I got DMCA'd by Anthropic (not a joke)

Yesterday, I published a video about Anthropic's leak of the Claude Code source. And this morning, I woke up to a DMCA notice. I'm not joking, by the way. This is legit. No April Fools here. I woke up to a message from my assistant telling me I was DMCA by Enthropic, and I just didn't believe it initially. So, I grabbed my laptop, went and looked, and sure indeed, I did. This is really, really bad. Want to make a few things clear before we start. First off, this DMCA is on GitHub. It has nothing to do with my YouTube video. I have received no notice whatsoever for the coverage that I've made thus far. The much more interesting detail here, however, is that you probably would assume the DMCA is on the fork of the source code that we were using. You know, the build that Ben and I had, so we could actually play with it. No, that is not what was DMCA. The repo that was DMCA was my fork of the official Claude Code repo. You know, the one that doesn't have their source. It just has their skills and some markdown files. The repo that got DMCA is the one I filed this PR with where I changed one word in one file. Yes, really. A one line of code change got me my first ever DMCA strike on GitHub. This is absolute absurdity, but I actually don't think I blame Anthropic that much. This video is going to be full of some very fun twists. I came in heated and pissed. And after doing more looking into it, talking to others, and figuring things out, I realized this is all much more complex than you might think. But do you know what I love more than a good plot twist? Today's sponsor. It's kind of crazy how AI has made so much stuff easier, but O is still obnoxious. Getting all of the details right, setting things up so the off flows work, and then trying to get preview builds working, god forbid you have to add billing to your service, all of this stuff is obnoxious no matter how good the AI gets. Unless you're using today's sponsor, Clerk. Whenever I'm not using Clerk, I honestly miss it. Their DX is unmatched. Not only do they have all of the components and libraries you would need for any modern tool you might find yourself using, they make it incredibly easy to integrate by literally just copying a prompt, going into your project, hitting paste, and it's implemented. I pushed this to the limit a little bit. I had a project I was working on that y'all might have seen. It's called T3 Canvas. We ultimately folded this into the T3 chat product. When I was working on it, I did a lot of experimentation. I had both a web and mobile app and they were working great, but I had no O experience because they were built just as an experiment for me. As more people wanted it, I wanted to add O. So, I decided to push the limits of Opus 4.5, not even 4.6. And I asked to add O to my monor repo using Clerk. I you not got it first try across both web and mobile with fully synced experiences across both platforms with all of the integrations you would expect from Clerk, including their super helpful user component that lets you manage your account and do all of the things a user would want to do. This also includes organizations for companies that want to have everybody on their team in a single group that's just built into Clerk. And yes, they even give you a component you can mount in your React app to do it. And if I was to tell you how much I love the billing product, I would need a lot more than the ad time here. So keep an eye out for the next one because Clerk has some really cool stuff cooking. Add off that agents love at soy.link/clerk. So as I mentioned at the start, I got DMCA by Anthropic. The thing I didn't mention about this DMCA is the time it happened at. actually happened while I was filming my video, but since I wasn't checking my email, I didn't see it when it happened. This is the official DMCA notice that Anthropic filed with GitHub because GitHub actually takes the time to publish all of them. It's also worth noting this is not the first time this has happened. Before we go in, we should probably understand the DMCA, the Digital Millennium Copyright Act. The goal of the DMCA was simple. Try to keep platforms from dying when there are different copyright holders trying to get their content taken off of them. So, if a random user uploads Avengers Endgame to YouTube, the owners of Avengers Endgame can't sue YouTube. They have to sue the person who uploaded it. And part of how this works is the safe harbor, which is that companies that own the rights to content can report infringing content to the platform with the expectation that the platform takes it down without the platform having to do their own due diligence. As it says here, the Title 17 extension limited the liability of providers of online services for copyright infringement by their users. It's also worth noting that the DMCA is often used for abusive purposes. If you have copyright for specific things that might seem inconsequential, you can use that to get stuff taken down. Reverse engineering is a protected right for us as Americans, which is why making an emulator is fully legal. But with the Wii, there was an interesting problem. Nintendo hard-coded a specific string that is used to decrypt games, and it's a very short string. It's like 15 or so characters. That string is needed to actually play the games, to get them off the disc and into memory. And without that, you can't play Wii games. So emulators like Dolphin have to include that or expect you to include it. And as such, whenever a version of Dolphin that has that included is distributed, Nintendo can go after them for copyright for including that 15 character string. There are things you can abuse with the DMCA, but thankfully we have protections as well. It is important with DMCA's that you're able to respond and file a counter notice. If I don't believe my content is infringing, is my right to inform the accusatory party that I believe my content is within fair use or isn't their content or within their copyright? And then we can fight it in the courts or they can say, "Okay, fine. I didn't know you were going to push back and drop it." And I had every intention of doing that here. And that is why I blasted them on Twitter this morning for how absurd this DMCA request was because again, the request wasn't for something that was copywritten. It wasn't for the source code that leaked yesterday. The DMCA was for my fork where I removed one word from a skill. And that makes sense when you look at the DMCA request that Anthropic filed because the notice was against a network of 8.1,000 repositories inclusive of the parent repo. That's a lot of repositories. I know a lot of people were playing with the repos that forked the illegal source that leaked of Cloud Code, but I don't think there were 8.1,000 of them. That sounds a little high, which means that what happened here was their attempt to flag the right repos went rogue and hit way more than it was supposed to. GitHub only lists up to a 100 of the repos that were hit in these DMCA requests. So, we only see a 100 of the 8,100 or so. So, it's hard to know how many other things were hit. But, I'm just going to scroll and click a random one. That one is DMCA still. Let's click another random one. DMCA. Interesting. According to chat, GitHub does list more than a hundred, but GitHub didn't include them all in the notice. Hard to know. I am not familiar with the details here. But this is where things start to break down a bit because my repo isn't gone anymore. My fork is still here. It's a little behind main, but again, all I changed here is that one skