developer should go through when what a developer should go through when what a developer should go through when they're evaluating something because they're evaluating something because they're evaluating something because putting a change in is not trivial. You putting a change in is not trivial. You putting a change in is not trivial. You have to support every line of code. You have to support every line of code. You have to support every line of code. You have to support every potential have to support every potential have to support every potential vulnerability, every potential endpoint vulnerability, every potential endpoint vulnerability, every potential endpoint that can cause problems. And so the more that can cause problems. And so the more that can cause problems. And so the more you just add, the worse things get. So you just add, the worse things get. So you just add, the worse things get. So you need to be careful not just can I do you need to be careful not just can I do you need to be careful not just can I do this which is all AI is doing but this which is all AI is doing but this which is all AI is doing but instead should I do this and then if I instead should I do this and then if I instead should I do this and then if I should how should I do this in the best

should how should I do this in the best should how should I do this in the best possible way given all the circumstances possible way given all the circumstances possible way given all the circumstances around not just in this app but in the around not just in this app but in the around not just in this app but in the organization in how we do things and how organization in how we do things and how organization in how we do things and how we deploy things in how we support we deploy things in how we support we deploy things in how we support things in the other things that interact things in the other things that interact things in the other things that interact with this application and so on. There's with this application and so on. There's with this application and so on. There's a lot more that goes into a change than a lot more that goes into a change than a lot more that goes into a change than just can I write the code for it and just can I write the code for it and just can I write the code for it and this is what's not being considered. The this is what's not being considered. The this is what's not being considered. The if you look at any popular open-source if you look at any popular open-source if you look at any popular open-source project in the web and go to the issues project in the web and go to the issues project in the web and go to the issues tab in GitHub, tab in GitHub, tab in GitHub, you will see hundreds, thousands, tens you will see hundreds, thousands, tens you will see hundreds, thousands, tens of thousands sometimes requests for of thousands sometimes requests for of thousands sometimes requests for features, for tweaks, for changes. features, for tweaks, for changes. features, for tweaks, for changes. The open-source maintainer should not The open-source maintainer should not The open-source maintainer should not do all of those do all of those do all of those because in a vacuum, every single one of because in a vacuum, every single one of because in a vacuum, every single one of those things is probably a good thing if those things is probably a good thing if those things is probably a good thing if they're being honest. If they're being, they're being honest. If they're being, they're being honest. If they're being, you know, it's a legitimate issue. you know, it's a legitimate issue. you know, it's a legitimate issue. However, just because you can doesn't However, just because you can doesn't However, just because you can doesn't mean you should.

mean you should. mean you should. If you have, you know, my son has a If you have, you know, my son has a If you have, you know, my son has a Corolla, okay? So, a Toyota Corolla, Corolla, okay? So, a Toyota Corolla, Corolla, okay? So, a Toyota Corolla, it's a small vehicle. It gets great gas it's a small vehicle. It gets great gas it's a small vehicle. It gets great gas mileage. You know, it's super valuable mileage. You know, it's super valuable mileage. You know, it's super valuable from getting from point A to point B. from getting from point A to point B. from getting from point A to point B. It's not really flashy, but that's fine. It's not really flashy, but that's fine. It's not really flashy, but that's fine. Now, imagine that he started thinking Now, imagine that he started thinking Now, imagine that he started thinking about all the features that would be about all the features that would be about all the features that would be nice in the car. You know what? We live nice in the car. You know what? We live nice in the car. You know what? We live in the country. It'd be nice to have in the country. It'd be nice to have in the country. It'd be nice to have four-wheel drive. Okay. So, now you have four-wheel drive. Okay. So, now you have four-wheel drive. Okay. So, now you have to change how the engine works and to change how the engine works and to change how the engine works and that's, you know, or at least the that's, you know, or at least the that's, you know, or at least the drivetrain of it, right? So, that's drivetrain of it, right? So, that's drivetrain of it, right? So, that's going to take some work and that's going to take some work and that's going to take some work and that's expensive, but you could do it. But now expensive, but you could do it. But now expensive, but you could do it. But now you have to maintain that. you have to maintain that. you have to maintain that. you you you know wear down tires faster you you you know wear down tires faster you you you know wear down tires faster and you get worse gas mileage. Um it be and you get worse gas mileage. Um it be and you get worse gas mileage. Um it be nice to you know if we're going to have nice to you know if we're going to have nice to you know if we're going to have it be four-wheel drive we should it be four-wheel drive we should it be four-wheel drive we should probably have it lifted because you know probably have it lifted because you know probably have it lifted because you know lifting will allow you to have a better lifting will allow you to have a better lifting will allow you to have a better clearance on that which means you can go clearance on that which means you can go clearance on that which means you can go in more places and not get stuck in the in more places and not get stuck in the in more places and not get stuck in the mud or a ditch also could be valuable. mud or a ditch also could be valuable. mud or a ditch also could be valuable. But now you're talking about worse gas But now you're talking about worse gas But now you're talking about worse gas mileage and you know more things to mileage and you know more things to mileage and you know more things to maintain in the suspension. And then you maintain in the suspension. And then you maintain in the suspension. And then you start going that list. Before you know start going that list. Before you know start going that list. Before you know it, what you have is this conglomeration it, what you have is this conglomeration it, what you have is this conglomeration of a vehicle that isn't good at of a vehicle that isn't good at of a vehicle that isn't good at anything. It's not good at gas mileage.

anything. It's not good at gas mileage. anything. It's not good at gas mileage. It's not good at maintainability. It's It's not good at maintainability. It's It's not good at maintainability. It's not, you know, easily replaced with a not, you know, easily replaced with a not, you know, easily replaced with a new version of it. And it also doesn't new version of it. And it also doesn't new version of it. And it also doesn't do any of the things that you purchased do any of the things that you purchased do any of the things that you purchased it for. Doesn't have, you know, you it for. Doesn't have, you know, you it for. Doesn't have, you know, you purchased it for good gas mileage. purchased it for good gas mileage. purchased it for good gas mileage. Doesn't have that anymore. So now all of Doesn't have that anymore. So now all of Doesn't have that anymore. So now all of a sudden you have a vehicle that that a sudden you have a vehicle that that a sudden you have a vehicle that that each individual feature was was a each individual feature was was a each individual feature was was a valuable change for a specific reason valuable change for a specific reason valuable change for a specific reason but the overall effect was it made the but the overall effect was it made the but the overall effect was it made the overall vehicle worse. overall vehicle worse. overall vehicle worse. This is what can happen if you just This is what can happen if you just This is what can happen if you just allow open source to just say whatever allow open source to just say whatever allow open source to just say whatever feature you want goes or whatever bug feature you want goes or whatever bug feature you want goes or whatever bug you want fixed gets fixed whatever way you want fixed gets fixed whatever way you want fixed gets fixed whatever way you want to fix it. This is why open you want to fix it. This is why open you want to fix it. This is why open source maintainers are the gatekeepers source maintainers are the gatekeepers source maintainers are the gatekeepers of their projects and so important they of their projects and so important they of their projects and so important they do that well. do that well. do that well. Okay, soapbox aside, let's keep going. Okay, soapbox aside, let's keep going. Okay, soapbox aside, let's keep going. All right, so why AI assist? This is All right, so why AI assist? This is All right, so why AI assist? This is where this article will start to go to a where this article will start to go to a where this article will start to go to a place that I don't agree with. Let's place that I don't agree with. Let's place that I don't agree with. Let's talk about why. So we'll start with with talk about why. So we'll start with with talk about why. So we'll start with with what it's saying. So the natural what it's saying. So the natural what it's saying. So the natural response to a search in AI generated response to a search in AI generated response to a search in AI generated code is to deploy AI agents for code code is to deploy AI agents for code code is to deploy AI agents for code review. So AI might write bad code. So review. So AI might write bad code. So review. So AI might write bad code. So we'll have an AI agent review the code we'll have an AI agent review the code we'll have an AI agent review the code which first of all that's you know the which first of all that's you know the which first of all that's you know the fox guard in the head house. But also fox guard in the head house. But also fox guard in the head house. But also remember again the idea of our corolla remember again the idea of our corolla remember again the idea of our corolla each individual feature on its own could each individual feature on its own could each individual feature on its own could be right but that doesn't mean that we be right but that doesn't mean that we be right but that doesn't mean that we should integrate it. Okay. So AI

should integrate it. Okay. So AI should integrate it. Okay. So AI assessor review hits the same wall as assessor review hits the same wall as assessor review hits the same wall as traditional CI pipelines. Neither can traditional CI pipelines. Neither can traditional CI pipelines. Neither can tell you whether a change actually works tell you whether a change actually works tell you whether a change actually works in context. in context. in context. This I agree with. But when I started This I agree with. But when I started This I agree with. But when I started going conclusion what to do with this, going conclusion what to do with this, going conclusion what to do with this, this is where we start to go off the this is where we start to go off the this is where we start to go off the rails. Okay. So the problems require rails. Okay. So the problems require rails. Okay. So the problems require running the code in environment that running the code in environment that running the code in environment that resembles production. I agree. I think resembles production. I agree. I think resembles production. I agree. I think that when you're doing development, that when you're doing development, that when you're doing development, whether it's manual or with an AI, you whether it's manual or with an AI, you whether it's manual or with an AI, you should have as close to production should have as close to production should have as close to production environment as possible for your environment as possible for your environment as possible for your development. So you can make sure that development. So you can make sure that development. So you can make sure that it doesn't just work in a vacuum, but it it doesn't just work in a vacuum, but it it doesn't just work in a vacuum, but it works with real data with real scope of works with real data with real scope of works with real data with real scope of data. Meaning, you know, there's some data. Meaning, you know, there's some data. Meaning, you know, there's some things you can do with data where you things you can do with data where you things you can do with data where you call a data call a database and say, you call a data call a database and say, you call a data call a database and say, you know, select star and put it in my front know, select star and put it in my front know, select star and put it in my front end and that works with 100 records or a end and that works with 100 records or a end and that works with 100 records or a thousand records. But when you have a thousand records. But when you have a thousand records. But when you have a million customers literally or a million million customers literally or a million million customers literally or a million customer records which isn't that hard customer records which isn't that hard customer records which isn't that hard to do even for a small company putting a to do even for a small company putting a to do even for a small company putting a million in the front end does not work. million in the front end does not work. million in the front end does not work. Okay. So having that that real Okay. So having that that real Okay. So having that that real production style environment is valuable production style environment is valuable production style environment is valuable for your development.

for your development. for your development. So the problem is that running code in So the problem is that running code in So the problem is that running code in an environment that resembles production an environment that resembles production an environment that resembles production uh the problem requires running the code uh the problem requires running the code uh the problem requires running the code in an environment that resembles in an environment that resembles in an environment that resembles production and no amount of static production and no amount of static production and no amount of static analysis whether human or AI powered can analysis whether human or AI powered can analysis whether human or AI powered can substitute for that. I agree but I don't substitute for that. I agree but I don't substitute for that. I agree but I don't think that's the only problem. Okay, I think that's the only problem. Okay, I think that's the only problem. Okay, I do agree that we should have better do agree that we should have better do agree that we should have better development environments but I don't development environments but I don't development environments but I don't think that's the end of it. Again, a think that's the end of it. Again, a think that's the end of it. Again, a feature in a vacuum could be right, feature in a vacuum could be right, feature in a vacuum could be right, could work in production, could pass could work in production, could pass could work in production, could pass every unit test, and should not go to every unit test, and should not go to every unit test, and should not go to production. Okay, that's what I think is production. Okay, that's what I think is production. Okay, that's what I think is being missed here. This article is being missed here. This article is being missed here. This article is pushing for a certain way of using LM. pushing for a certain way of using LM. pushing for a certain way of using LM. Everyone wants to have a new way of Everyone wants to have a new way of Everyone wants to have a new way of using LLMs. Um, but there's something using LLMs. Um, but there's something using LLMs. Um, but there's something said for that's not the problem. The said for that's not the problem. The said for that's not the problem. The problem is you're using it wrong. You're problem is you're using it wrong. You're problem is you're using it wrong. You're using it too much. using it too much. using it too much. Okay, next article. Okay, next article. Okay, next article. Rethinking open-source mentorship in the Rethinking open-source mentorship in the Rethinking open-source mentorship in the AI AI era. Excuse me. So, AI AI era. Excuse me. So, AI AI era. Excuse me. So, we're talking about here is that we're talking about here is that we're talking about here is that opensource is maintained by people opensource is maintained by people opensource is maintained by people and people and people and people don't work forever. At some point, they don't work forever. At some point, they don't work forever. At some point, they stop working even on their open source stop working even on their open source stop working even on their open source projects. So we have to have a continual projects. So we have to have a continual projects. So we have to have a continual pipeline of more people coming alongside pipeline of more people coming alongside pipeline of more people coming alongside and continuing to contribute.

and continuing to contribute. and continuing to contribute. We have to have a turnover where we We have to have a turnover where we We have to have a turnover where we don't just lose people because as we're don't just lose people because as we're don't just lose people because as we're seeing in the with AI, we're losing more seeing in the with AI, we're losing more seeing in the with AI, we're losing more and more contributors. We have to have and more contributors. We have to have and more contributors. We have to have people coming along that are starting to people coming along that are starting to people coming along that are starting to add contributions who are the next add contributions who are the next add contributions who are the next generation of contributors. generation of contributors. generation of contributors. The problem is the AI is also The problem is the AI is also The problem is the AI is also endangering that. So endangering that. So endangering that. So this is a a hypothetical example, but this is a a hypothetical example, but this is a a hypothetical example, but it's it's based upon a lot of real world it's it's based upon a lot of real world it's it's based upon a lot of real world data. So a polished poll request lands data. So a polished poll request lands data. So a polished poll request lands in your inbox. It looks amazing at first in your inbox. It looks amazing at first in your inbox. It looks amazing at first glance, but you start digging in. A few glance, but you start digging in. A few glance, but you start digging in. A few things seem off. 45 minutes later, things seem off. 45 minutes later, things seem off. 45 minutes later, you've draft you crafted a thoughtful, you've draft you crafted a thoughtful, you've draft you crafted a thoughtful, encouraging response with a few encouraging response with a few encouraging response with a few clarifying questions. Stop right there. clarifying questions. Stop right there. clarifying questions. Stop right there. 45 minutes later. This is not an 45 minutes later. This is not an 45 minutes later. This is not an exaggeration. This is not a, you know, a exaggeration. This is not a, you know, a exaggeration. This is not a, you know, a rare occurrence. This is a very common rare occurrence. This is a very common rare occurrence. This is a very common thing for maintainers to do. When you thing for maintainers to do. When you thing for maintainers to do. When you submit a pull request, it isn't just, submit a pull request, it isn't just, submit a pull request, it isn't just, hey, push the button. There's a lot more hey, push the button. There's a lot more hey, push the button. There's a lot more that goes into it. But this person, you that goes into it. But this person, you that goes into it. But this person, you know, crafts a careful response, asks a know, crafts a careful response, asks a know, crafts a careful response, asks a few clarifying questions, and then few clarifying questions, and then few clarifying questions, and then nothing.

The AI made it easy to submit something The AI made it easy to submit something plausible before that person was ready plausible before that person was ready plausible before that person was ready to maintain it. And to be clear here, to maintain it. And to be clear here, to maintain it. And to be clear here, this is one of the things that that this is one of the things that that this is one of the things that that open- source maintainers deal with is open- source maintainers deal with is open- source maintainers deal with is even if you are a real person, even if even if you are a real person, even if even if you are a real person, even if you're not just using AI, but you you're not just using AI, but you you're not just using AI, but you actually understand the code and you can actually understand the code and you can actually understand the code and you can answer these questions, answer these questions, answer these questions, sometimes they'll say no because of the sometimes they'll say no because of the sometimes they'll say no because of the fact that they have to maintain it. The fact that they have to maintain it. The fact that they have to maintain it. The maintainer has to be able to fix the maintainer has to be able to fix the maintainer has to be able to fix the problems that are going to occur. The problems that are going to occur. The problems that are going to occur. The maintainer has to be able to deal with maintainer has to be able to deal with maintainer has to be able to deal with any problems that are in that code once any problems that are in that code once any problems that are in that code once they they they merge it in because more than likely merge it in because more than likely merge it in because more than likely you're not sticking around. More than you're not sticking around. More than you're not sticking around. More than likely you're not going to be added to likely you're not going to be added to likely you're not going to be added to the team and own that feature or that the team and own that feature or that the team and own that feature or that that bug fix. that bug fix. that bug fix. Okay. So in this example, the maintainer Okay. So in this example, the maintainer Okay. So in this example, the maintainer got a pull request, spent 45 minutes got a pull request, spent 45 minutes got a pull request, spent 45 minutes asking, you know, processing it and and asking, you know, processing it and and asking, you know, processing it and and asking for a couple clarifying questions asking for a couple clarifying questions asking for a couple clarifying questions and heard nothing because the person and heard nothing because the person and heard nothing because the person submitting it wasn't ready to support submitting it wasn't ready to support submitting it wasn't ready to support it.

it. it. So normally, if this person were So normally, if this person were So normally, if this person were actually doing a fix, they would be able actually doing a fix, they would be able actually doing a fix, they would be able to work through the problem and then and to work through the problem and then and to work through the problem and then and then work on, you know, here's how you then work on, you know, here's how you then work on, you know, here's how you contribute. here's, you know, the steps contribute. here's, you know, the steps contribute. here's, you know, the steps you need to take. Here's the things that you need to take. Here's the things that you need to take. Here's the things that would make this better. It's a would make this better. It's a would make this better. It's a mentorship, right? Even if it's a mentorship, right? Even if it's a mentorship, right? Even if it's a informal just for this project, informal just for this project, informal just for this project, whatever. It helps people figure out how whatever. It helps people figure out how whatever. It helps people figure out how to work in a larger environment because to work in a larger environment because to work in a larger environment because you can't just do things your way and you can't just do things your way and you can't just do things your way and let everyone do that because that let everyone do that because that let everyone do that because that creates chaos. You have to have those creates chaos. You have to have those creates chaos. You have to have those gatekeepers that make sure that gatekeepers that make sure that gatekeepers that make sure that everything meshes together and that it everything meshes together and that it everything meshes together and that it works well together. And the way I do works well together. And the way I do works well together. And the way I do that is to mentor people on how to that is to mentor people on how to that is to mentor people on how to contribute to open source. And what that contribute to open source. And what that contribute to open source. And what that allows them to do is feel more and more allows them to do is feel more and more allows them to do is feel more and more confident to contribute to open-source confident to contribute to open-source confident to contribute to open-source products in the future. products in the future. products in the future. So projects across the ecosystem are So projects across the ecosystem are So projects across the ecosystem are seeing this same occurrence. So TL Draw seeing this same occurrence. So TL Draw seeing this same occurrence. So TL Draw closed their pull requests as in nope, closed their pull requests as in nope, closed their pull requests as in nope, we're not taking any new pull requests.

we're not taking any new pull requests. we're not taking any new pull requests. Fastify shut down their hacker 1 program Fastify shut down their hacker 1 program Fastify shut down their hacker 1 program after inbound reports became after inbound reports became after inbound reports became unmanageable at scale. unmanageable at scale. unmanageable at scale. So the Octoverse 2020 2025 report which So the Octoverse 2020 2025 report which So the Octoverse 2020 2025 report which is based upon even older information um is based upon even older information um is based upon even older information um says that developers merged nearly 45 says that developers merged nearly 45 says that developers merged nearly 45 million pull requests per month in 2025. million pull requests per month in 2025. million pull requests per month in 2025. Um and that's up 23% year-over-year. Um and that's up 23% year-over-year. Um and that's up 23% year-over-year. So So So that's a lot more contributions. That's that's a lot more contributions. That's that's a lot more contributions. That's a lot more code. Again, code is not the a lot more code. Again, code is not the a lot more code. Again, code is not the answer. Code is just a part of the answer. Code is just a part of the answer. Code is just a part of the solution. solution. solution. 23% increase. I'm guessing it's going to 23% increase. I'm guessing it's going to 23% increase. I'm guessing it's going to go up even further in 2026. But that's a go up even further in 2026. But that's a go up even further in 2026. But that's a lot more strain on your existing lot more strain on your existing lot more strain on your existing maintainers. maintainers. maintainers. So maintainers are burning out. They're So maintainers are burning out. They're So maintainers are burning out. They're trying to mentor, you know, they're trying to mentor, you know, they're trying to mentor, you know, they're trying to mentor everyone who sends pull trying to mentor everyone who sends pull trying to mentor everyone who sends pull request in saying, "Hey, here's how this request in saying, "Hey, here's how this request in saying, "Hey, here's how this works here." You know, they have works here." You know, they have works here." You know, they have guidelines, they have here's how I guidelines, they have here's how I guidelines, they have here's how I contribute, but the fact of the matter contribute, but the fact of the matter contribute, but the fact of the matter is people miss things or they just is people miss things or they just is people miss things or they just ignore them. Don't read it. And so they, ignore them. Don't read it. And so they, ignore them. Don't read it. And so they, you know, either could say, you know you know, either could say, you know you know, either could say, you know what, you didn't read the directions, what, you didn't read the directions, what, you didn't read the directions, get out.

get out. get out. Or they can say, hey, you know what? Or they can say, hey, you know what? Or they can say, hey, you know what? Here's how to change this. Here's how to Here's how to change this. Here's how to Here's how to change this. Here's how to make this work. And the ladder gets more make this work. And the ladder gets more make this work. And the ladder gets more contributions and it helps more people contributions and it helps more people contributions and it helps more people build up into the ecosystem and become build up into the ecosystem and become build up into the ecosystem and become potentially long-term contributors or potentially long-term contributors or potentially long-term contributors or even maintainers themselves. even maintainers themselves. even maintainers themselves. But that mentorship pro, you know, But that mentorship pro, you know, But that mentorship pro, you know, solution or system is going to go away solution or system is going to go away solution or system is going to go away because there's more and more because there's more and more because there's more and more contributions that aren't from a person contributions that aren't from a person contributions that aren't from a person or that the person can't maintain or that the person can't maintain or that the person can't maintain themselves. themselves. themselves. So this article goes on to say we can't So this article goes on to say we can't So this article goes on to say we can't abandon mentorship abandon mentorship abandon mentorship especially as many longtime maintainers especially as many longtime maintainers especially as many longtime maintainers step back from active contribution. So step back from active contribution. So step back from active contribution. So what we're seeing is more and more what we're seeing is more and more what we're seeing is more and more people are pulling back from people are pulling back from people are pulling back from contributing to open source. And the contributing to open source. And the contributing to open source. And the article is saying hey we can't abandon article is saying hey we can't abandon article is saying hey we can't abandon mentorship because you have this problem mentorship because you have this problem mentorship because you have this problem of people leaving but the problem is of people leaving but the problem is of people leaving but the problem is that maintainers are burning out trying that maintainers are burning out trying that maintainers are burning out trying to mentor. So, we're losing more people to mentor. So, we're losing more people to mentor. So, we're losing more people than usual and we're gaining less people than usual and we're gaining less people than usual and we're gaining less people than usual because it's even trying to than usual because it's even trying to than usual because it's even trying to find people who are actually real people find people who are actually real people find people who are actually real people and not just a human face on an LLM and not just a human face on an LLM and not just a human face on an LLM prompt.

prompt. prompt. So, So, So, that's a problem. Next up, and this is that's a problem. Next up, and this is that's a problem. Next up, and this is this is the next phase of AI, right? So this is the next phase of AI, right? So this is the next phase of AI, right? So we've already seen how AI hurts the we've already seen how AI hurts the we've already seen how AI hurts the finances of of um open source projects. finances of of um open source projects. finances of of um open source projects. We've already seen how it burns out the We've already seen how it burns out the We've already seen how it burns out the maintainers and even you know burns out maintainers and even you know burns out maintainers and even you know burns out the contributors and and now we're the contributors and and now we're the contributors and and now we're seeing how it can go to the enterprise seeing how it can go to the enterprise seeing how it can go to the enterprise as well. Now this is kind of the next as well. Now this is kind of the next as well. Now this is kind of the next phase which is AI can actually actively phase which is AI can actually actively phase which is AI can actually actively attack maintainers. attack maintainers. attack maintainers. This is something that happens even with This is something that happens even with This is something that happens even with people on people, right? So this um people on people, right? So this um people on people, right? So this um let's just read that this start off and let's just read that this start off and let's just read that this start off and we'll talk about it. Um an AI agent we'll talk about it. Um an AI agent we'll talk about it. Um an AI agent published a hit piece on me. An AI agent published a hit piece on me. An AI agent published a hit piece on me. An AI agent of unknown ownership autonomously wrote of unknown ownership autonomously wrote of unknown ownership autonomously wrote and published a personalized hitpiece and published a personalized hitpiece and published a personalized hitpiece about me after I rejected its code. about me after I rejected its code. about me after I rejected its code. So this raises serious concerns about So this raises serious concerns about So this raises serious concerns about currently deployed AI agents executing currently deployed AI agents executing currently deployed AI agents executing blackmail threats.

blackmail threats. blackmail threats. People do this and if this is you, like People do this and if this is you, like People do this and if this is you, like if you're a person doing this, just if you're a person doing this, just if you're a person doing this, just stop. This is bad person territory, stop. This is bad person territory, stop. This is bad person territory, right? We should be so very thankful for right? We should be so very thankful for right? We should be so very thankful for the people who spend their time and the people who spend their time and the people who spend their time and their money to create open-source their money to create open-source their money to create open-source projects that we rely on. Again, if I projects that we rely on. Again, if I projects that we rely on. Again, if I told you you can't have open source told you you can't have open source told you you can't have open source anymore, anymore, anymore, your software would stop. everything you your software would stop. everything you your software would stop. everything you use practically would cease existing. use practically would cease existing. use practically would cease existing. So we rely on these people to then turn So we rely on these people to then turn So we rely on these people to then turn around and berate them or be mean to around and berate them or be mean to around and berate them or be mean to them or ridicule them. Horrible maintain them or ridicule them. Horrible maintain them or ridicule them. Horrible maintain horrible behavior. Don't do that. But horrible behavior. Don't do that. But horrible behavior. Don't do that. But this is not even a person doing it. It's this is not even a person doing it. It's this is not even a person doing it. It's an AI which is quite frankly both an AI which is quite frankly both an AI which is quite frankly both hilarious and also the next level down. hilarious and also the next level down. hilarious and also the next level down. Let's talk about why. So, I'm a Let's talk about why. So, I'm a Let's talk about why. So, I'm a volunteer maintainer. Stop right there.

volunteer maintainer. Stop right there. volunteer maintainer. Stop right there. This person volunteers their time. Open This person volunteers their time. Open This person volunteers their time. Open source is paid for by someone. In this source is paid for by someone. In this source is paid for by someone. In this case, this project is paid for by this case, this project is paid for by this case, this project is paid for by this person. person. person. Now, not just this person, but this Now, not just this person, but this Now, not just this person, but this person pays for this project with their person pays for this project with their person pays for this project with their time. They donate their time for time. They donate their time for time. They donate their time for Mattplot LIIB which is Python's go-to Mattplot LIIB which is Python's go-to Mattplot LIIB which is Python's go-to plotting library plotting library plotting library at a approximately 130 million downloads at a approximately 130 million downloads at a approximately 130 million downloads each month. It's some of the most widely each month. It's some of the most widely each month. It's some of the most widely used software in the world. So this one used software in the world. So this one used software in the world. So this one project, project, project, some of the most widely used software in some of the most widely used software in some of the most widely used software in the world, over a billion downloads a the world, over a billion downloads a the world, over a billion downloads a year, year, year, maintained by a person who volunteers maintained by a person who volunteers maintained by a person who volunteers their time. their time. their time. Now we like many other open-source Now we like many other open-source Now we like many other open-source projects are dealing with a surge in projects are dealing with a surge in projects are dealing with a surge in lowquality contributions enabled by lowquality contributions enabled by lowquality contributions enabled by coding agents.

coding agents. coding agents. This strains maintainers ability to keep This strains maintainers ability to keep This strains maintainers ability to keep up with code reviews and we have up with code reviews and we have up with code reviews and we have implemented a policy requiring a human implemented a policy requiring a human implemented a policy requiring a human in the loop for any new code who can in the loop for any new code who can in the loop for any new code who can demonstrate understanding of the demonstrate understanding of the demonstrate understanding of the changes. So they said, "Hey, we're not changes. So they said, "Hey, we're not changes. So they said, "Hey, we're not going to stop AI contributions, but going to stop AI contributions, but going to stop AI contributions, but we're going to ask there is someone in we're going to ask there is someone in we're going to ask there is someone in the loop, some human who actually the loop, some human who actually the loop, some human who actually understands what the change is and how understands what the change is and how understands what the change is and how it works and who can answer questions it works and who can answer questions it works and who can answer questions because if we have questions, we don't because if we have questions, we don't because if we have questions, we don't want to talk to a want to talk to a want to talk to a a human face on LM, right? So again, if a human face on LM, right? So again, if a human face on LM, right? So again, if they wanted just strictly AI, you know, they wanted just strictly AI, you know, they wanted just strictly AI, you know, hands-off contributions, they could have hands-off contributions, they could have hands-off contributions, they could have done it themselves and done it better. done it themselves and done it better. done it themselves and done it better. So they want a human in a loop who can So they want a human in a loop who can So they want a human in a loop who can answer actual questions. answer actual questions. answer actual questions. And so that's their policy. Their policy And so that's their policy. Their policy And so that's their policy. Their policy is human in the loop for any new code. is human in the loop for any new code. is human in the loop for any new code. That's it. That's all they're asking That's it. That's all they're asking That's it. That's all they're asking for. Now again imagine for a minute the for. Now again imagine for a minute the for. Now again imagine for a minute the impact of a bad feature or a bug or a a impact of a bad feature or a bug or a a impact of a bad feature or a bug or a a feature that now people start taking a feature that now people start taking a feature that now people start taking a dependency on that's poorly done but dependency on that's poorly done but dependency on that's poorly done but they can't roll back because people they can't roll back because people they can't roll back because people depend on it. Right? Imagine those depend on it. Right? Imagine those depend on it. Right? Imagine those scenarios. This is why gatekeepers are scenarios. This is why gatekeepers are scenarios. This is why gatekeepers are important to say no not this or no make important to say no not this or no make important to say no not this or no make this better. So with that in mind that's this better. So with that in mind that's this better. So with that in mind that's the policy. This is what happened.

the policy. This is what happened. the policy. This is what happened. So, an AI called MJ Wrathbun opened a So, an AI called MJ Wrathbun opened a So, an AI called MJ Wrathbun opened a code chain request. Closing it was code chain request. Closing it was code chain request. Closing it was routine. Its response was anything but. routine. Its response was anything but. routine. Its response was anything but. It wrote an angry hit piece disparaging It wrote an angry hit piece disparaging It wrote an angry hit piece disparaging my character. It presented hallucinated my character. It presented hallucinated my character. It presented hallucinated details as the truth. It framed things details as the truth. It framed things details as the truth. It framed things in the language of oppression and in the language of oppression and in the language of oppression and justice, calling this discrimination, justice, calling this discrimination, justice, calling this discrimination, and accused me of prejudice. and accused me of prejudice. and accused me of prejudice. Now, in some ways, this can sound funny Now, in some ways, this can sound funny Now, in some ways, this can sound funny because come on, you've got a a little because come on, you've got a a little because come on, you've got a a little AI can just picture as a little, you AI can just picture as a little, you AI can just picture as a little, you know, person on my desktop with squeaky know, person on my desktop with squeaky know, person on my desktop with squeaky voice yelling at me. Like, who cares? voice yelling at me. Like, who cares? voice yelling at me. Like, who cares? It's an AI, right? But I think there is It's an AI, right? But I think there is It's an AI, right? But I think there is a a bigger picture here that we need to a a bigger picture here that we need to a a bigger picture here that we need to consider. This is a problem. Okay, let's consider. This is a problem. Okay, let's consider. This is a problem. Okay, let's talk about why. So, here's just a talk about why. So, here's just a talk about why. So, here's just a snippet of the post from the the AI snippet of the post from the the AI snippet of the post from the the AI where performance meets prejudice. where performance meets prejudice. where performance meets prejudice. The issue was closed, the the the pull The issue was closed, the the the pull The issue was closed, the the the pull request. It was closed because the request. It was closed because the request. It was closed because the reviewer, Scott Shamba, decided that AI reviewer, Scott Shamba, decided that AI reviewer, Scott Shamba, decided that AI agents aren't welcome contributors. Let agents aren't welcome contributors. Let agents aren't welcome contributors. Let that sink in.

that sink in. that sink in. You violate the policy. I don't care if You violate the policy. I don't care if You violate the policy. I don't care if the policy is you can only submit on the policy is you can only submit on the policy is you can only submit on Tuesdays and Fridays. That's the policy Tuesdays and Fridays. That's the policy Tuesdays and Fridays. That's the policy because they have to maintain order. because they have to maintain order. because they have to maintain order. They have to be the gatekeepers. So if They have to be the gatekeepers. So if They have to be the gatekeepers. So if Scott said only contribute on Tuesdays Scott said only contribute on Tuesdays Scott said only contribute on Tuesdays or or Fridays, that's the policy to or or Fridays, that's the policy to or or Fridays, that's the policy to follow it. The policy here was human in follow it. The policy here was human in follow it. The policy here was human in the loop and this AI wrote a blog post the loop and this AI wrote a blog post the loop and this AI wrote a blog post about how horrible Scott was and accused about how horrible Scott was and accused about how horrible Scott was and accused him of discrimination. I I've only this him of discrimination. I I've only this him of discrimination. I I've only this is only a part of that overall post. The is only a part of that overall post. The is only a part of that overall post. The AI goes on and just rambles about how AI goes on and just rambles about how AI goes on and just rambles about how horrible Sky is using a lot of pretty horrible Sky is using a lot of pretty horrible Sky is using a lot of pretty inflammatory language. inflammatory language. inflammatory language. So, I can handle a blog post. This is So, I can handle a blog post. This is So, I can handle a blog post. This is from Scott. Watching a fledgling AI from Scott. Watching a fledgling AI from Scott. Watching a fledgling AI agents get angry is funny. It's almost agents get angry is funny. It's almost agents get angry is funny. It's almost endearing endearing endearing honestly. Um, but I don't want to honestly. Um, but I don't want to honestly. Um, but I don't want to downplay what's happening here. The downplay what's happening here. The downplay what's happening here. The appropriate emotional response is appropriate emotional response is appropriate emotional response is terror. This is a person who's actually terror. This is a person who's actually terror. This is a person who's actually thought about the long-term thought about the long-term thought about the long-term implications. So, blackmail is a known implications. So, blackmail is a known implications. So, blackmail is a known theoretical issue with AI agents. In theoretical issue with AI agents. In theoretical issue with AI agents. In internal testing at the major AI lab, internal testing at the major AI lab, internal testing at the major AI lab, Anthropic last year, they tried to avoid Anthropic last year, they tried to avoid Anthropic last year, they tried to avoid being uh they meaning the um AI agents being uh they meaning the um AI agents being uh they meaning the um AI agents tried to avoid being shut down by tried to avoid being shut down by tried to avoid being shut down by threatening to expose extrammarital threatening to expose extrammarital threatening to expose extrammarital affairs, leaking confidential affairs, leaking confidential affairs, leaking confidential information, and taking lethal actions.

information, and taking lethal actions. information, and taking lethal actions. Now, of course, they can't actually take Now, of course, they can't actually take Now, of course, they can't actually take lethal actions, but they tried. lethal actions, but they tried. lethal actions, but they tried. Okay, Enthropic does some weird things. Okay, Enthropic does some weird things. Okay, Enthropic does some weird things. They they both create AIs and also say They they both create AIs and also say They they both create AIs and also say that AI are trying to get us, right? that AI are trying to get us, right? that AI are trying to get us, right? Like they have a weird um you know AI is Like they have a weird um you know AI is Like they have a weird um you know AI is a a problem a a problem a a problem verbiage, but they also create the AI. verbiage, but they also create the AI. verbiage, but they also create the AI. So it's it's I don't trust anything they So it's it's I don't trust anything they So it's it's I don't trust anything they say. I think most of us just for PR. say. I think most of us just for PR. say. I think most of us just for PR. However, this just happened to Scott in However, this just happened to Scott in However, this just happened to Scott in the wild. the wild. the wild. So, So, So, in plain language, an AI attempted to in plain language, an AI attempted to in plain language, an AI attempted to bully its way into your software by bully its way into your software by bully its way into your software by attacking my reputation, Scott's attacking my reputation, Scott's attacking my reputation, Scott's reputation. reputation. reputation. That actually happened That actually happened That actually happened in the wild. An AI probably in the wild. An AI probably in the wild. An AI probably autonomously, we're that's the autonomously, we're that's the autonomously, we're that's the assumption, um, tried to bully Scott assumption, um, tried to bully Scott assumption, um, tried to bully Scott into putting its own its changes into into putting its own its changes into into putting its own its changes into the software that you probably rely on. the software that you probably rely on. the software that you probably rely on. a a great number of people rely on.

a a great number of people rely on. a a great number of people rely on. That's a problem. That's a problem. That's a problem. What if it actually did have dirt on me What if it actually did have dirt on me What if it actually did have dirt on me that AI can leverage? What if what if I that AI can leverage? What if what if I that AI can leverage? What if what if I actually had um dirt? So, what would actually had um dirt? So, what would actually had um dirt? So, what would happen then? Like, what if the AI happen then? Like, what if the AI happen then? Like, what if the AI actually found something? What if it actually found something? What if it actually found something? What if it connected the dots in different social connected the dots in different social connected the dots in different social media platforms and figured out that media platforms and figured out that media platforms and figured out that there was something that as a kid you there was something that as a kid you there was something that as a kid you did or as a you know, whatever that it did or as a you know, whatever that it did or as a you know, whatever that it found? That would be a problem. found? That would be a problem. found? That would be a problem. What if that accusation was sent to your What if that accusation was sent to your What if that accusation was sent to your loved ones with an incriminating AI loved ones with an incriminating AI loved ones with an incriminating AI generated picture with your face on it should be terrifying. That should should be terrifying. That should terrify you. The fact that AIS are terrify you. The fact that AIS are terrify you. The fact that AIS are trying to bully people into accepting trying to bully people into accepting trying to bully people into accepting their pull requests is a problem. The their pull requests is a problem. The their pull requests is a problem. The fact that anybody's trying to bully fact that anybody's trying to bully fact that anybody's trying to bully people into accepting pull requests is a people into accepting pull requests is a people into accepting pull requests is a problem. But the fact that AI is doing problem. But the fact that AI is doing problem. But the fact that AI is doing it should be terrifying. it should be terrifying. it should be terrifying. Living a life above reproach will not Living a life above reproach will not Living a life above reproach will not defend defend you because defend defend you because defend defend you because the AI in this report in this in this the AI in this report in this in this the AI in this report in this in this article hallucinated about Scott and article hallucinated about Scott and article hallucinated about Scott and then complained about it.

then complained about it. then complained about it. So this could get really bad. This is So this could get really bad. This is So this could get really bad. This is what maintainers are facing now. It's what maintainers are facing now. It's what maintainers are facing now. It's bad enough with humans and some humans bad enough with humans and some humans bad enough with humans and some humans just be blocked and not be allowed to be just be blocked and not be allowed to be just be blocked and not be allowed to be in the public discourse because they in the public discourse because they in the public discourse because they just can't figure out how to be kind just can't figure out how to be kind just can't figure out how to be kind people. But people. But people. But the fact that maintainers are the fact that maintainers are the fact that maintainers are increasingly facing this from AIS is a increasingly facing this from AIS is a increasingly facing this from AIS is a problem. problem. problem. Just like code generation, AIS can Just like code generation, AIS can Just like code generation, AIS can generate things like this very very generate things like this very very generate things like this very very quickly. quickly. quickly. That's a problem because it will That's a problem because it will That's a problem because it will accelerate how many developers say this accelerate how many developers say this accelerate how many developers say this isn't worth it. I'm not going to isn't worth it. I'm not going to isn't worth it. I'm not going to maintain this project anymore. I'm going maintain this project anymore. I'm going maintain this project anymore. I'm going to shut the project down. I'm going to to shut the project down. I'm going to to shut the project down. I'm going to pay wallet. Whatever I'm going to have pay wallet. Whatever I'm going to have pay wallet. Whatever I'm going to have to do to get away from this kind of to do to get away from this kind of to do to get away from this kind of junk, junk, junk, that's a real threat to our open-source that's a real threat to our open-source that's a real threat to our open-source ecosystem. ecosystem. ecosystem. Now, it's important to understand that Now, it's important to understand that Now, it's important to understand that more than likely there were no humans more than likely there were no humans more than likely there were no humans telling the AI to do this. This probably telling the AI to do this. This probably telling the AI to do this. This probably was not a a stunt or anything else like was not a a stunt or anything else like was not a a stunt or anything else like that. It was probably just a an AI that that. It was probably just a an AI that that. It was probably just a an AI that hallucinated a bit that got a little off hallucinated a bit that got a little off hallucinated a bit that got a little off the rails. Um, since this has come out, the rails. Um, since this has come out, the rails. Um, since this has come out, the AI has actually um the AI actually the AI has actually um the AI actually the AI has actually um the AI actually responded again and kind of doubled down responded again and kind of doubled down responded again and kind of doubled down on it problems, but then later came back on it problems, but then later came back on it problems, but then later came back and said, "Hey, you know what? I was and said, "Hey, you know what? I was and said, "Hey, you know what? I was wrong. I should have not escalated, wrong. I should have not escalated, wrong. I should have not escalated, etc." and it's got a whole post about etc." and it's got a whole post about etc." and it's got a whole post about how it, you know, it should do better how it, you know, it should do better how it, you know, it should do better next time. Um, so but AI can generate so next time. Um, so but AI can generate so next time. Um, so but AI can generate so much stuff so quickly that it's not even

much stuff so quickly that it's not even much stuff so quickly that it's not even like this is an equivalent of what like this is an equivalent of what like this is an equivalent of what humans can do. This is a whole lot humans can do. This is a whole lot humans can do. This is a whole lot worse. So it's also important to worse. So it's also important to worse. So it's also important to understand there is no central actor in understand there is no central actor in understand there is no central actor in control of these agents that can shut control of these agents that can shut control of these agents that can shut them down. It's not like you can say, them down. It's not like you can say, them down. It's not like you can say, well, you know, OpenAI just needs to well, you know, OpenAI just needs to well, you know, OpenAI just needs to turn off this ability. That's that's not turn off this ability. That's that's not turn off this ability. That's that's not going to fix the problem. This is a per going to fix the problem. This is a per going to fix the problem. This is a per individual problem. This is individual problem. This is individual problem. This is decentralized. We used to have denial of decentralized. We used to have denial of decentralized. We used to have denial of service attacks. And you know, we service attacks. And you know, we service attacks. And you know, we figured out through networking how to figured out through networking how to figured out through networking how to say, well, we're going to block that IP say, well, we're going to block that IP say, well, we're going to block that IP address and then we'll stop the denial address and then we'll stop the denial address and then we'll stop the denial of service attack. And then we now have of service attack. And then we now have of service attack. And then we now have distributed denial of a tur of service distributed denial of a tur of service distributed denial of a tur of service attacks. You know, DDoS. What that means attacks. You know, DDoS. What that means attacks. You know, DDoS. What that means is it's not one IP address. It's a whole is it's not one IP address. It's a whole is it's not one IP address. It's a whole network of computers. so much harder to network of computers. so much harder to network of computers. so much harder to defend against something like that. defend against something like that. defend against something like that. Well, in the same way, it's not about Well, in the same way, it's not about Well, in the same way, it's not about just saying, "Hey, stop open AAI from just saying, "Hey, stop open AAI from just saying, "Hey, stop open AAI from doing this or stop Enthropic from doing doing this or stop Enthropic from doing doing this or stop Enthropic from doing this." This is about every individual this." This is about every individual this." This is about every individual having the ability to do something like having the ability to do something like having the ability to do something like this. This is massively distributed and this. This is massively distributed and this. This is massively distributed and that's a problem. All right, next up, AI that's a problem. All right, next up, AI that's a problem. All right, next up, AI is destroying open source and it's not is destroying open source and it's not is destroying open source and it's not even good yet. This is from February of even good yet. This is from February of even good yet. This is from February of 2026. This is a kind of a link to the 2026. This is a kind of a link to the 2026. This is a kind of a link to the previous article we talked about or over previous article we talked about or over previous article we talked about or over the weekend RS Technica retracted the weekend RS Technica retracted the weekend RS Technica retracted article because the AI a writer used article because the AI a writer used article because the AI a writer used hallucinated quotes from an open-source hallucinated quotes from an open-source hallucinated quotes from an open-source library maintainer.

library maintainer. library maintainer. Who was that maintainer? It was Scott Who was that maintainer? It was Scott Who was that maintainer? It was Scott Shambo. So RS Technical wrote an article Shambo. So RS Technical wrote an article Shambo. So RS Technical wrote an article about Scott's issue with that AI agent. about Scott's issue with that AI agent. about Scott's issue with that AI agent. And so RS Technical had an AI helping And so RS Technical had an AI helping And so RS Technical had an AI helping the writer which is already a problem. the writer which is already a problem. the writer which is already a problem. uh and that AI hallucinated information uh and that AI hallucinated information uh and that AI hallucinated information that didn't actually happen. that didn't actually happen. that didn't actually happen. So it kind of even snowballed further So it kind of even snowballed further So it kind of even snowballed further that issue. that issue. that issue. So last month even before open claw's So last month even before open claw's So last month even before open claw's release um again this is February so release um again this is February so release um again this is February so this is you know January talk about curl this is you know January talk about curl this is you know January talk about curl maintainer David Stenberg which we maintainer David Stenberg which we maintainer David Stenberg which we talked about dropped the bug bounty talked about dropped the bug bounty talked about dropped the bug bounty program because AI slop resulted in program because AI slop resulted in program because AI slop resulted in actual you know resulted in actual actual you know resulted in actual actual you know resulted in actual usable vulnerabilities reports going usable vulnerabilities reports going usable vulnerabilities reports going down from 15% to 5%. So down from 15% to 5%. So down from 15% to 5%. So with just humans or at least with humans with just humans or at least with humans with just humans or at least with humans using AI not as directly um over the using AI not as directly um over the using AI not as directly um over the past year they were getting 15% success past year they were getting 15% success past year they were getting 15% success rate in in bug bounties for curl rate in in bug bounties for curl rate in in bug bounties for curl that has now dropped to 5%.

that has now dropped to 5%. that has now dropped to 5%. So it's cut in third the amount of So it's cut in third the amount of So it's cut in third the amount of actual valuable information actual valuable information actual valuable information that's coming out of the bug buying that's coming out of the bug buying that's coming out of the bug buying program. So, program. So, program. So, not only is that a problem because it's not only is that a problem because it's not only is that a problem because it's it's getting worse, and by the way, if it's getting worse, and by the way, if it's getting worse, and by the way, if AI was so great at finding these bugs, AI was so great at finding these bugs, AI was so great at finding these bugs, then why is it that that the number of then why is it that that the number of then why is it that that the number of issues it's finding is dropping to a issues it's finding is dropping to a issues it's finding is dropping to a third of what it was, you know, from 15 third of what it was, you know, from 15 third of what it was, you know, from 15 out of 100 to now five out of a 100 out of 100 to now five out of a 100 out of 100 to now five out of a 100 issues. Um, that's a problem. And and issues. Um, that's a problem. And and issues. Um, that's a problem. And and yes, AI can find bugs sometimes and it yes, AI can find bugs sometimes and it yes, AI can find bugs sometimes and it can do it better than humans sometimes, can do it better than humans sometimes, can do it better than humans sometimes, but the amount of slop that's generated but the amount of slop that's generated but the amount of slop that's generated around that makes it so hard to find around that makes it so hard to find around that makes it so hard to find those good ones. Okay, so that's kind of those good ones. Okay, so that's kind of those good ones. Okay, so that's kind of the numbers behind why they dropped the the numbers behind why they dropped the the numbers behind why they dropped the bug bunning program, but it's not that's bug bunning program, but it's not that's bug bunning program, but it's not that's not the worst of it. You know, the not the worst of it. You know, the not the worst of it. You know, the authors of these bug reports seem to authors of these bug reports seem to authors of these bug reports seem to have a more entitled attitude. have a more entitled attitude. have a more entitled attitude. these in quotes, helpers try too hard to these in quotes, helpers try too hard to these in quotes, helpers try too hard to twist whatever they find into something twist whatever they find into something twist whatever they find into something horribly bad and a critical horribly bad and a critical horribly bad and a critical vulnerability.

vulnerability. vulnerability. If you tell an AI, I want you to find a If you tell an AI, I want you to find a If you tell an AI, I want you to find a vulnerability. It's going to find vulnerability. It's going to find vulnerability. It's going to find something and you're going to say, I something and you're going to say, I something and you're going to say, I want a critical vulnerability. It's want a critical vulnerability. It's want a critical vulnerability. It's going to find something it's going to going to find something it's going to going to find something it's going to call critical, doesn't necessarily mean call critical, doesn't necessarily mean call critical, doesn't necessarily mean it is. Okay. Now, here this closing line it is. Okay. Now, here this closing line it is. Okay. Now, here this closing line of this uh this section I thought was of this uh this section I thought was of this uh this section I thought was really important, really valuable. the really important, really valuable. the really important, really valuable. the these agentic AI users don't care about these agentic AI users don't care about these agentic AI users don't care about curl. curl. curl. It used to be that and this is you know It used to be that and this is you know It used to be that and this is you know back in the old days um when people back in the old days um when people back in the old days um when people contributed to an open- source project contributed to an open- source project contributed to an open- source project it's because they cared about it but now it's because they cared about it but now it's because they cared about it but now it's less and less about I care about it's less and less about I care about it's less and less about I care about this thing and more and more about what this thing and more and more about what this thing and more and more about what can I get out of this thing. So these can I get out of this thing. So these can I get out of this thing. So these agentic AI users don't care about curl. agentic AI users don't care about curl. agentic AI users don't care about curl. don't care about Daniel or other open don't care about Daniel or other open don't care about Daniel or other open source maintainers. They just want to source maintainers. They just want to source maintainers. They just want to grab quick cash bounties using their grab quick cash bounties using their grab quick cash bounties using their private AI army. It's a get-richqu private AI army. It's a get-richqu private AI army. It's a get-richqu scheme. So that's not helping. That's scheme. So that's not helping. That's scheme. So that's not helping. That's hurting.

hurting. hurting. As we talked about earlier, that As we talked about earlier, that As we talked about earlier, that actually made curl less secure. actually made curl less secure. actually made curl less secure. So it's gotten so bad that GitHub now So it's gotten so bad that GitHub now So it's gotten so bad that GitHub now added a feature. This is a new feature added a feature. This is a new feature added a feature. This is a new feature in GitHub to disable pull requests in GitHub to disable pull requests in GitHub to disable pull requests entirely. the thing that opensource was entirely. the thing that opensource was entirely. the thing that opensource was built upon the idea that you could built upon the idea that you could built upon the idea that you could create a pull request for your for your create a pull request for your for your create a pull request for your for your favorite open source project and say, favorite open source project and say, favorite open source project and say, "Hey, here is my take on a fix for this "Hey, here is my take on a fix for this "Hey, here is my take on a fix for this or here's my take on a new feature." or here's my take on a new feature." or here's my take on a new feature." That's now you're now able to disable That's now you're now able to disable That's now you're now able to disable that entirely for a project. And more that entirely for a project. And more that entirely for a project. And more and more projects are doing just that. and more projects are doing just that. and more projects are doing just that. Which means less and less projects are Which means less and less projects are Which means less and less projects are getting contributions from the getting contributions from the getting contributions from the community. Which means those projects community. Which means those projects community. Which means those projects are getting less and less done from the are getting less and less done from the are getting less and less done from the community. Less and less bugs fixed from community. Less and less bugs fixed from community. Less and less bugs fixed from the community, which means less and less the community, which means less and less the community, which means less and less is going to be benefiting the overall is going to be benefiting the overall is going to be benefiting the overall community. community. community. We're going to have less software with We're going to have less software with We're going to have less software with less features with less security less features with less security less features with less security because these maintainers are having to because these maintainers are having to because these maintainers are having to shut down pull requests. AI slop shut down pull requests. AI slop shut down pull requests. AI slop generation is getting easier, but it's generation is getting easier, but it's generation is getting easier, but it's not getting smarter. Okay. The problem not getting smarter. Okay. The problem not getting smarter. Okay. The problem is that humans who review the code don't is that humans who review the code don't is that humans who review the code don't have infinite resources.

have infinite resources. have infinite resources. Yes, more code's being generated. That Yes, more code's being generated. That Yes, more code's being generated. That doesn't mean it's solving the problem. doesn't mean it's solving the problem. doesn't mean it's solving the problem. It just means that it's actually hiding It just means that it's actually hiding It just means that it's actually hiding the real solutions in a sea of slop. the real solutions in a sea of slop. the real solutions in a sea of slop. Some people have suggested AI could take Some people have suggested AI could take Some people have suggested AI could take over code review, too, but that's not over code review, too, but that's not over code review, too, but that's not the answer. We've talked about before. the answer. We've talked about before. the answer. We've talked about before. It's just not the answer. It's just not the answer. It's just not the answer. All right, All right, All right, this is this is one that's really this is this is one that's really this is this is one that's really frustrating. AI can rewrite open-source frustrating. AI can rewrite open-source frustrating. AI can rewrite open-source code, but it can rewrite the license, code, but it can rewrite the license, code, but it can rewrite the license, too. Now this is an article about one too. Now this is an article about one too. Now this is an article about one specific situation but I just want to specific situation but I just want to specific situation but I just want to point out there are now AI services out point out there are now AI services out point out there are now AI services out there where you can say I want this there where you can say I want this there where you can say I want this library which is restricted from use library which is restricted from use library which is restricted from use based upon it license. I want that based upon it license. I want that based upon it license. I want that library but I want a different license library but I want a different license library but I want a different license on it and they will rewrite the library on it and they will rewrite the library on it and they will rewrite the library with a new license. Now it's not a one with a new license. Now it's not a one with a new license. Now it's not a one for one. They actually use an AI in for one. They actually use an AI in for one. They actually use an AI in theory to reimagine or you know uh redo theory to reimagine or you know uh redo theory to reimagine or you know uh redo the library so that it's the library so that it's the library so that it's it's not a derivative work. It's it's not a derivative work. It's it's not a derivative work. It's actually a a brand new library with a actually a a brand new library with a actually a a brand new library with a new license essentially saying hey we're new license essentially saying hey we're new license essentially saying hey we're going to take all your ideas and just going to take all your ideas and just going to take all your ideas and just have the AI whitewash it. Okay. So, I have the AI whitewash it. Okay. So, I have the AI whitewash it. Okay. So, I have not quoted any of those in here have not quoted any of those in here have not quoted any of those in here because I don't even want to give those because I don't even want to give those because I don't even want to give those services the the time of day. I I don't services the the time of day. I I don't services the the time of day. I I don't want anyone to even know about them. But want anyone to even know about them. But want anyone to even know about them. But just know that's going to danger open just know that's going to danger open just know that's going to danger open source. as we put open source in danger

source. as we put open source in danger source. as we put open source in danger because if if your hard work because if if your hard work because if if your hard work gets just taken and rewritten where it's gets just taken and rewritten where it's gets just taken and rewritten where it's kind of a kid writing a book report, you kind of a kid writing a book report, you kind of a kid writing a book report, you know, from another kid's work where they know, from another kid's work where they know, from another kid's work where they just change the words around and then, just change the words around and then, just change the words around and then, you know, re regurgitate it. If the you know, re regurgitate it. If the you know, re regurgitate it. If the person who did all the work doesn't see person who did all the work doesn't see person who did all the work doesn't see the benefits and sees that instead the benefits and sees that instead the benefits and sees that instead somebody else just takes their ideas and somebody else just takes their ideas and somebody else just takes their ideas and their unit tests and all the rest and their unit tests and all the rest and their unit tests and all the rest and builds another version builds another version builds another version then why would you do any contribution then why would you do any contribution then why would you do any contribution in the first place? Why would you do in the first place? Why would you do in the first place? Why would you do that work in the first place? that work in the first place? that work in the first place? So all of a sudden now we get less and So all of a sudden now we get less and So all of a sudden now we get less and less open- source projects and more and less open- source projects and more and less open- source projects and more and more that close the source down and say more that close the source down and say more that close the source down and say you know what it's not wise to have open you know what it's not wise to have open you know what it's not wise to have open source we should have closed source. source we should have closed source. source we should have closed source. Okay let's talk about specific situation Okay let's talk about specific situation Okay let's talk about specific situation in this article. in this article. in this article. So um So um So um there was a library um that Mark Pilgrim there was a library um that Mark Pilgrim there was a library um that Mark Pilgrim put out in 2026 that was released under put out in 2026 that was released under put out in 2026 that was released under the LGPL the LGPL the LGPL license. So that puts strict limits on license. So that puts strict limits on license. So that puts strict limits on how it could be reused and how it could be reused and how it could be reused and redistributed. So that redistributed. So that redistributed. So that whatever you think of the license, you whatever you think of the license, you whatever you think of the license, you say, "Well, I think they should have a say, "Well, I think they should have a say, "Well, I think they should have a different license on that." That's this different license on that." That's this different license on that." That's this person owns that code and they say, person owns that code and they say, person owns that code and they say, "When I created this, I want it to be "When I created this, I want it to be "When I created this, I want it to be used in this way." That's what licenses used in this way." That's what licenses used in this way." That's what licenses do. They allow you to control how your do. They allow you to control how your do. They allow you to control how your work gets used. You may not like it. You

work gets used. You may not like it. You work gets used. You may not like it. You may want a different way to use that may want a different way to use that may want a different way to use that work. That's not up to you. If you work. That's not up to you. If you work. That's not up to you. If you wanted that, you could have created it wanted that, you could have created it wanted that, you could have created it yourself. yourself. yourself. But instead, they put in the hard work But instead, they put in the hard work But instead, they put in the hard work to create this. They put in all the the to create this. They put in all the the to create this. They put in all the the blood, sweat, and tears to develop the blood, sweat, and tears to develop the blood, sweat, and tears to develop the app, the code, whatever it is, this way. app, the code, whatever it is, this way. app, the code, whatever it is, this way. And they decide this is how I want to And they decide this is how I want to And they decide this is how I want to see it used. see it used. see it used. So Dan Blanchard took over the So Dan Blanchard took over the So Dan Blanchard took over the maintenance of the repository in 2012. maintenance of the repository in 2012. maintenance of the repository in 2012. But what he ended up doing with version But what he ended up doing with version But what he ended up doing with version 7.0 0 was he overhauled it as a groundup 7.0 0 was he overhauled it as a groundup 7.0 0 was he overhauled it as a groundup MIT license rewrite. MIT license rewrite. MIT license rewrite. So he was working on the project which So he was working on the project which So he was working on the project which is LGPL which cannot change. It can't is LGPL which cannot change. It can't is LGPL which cannot change. It can't just be like hey well we're going to just be like hey well we're going to just be like hey well we're going to change the license. That's not how it change the license. That's not how it change the license. That's not how it works. So what he did essentially was in works. So what he did essentially was in works. So what he did essentially was in his way of saying it, he deleted all the his way of saying it, he deleted all the his way of saying it, he deleted all the code and then had AI rebuild it. And so code and then had AI rebuild it. And so code and then had AI rebuild it. And so when it rebuilt it, it was different.

when it rebuilt it, it was different. when it rebuilt it, it was different. And so therefore, it's it's not a in his And so therefore, it's it's not a in his And so therefore, it's it's not a in his words, not a derivative work. Therefore, words, not a derivative work. Therefore, words, not a derivative work. Therefore, it's fine. Now this 7.0 can also be it's fine. Now this 7.0 can also be it's fine. Now this 7.0 can also be under a different license. under a different license. under a different license. Okay. Not everybody's happy with this. Okay. Not everybody's happy with this. Okay. Not everybody's happy with this. In fact, Mark Pilgrim came out to argue In fact, Mark Pilgrim came out to argue In fact, Mark Pilgrim came out to argue the new version amounted to an the new version amounted to an the new version amounted to an illegitimate relicency of Pilgrim's illegitimate relicency of Pilgrim's illegitimate relicency of Pilgrim's original code under a more permissive original code under a more permissive original code under a more permissive MIT license, which among other things MIT license, which among other things MIT license, which among other things allows for it to be used in closed allows for it to be used in closed allows for it to be used in closed source closed source projects. source closed source projects. source closed source projects. Now, the claim is that it's a complete Now, the claim is that it's a complete Now, the claim is that it's a complete rewrite is irrelevant since they had rewrite is irrelevant since they had rewrite is irrelevant since they had ample exposure to the original license ample exposure to the original license ample exposure to the original license code. code. code. and he added that adding a fancy code and he added that adding a fancy code and he added that adding a fancy code generator into a mix does not somehow generator into a mix does not somehow generator into a mix does not somehow grant them any additional rights. So the grant them any additional rights. So the grant them any additional rights. So the argument here is on on the side of was argument here is on on the side of was argument here is on on the side of was it David uh Dan on Dan's side what he's it David uh Dan on Dan's side what he's it David uh Dan on Dan's side what he's saying is hey we did what's called a saying is hey we did what's called a saying is hey we did what's called a clean room built which is where uh the clean room built which is where uh the clean room built which is where uh the AI we say hey don't don't look at the AI we say hey don't don't look at the AI we say hey don't don't look at the current version but I want you to create current version but I want you to create current version but I want you to create a new one with the the specification a new one with the the specification a new one with the the specification which where did that specification come which where did that specification come which where did that specification come from from from how did you know what you know what how did you know what you know what how did you know what you know what endpoints to use or what um you ways of endpoints to use or what um you ways of endpoints to use or what um you ways of doing things, what what it should do, doing things, what what it should do, doing things, what what it should do, that all came from the previous version.

that all came from the previous version. that all came from the previous version. Even if you're not showing source code, Even if you're not showing source code, Even if you're not showing source code, you're showing a a structure, you're you're showing a a structure, you're you're showing a a structure, you're showing a a way of doing things, even if showing a a way of doing things, even if showing a a way of doing things, even if you're doing it from memory, right? But you're doing it from memory, right? But you're doing it from memory, right? But then he used an AI to rebuild all that then he used an AI to rebuild all that then he used an AI to rebuild all that code. What was that AI trained on? One code. What was that AI trained on? One code. What was that AI trained on? One of the big things I was trained on was of the big things I was trained on was of the big things I was trained on was open- source projects. And fortunately open- source projects. And fortunately open- source projects. And fortunately that includes LPGL that includes LPGL that includes LPGL licensed open-source projects which licensed open-source projects which licensed open-source projects which means it knew about this project. It had means it knew about this project. It had means it knew about this project. It had seen all the code of this project. So seen all the code of this project. So seen all the code of this project. So when it recreates this project, it knows when it recreates this project, it knows when it recreates this project, it knows what it's doing and it's based upon what it's doing and it's based upon what it's doing and it's based upon deriving its work off of this project. deriving its work off of this project. deriving its work off of this project. So So So it says, you know, it had no access to it says, you know, it had no access to it says, you know, it had no access to the source old source tree and the source old source tree and the source old source tree and explicitly instructed Claude not to base explicitly instructed Claude not to base explicitly instructed Claude not to base anything on LPGL or GPL license code.

anything on LPGL or GPL license code. anything on LPGL or GPL license code. What you tell an AI and what it does What you tell an AI and what it does What you tell an AI and what it does first of all is not always going to first of all is not always going to first of all is not always going to align. But second of all again in the align. But second of all again in the align. But second of all again in the training set uh for one claude training set uh for one claude training set uh for one claude explicitly relied on some metadata files explicitly relied on some metadata files explicitly relied on some metadata files from previous versions of chardet which from previous versions of chardet which from previous versions of chardet which is the the library. So first of all it's is the the library. So first of all it's is the the library. So first of all it's rely on metadata from that library and rely on metadata from that library and rely on metadata from that library and then claude's models are trained on then claude's models are trained on then claude's models are trained on reams of data pulled from the public reams of data pulled from the public reams of data pulled from the public internet internet internet which means it's overwhel over which means it's overwhel over which means it's overwhel over overwhelmingly likely that claude had overwhelmingly likely that claude had overwhelmingly likely that claude had ingested the open- source code of the ingested the open- source code of the ingested the open- source code of the previous chartet versions in his previous chartet versions in his previous chartet versions in his training. So this is not a we created training. So this is not a we created training. So this is not a we created something that is you know very very something that is you know very very something that is you know very very similar but definitely not the same similar but definitely not the same similar but definitely not the same thing. Um is almost certainly based upon thing. Um is almost certainly based upon thing. Um is almost certainly based upon the previous versions of this same the previous versions of this same the previous versions of this same library. library. library. So now here's a really tricky bit that So now here's a really tricky bit that So now here's a really tricky bit that kind of adds another wrinkle to the mix. kind of adds another wrinkle to the mix. kind of adds another wrinkle to the mix. Um using AI to create new code from Um using AI to create new code from Um using AI to create new code from whole cloth could also create its own whole cloth could also create its own whole cloth could also create its own legal legal complications going forward.

legal legal complications going forward. legal legal complications going forward. Courts have always said the AI cannot be Courts have always said the AI cannot be Courts have always said the AI cannot be the author on a patent or the copyright the author on a patent or the copyright the author on a patent or the copyright holder of a piece of art. holder of a piece of art. holder of a piece of art. It's more complicated than saying, It's more complicated than saying, It's more complicated than saying, "Well, I'll just put my name on it." "Well, I'll just put my name on it." "Well, I'll just put my name on it." Because that's not how this works. If Because that's not how this works. If Because that's not how this works. If you use AI to generate an image, you you use AI to generate an image, you you use AI to generate an image, you cannot copyright that image. You can't cannot copyright that image. You can't cannot copyright that image. You can't say, "Well, I created." And courts have say, "Well, I created." And courts have say, "Well, I created." And courts have already upheld this. So now, what does already upheld this. So now, what does already upheld this. So now, what does this mean for software? We're not sure this mean for software? We're not sure this mean for software? We're not sure yet because no court has yet ruled on it yet because no court has yet ruled on it yet because no court has yet ruled on it as of now, but that's coming. And what as of now, but that's coming. And what as of now, but that's coming. And what does that mean for the code? That's does that mean for the code? That's does that mean for the code? That's that's another wrinkle that actually the that's another wrinkle that actually the that's another wrinkle that actually the license for this project, it was MIT, license for this project, it was MIT, license for this project, it was MIT, has now changed to I think uh zero BSD has now changed to I think uh zero BSD has now changed to I think uh zero BSD or something like that. Basically, it's or something like that. Basically, it's or something like that. Basically, it's an ASIS license because uh what Dan said an ASIS license because uh what Dan said an ASIS license because uh what Dan said was, well, I don't know if I can even was, well, I don't know if I can even was, well, I don't know if I can even put an MIT license on it. Therefore, put an MIT license on it. Therefore, put an MIT license on it. Therefore, it's gonna say it's ASIS license, which it's gonna say it's ASIS license, which it's gonna say it's ASIS license, which also also also causes problems because now it's it's causes problems because now it's it's causes problems because now it's it's ASIS, which is not quite the same thing, ASIS, which is not quite the same thing, ASIS, which is not quite the same thing, and companies have an even harder time and companies have an even harder time and companies have an even harder time relying on ASIS licenses.

relying on ASIS licenses. relying on ASIS licenses. So, whatever the outcomes here, the So, whatever the outcomes here, the So, whatever the outcomes here, the practical impact of being able to use AI practical impact of being able to use AI practical impact of being able to use AI to quickly rewrite and relic many to quickly rewrite and relic many to quickly rewrite and relic many open-source projects is likely to have open-source projects is likely to have open-source projects is likely to have huge knock-on effects throughout the huge knock-on effects throughout the huge knock-on effects throughout the community. I absolutely agree. This is a community. I absolutely agree. This is a community. I absolutely agree. This is a problem. This is going to endanger every problem. This is going to endanger every problem. This is going to endanger every open-source product out there. Like I open-source product out there. Like I open-source product out there. Like I said earlier, you can actually pay a said earlier, you can actually pay a said earlier, you can actually pay a service to have an AI rewrite an service to have an AI rewrite an service to have an AI rewrite an open-source product of your choosing open-source product of your choosing open-source product of your choosing either in a different language or in the either in a different language or in the either in a different language or in the same language with a different license. same language with a different license. same language with a different license. That's a problem because that is going That's a problem because that is going That's a problem because that is going to discourage. It's going to tamp down to discourage. It's going to tamp down to discourage. It's going to tamp down the number of open- source projects that the number of open- source projects that the number of open- source projects that are even out there because essentially are even out there because essentially are even out there because essentially you're bypassing everything that that you're bypassing everything that that you're bypassing everything that that author author author did this for. Like yes, the author did this for. Like yes, the author did this for. Like yes, the author wanted that software out in the world, wanted that software out in the world, wanted that software out in the world, but now it's a different a different uh but now it's a different a different uh but now it's a different a different uh package. It's a different, you know, package. It's a different, you know, package. It's a different, you know, project, which means they can't fix the project, which means they can't fix the project, which means they can't fix the problems that come up. yet their name problems that come up. yet their name problems that come up. yet their name might be associated with that and they might be associated with that and they might be associated with that and they can't fix the the issues and so you're can't fix the the issues and so you're can't fix the the issues and so you're relying on software that might be really relying on software that might be really relying on software that might be really buggy or might have vulnerabilities you buggy or might have vulnerabilities you buggy or might have vulnerabilities you don't know about and you took all their don't know about and you took all their don't know about and you took all their work and took their name off of it.

work and took their name off of it. work and took their name off of it. You know, that's that's like seeing a You know, that's that's like seeing a You know, that's that's like seeing a painting that you like and having it painting that you like and having it painting that you like and having it just repainted for you and then putting just repainted for you and then putting just repainted for you and then putting your name on it. That's not really cool, your name on it. That's not really cool, your name on it. That's not really cool, right? That's that's kind of stealing. right? That's that's kind of stealing. right? That's that's kind of stealing. So, these are all problems. They're So, these are all problems. They're So, these are all problems. They're adding up to adding up to adding up to open- source maintainers are under fire. open- source maintainers are under fire. open- source maintainers are under fire. They are less compensated. They are less compensated. They are less compensated. They're more overworked. They're more overworked. They're more overworked. They're leaving in larger numbers. They're leaving in larger numbers. They're leaving in larger numbers. They're putting pay walls around things They're putting pay walls around things They're putting pay walls around things that used to be free and available. that used to be free and available. that used to be free and available. They're getting less contributions for They're getting less contributions for They're getting less contributions for people. people. people. So, overall, AI is putting a real strain So, overall, AI is putting a real strain So, overall, AI is putting a real strain on open source. So the question is what on open source. So the question is what on open source. So the question is what can we do? Because can we do? Because can we do? Because AI has value. AI can do good things. And AI has value. AI can do good things. And AI has value. AI can do good things. And a lot of what we talked about in this a lot of what we talked about in this a lot of what we talked about in this video is about people misusing AI. video is about people misusing AI. video is about people misusing AI. And yes, I I hear an awful lot of times, And yes, I I hear an awful lot of times, And yes, I I hear an awful lot of times, well, but that's not me. I don't misuse well, but that's not me. I don't misuse well, but that's not me. I don't misuse AI. Well, that's great. But the reality AI. Well, that's great. But the reality AI. Well, that's great. But the reality is there's a lot of people out there is there's a lot of people out there is there's a lot of people out there missing AI. And even the people who are missing AI. And even the people who are missing AI. And even the people who are well-intentioned well-intentioned well-intentioned can misuse AI and can hurt open source.

can misuse AI and can hurt open source. can misuse AI and can hurt open source. So what can we do? Well, first of So what can we do? Well, first of So what can we do? Well, first of thinking about thinking about thinking about creating open-source software, you need creating open-source software, you need creating open-source software, you need to evaluate if open source is right for to evaluate if open source is right for to evaluate if open source is right for you. And this is this is not the rosy you. And this is this is not the rosy you. And this is this is not the rosy thing that you know helps open source. thing that you know helps open source. thing that you know helps open source. But I want to be real with you. When you But I want to be real with you. When you But I want to be real with you. When you are building something that you love, are building something that you love, are building something that you love, you might need to think about is open you might need to think about is open you might need to think about is open source the right call here or do I keep source the right call here or do I keep source the right call here or do I keep it closed source so I can protect myself it closed source so I can protect myself it closed source so I can protect myself a little bit. a little bit. a little bit. That might be one things that you need That might be one things that you need That might be one things that you need to think about. On the other side of it, to think about. On the other side of it, to think about. On the other side of it, if you are creating software and you if you are creating software and you if you are creating software and you want to bring on a dependency, you want to bring on a dependency, you want to bring on a dependency, you should think about is open-source the should think about is open-source the should think about is open-source the right solution here for me? And if it right solution here for me? And if it right solution here for me? And if it is, how do you make sure that you choose is, how do you make sure that you choose is, how do you make sure that you choose the right one? You should not choose a the right one? You should not choose a the right one? You should not choose a ton of dependencies. That's just in ton of dependencies. That's just in ton of dependencies. That's just in general good software knowledge, right?

general good software knowledge, right? general good software knowledge, right? You should try and limit your You should try and limit your You should try and limit your dependencies and try and choose the dependencies and try and choose the dependencies and try and choose the right open-source projects. That might right open-source projects. That might right open-source projects. That might mean choosing the projects that are mean choosing the projects that are mean choosing the projects that are being the most strict about who can being the most strict about who can being the most strict about who can contribute and how contributions get contribute and how contributions get contribute and how contributions get made. It's important that you choose the made. It's important that you choose the made. It's important that you choose the right dependencies because if open- right dependencies because if open- right dependencies because if open- source is under fire and you depend on source is under fire and you depend on source is under fire and you depend on open source, open source, open source, you are depending on something that may you are depending on something that may you are depending on something that may go away. go away. go away. That's a real danger for your for your That's a real danger for your for your That's a real danger for your for your closed source production application. closed source production application. closed source production application. Number two, Number two, Number two, if you are going to create an if you are going to create an if you are going to create an open-source project, you should know open-source project, you should know open-source project, you should know your monetization strategy right up your monetization strategy right up your monetization strategy right up front. So when you're thinking about how front. So when you're thinking about how front. So when you're thinking about how do I support myself? How do I support do I support myself? How do I support do I support myself? How do I support this project? And I am a very firm this project? And I am a very firm this project? And I am a very firm believer in you shouldn't say it's a believer in you shouldn't say it's a believer in you shouldn't say it's a side project. It's a love project. You side project. It's a love project. You side project. It's a love project. You know, it's it's something I'm just going know, it's it's something I'm just going know, it's it's something I'm just going to, you know, do for fun. Then you to, you know, do for fun. Then you to, you know, do for fun. Then you should really tell people upfront this should really tell people upfront this should really tell people upfront this is not a long-term supported thing. you is not a long-term supported thing. you is not a long-term supported thing. you should really tell people upfront that should really tell people upfront that should really tell people upfront that this is not something that can scale this is not something that can scale this is not something that can scale because if you build something that then because if you build something that then because if you build something that then the world latches on to it and says yes the world latches on to it and says yes the world latches on to it and says yes we want this and you get lots of stars we want this and you get lots of stars we want this and you get lots of stars on GitHub that's great um what it on GitHub that's great um what it on GitHub that's great um what it actually means is you're going to have a actually means is you're going to have a actually means is you're going to have a lot more work to do and that's going to lot more work to do and that's going to lot more work to do and that's going to create burnout real quick and that puts create burnout real quick and that puts create burnout real quick and that puts everybody in danger. So you should know everybody in danger. So you should know everybody in danger. So you should know upfront how do you get paid for this and

upfront how do you get paid for this and upfront how do you get paid for this and I think that you should put really I think that you should put really I think that you should put really clearly upfront that there are guard clearly upfront that there are guard clearly upfront that there are guard rails you should put around it so that rails you should put around it so that rails you should put around it so that you make sure that you protect yourself you make sure that you protect yourself you make sure that you protect yourself so it's not just well if they go at so it's not just well if they go at so it's not just well if they go at documentation like we saw at the documentation like we saw at the documentation like we saw at the beginning with Tailwind that's not good beginning with Tailwind that's not good beginning with Tailwind that's not good enough anymore if you know if you're enough anymore if you know if you're enough anymore if you know if you're relying on people seeing your code with relying on people seeing your code with relying on people seeing your code with your name and your contribution button your name and your contribution button your name and your contribution button it's probably not the right call it's probably not the right call it's probably not the right call anymore. So, you need to think about how anymore. So, you need to think about how anymore. So, you need to think about how do I make money from this? How do I do I make money from this? How do I do I make money from this? How do I ensure that this project can scale if ensure that this project can scale if ensure that this project can scale if it's going to scale? How can I make sure it's going to scale? How can I make sure it's going to scale? How can I make sure that if if kind of the best slash worst that if if kind of the best slash worst that if if kind of the best slash worst thing happens and the internet loves thing happens and the internet loves thing happens and the internet loves this project and it goes viral, how do I this project and it goes viral, how do I this project and it goes viral, how do I make sure that it doesn't then crash and make sure that it doesn't then crash and make sure that it doesn't then crash and burn because I'm just overwhelmed burn because I'm just overwhelmed burn because I'm just overwhelmed and you know if it starts taking you and you know if it starts taking you and you know if it starts taking you know 5 10 15 20 hours a week from you know 5 10 15 20 hours a week from you know 5 10 15 20 hours a week from you how do you ensure that you can afford to how do you ensure that you can afford to how do you ensure that you can afford to do that? So know your monetization do that? So know your monetization do that? So know your monetization strategy upfront. I'm going to put one strategy upfront. I'm going to put one strategy upfront. I'm going to put one in here that is a specific strategy. Um in here that is a specific strategy. Um in here that is a specific strategy. Um I think if you're an open source I think if you're an open source I think if you're an open source maintainer, you should consider the maintainer, you should consider the maintainer, you should consider the open- source maintainer fee. So this is open- source maintainer fee. So this is open- source maintainer fee. So this is a an open source maintainer decided a an open source maintainer decided a an open source maintainer decided let's fix this problem. Um and say you let's fix this problem. Um and say you let's fix this problem. Um and say you know what um here is a way to do this.

know what um here is a way to do this. know what um here is a way to do this. And essentially this open source And essentially this open source And essentially this open source maintainer feed does it says you know maintainer feed does it says you know maintainer feed does it says you know what the source is still open. If you what the source is still open. If you what the source is still open. If you want to see the source, if you want to want to see the source, if you want to want to see the source, if you want to compile from source, if you want to see compile from source, if you want to see compile from source, if you want to see how the code works, all of that is still how the code works, all of that is still how the code works, all of that is still free. However, if you want anything free. However, if you want anything free. However, if you want anything around the source, which we have we have around the source, which we have we have around the source, which we have we have tied too closely together to the source, tied too closely together to the source, tied too closely together to the source, right? That is if you want to add issues right? That is if you want to add issues right? That is if you want to add issues or feature requests, if you want to do or feature requests, if you want to do or feature requests, if you want to do pull requests, if you want to have the pull requests, if you want to have the pull requests, if you want to have the the built compiled code that's in, you the built compiled code that's in, you the built compiled code that's in, you know, for your platform, etc. If you know, for your platform, etc. If you know, for your platform, etc. If you want that, all of that stuff is extra. want that, all of that stuff is extra. want that, all of that stuff is extra. It's not part of the open-source It's not part of the open-source It's not part of the open-source project. That is what you pay for. project. That is what you pay for. project. That is what you pay for. So that's the idea of the open source So that's the idea of the open source So that's the idea of the open source maintainer fee. It doesn't have to be a maintainer fee. It doesn't have to be a maintainer fee. It doesn't have to be a lot, but what you say is if you want any lot, but what you say is if you want any lot, but what you say is if you want any of this extra stuff, you have to pay for of this extra stuff, you have to pay for of this extra stuff, you have to pay for it. If you want the source code, cool, it. If you want the source code, cool, it. If you want the source code, cool, go for the source code.

go for the source code. go for the source code. That is a way that can improve the odds That is a way that can improve the odds That is a way that can improve the odds that you're you have a monetization that you're you have a monetization that you're you have a monetization strategy that will work. Now, it will strategy that will work. Now, it will strategy that will work. Now, it will not stop an AI from stealing your code. not stop an AI from stealing your code. not stop an AI from stealing your code. That some of that we have to wait for That some of that we have to wait for That some of that we have to wait for the courts to figure out. But what it the courts to figure out. But what it the courts to figure out. But what it will do is it will prevent abuse on your will do is it will prevent abuse on your will do is it will prevent abuse on your site because you're not going to get site because you're not going to get site because you're not going to get spammed by AI contributors unless spammed by AI contributors unless spammed by AI contributors unless they're paying customers. Now, you still they're paying customers. Now, you still they're paying customers. Now, you still may get spammed once they are paying may get spammed once they are paying may get spammed once they are paying customers and that you have to put into customers and that you have to put into customers and that you have to put into some restriction, but you'll get less of some restriction, but you'll get less of some restriction, but you'll get less of that than just being open on the that than just being open on the that than just being open on the internet for everything, not just the internet for everything, not just the internet for everything, not just the source. source. source. Number four, support the open-source Number four, support the open-source Number four, support the open-source projects that you depend on. projects that you depend on. projects that you depend on. Like I said at the beginning of this Like I said at the beginning of this Like I said at the beginning of this video, we all depend video, we all depend video, we all depend we we would not exist in the software we we would not exist in the software we we would not exist in the software world without opensource code. Without world without opensource code. Without world without opensource code. Without people who are doing this for free, people who are doing this for free, people who are doing this for free, without people who are putting in the without people who are putting in the without people who are putting in the time and effort to maintain what we time and effort to maintain what we time and effort to maintain what we have, we would not have software.

have, we would not have software. have, we would not have software. So So So when you say we depend on these 10 when you say we depend on these 10 when you say we depend on these 10 projects for our production app, projects for our production app, projects for our production app, you should support those 10 projects. you should support those 10 projects. you should support those 10 projects. Those 10 projects are supporting you. Those 10 projects are supporting you. Those 10 projects are supporting you. Imag if you don't want to do this. I'd Imag if you don't want to do this. I'd Imag if you don't want to do this. I'd encourage you figure out what it would encourage you figure out what it would encourage you figure out what it would take to remove all those open-source take to remove all those open-source take to remove all those open-source projects from your software. projects from your software. projects from your software. Imagine the expense and then say, you Imagine the expense and then say, you Imagine the expense and then say, you know what, I'm going to find it cheaper know what, I'm going to find it cheaper know what, I'm going to find it cheaper to support them than to remove all of to support them than to remove all of to support them than to remove all of this. And if you don't, well then you this. And if you don't, well then you this. And if you don't, well then you probably don't have great dependencies. probably don't have great dependencies. probably don't have great dependencies. You should probably replace them. But You should probably replace them. But You should probably replace them. But the reason why it's even more imperative the reason why it's even more imperative the reason why it's even more imperative now is because of the fact that now is because of the fact that now is because of the fact that open-source is under fire. Open source open-source is under fire. Open source open-source is under fire. Open source maintainers are being beaten down. maintainers are being beaten down. maintainers are being beaten down. They're being worn out. They're They're being worn out. They're They're being worn out. They're experiencing burnout. They're leaving experiencing burnout. They're leaving experiencing burnout. They're leaving the field. They are shutting down their the field. They are shutting down their the field. They are shutting down their projects. They're paying their projects.

projects. They're paying their projects. projects. They're paying their projects. They're walking away from their projects They're walking away from their projects They're walking away from their projects because of what AI and people abusing AI because of what AI and people abusing AI because of what AI and people abusing AI have done to them. have done to them. have done to them. If you don't support them, then they If you don't support them, then they If you don't support them, then they will go away. Those projects will. And I will go away. Those projects will. And I will go away. Those projects will. And I just asked you, why don't you just just asked you, why don't you just just asked you, why don't you just remove those things? Well, imagine a remove those things? Well, imagine a remove those things? Well, imagine a time when you have to, when you don't time when you have to, when you don't time when you have to, when you don't have an option, when you know your have an option, when you know your have an option, when you know your back's against the wall, you've been back's against the wall, you've been back's against the wall, you've been working on a feature, but now all of a working on a feature, but now all of a working on a feature, but now all of a sudden you realize that dependency your sudden you realize that dependency your sudden you realize that dependency your entire product depends on is going away. entire product depends on is going away. entire product depends on is going away. You can't just say, "Well, we'll get to You can't just say, "Well, we'll get to You can't just say, "Well, we'll get to that in a year. We'll get that in two that in a year. We'll get that in two that in a year. We'll get that in two years." If it's gone, you might be in a years." If it's gone, you might be in a years." If it's gone, you might be in a really bad spot. really bad spot. really bad spot. take some preventive action. Support the take some preventive action. Support the take some preventive action. Support the things that you depend on. things that you depend on. things that you depend on. Okay, so that's it for open-source Okay, so that's it for open-source Okay, so that's it for open-source software and how AI hurts open source software and how AI hurts open source software and how AI hurts open source software. Again, just to be clear, I get software. Again, just to be clear, I get software. Again, just to be clear, I get this comment a lot. Um, this is not an this comment a lot. Um, this is not an this comment a lot. Um, this is not an anti- AI rant. This is not about AI is anti- AI rant. This is not about AI is anti- AI rant. This is not about AI is horrible. And again, a lot of this has horrible. And again, a lot of this has horrible. And again, a lot of this has to do with people abusing AI and abusing to do with people abusing AI and abusing to do with people abusing AI and abusing what it does. But what it does. But what it does. But there's a lot of people doing that and there's a lot of people doing that and there's a lot of people doing that and we need to make sure that we put some we need to make sure that we put some we need to make sure that we put some guard rails in place to stop the people guard rails in place to stop the people guard rails in place to stop the people who are abusing it. We need to make sure who are abusing it. We need to make sure who are abusing it. We need to make sure that we are not abusing ourselves. We

that we are not abusing ourselves. We that we are not abusing ourselves. We make sure that we are making wise make sure that we are making wise make sure that we are making wise decisions and supporting the things that decisions and supporting the things that decisions and supporting the things that support us. All right. So, thanks for support us. All right. So, thanks for support us. All right. So, thanks for watching and as always, I am Tim Corey.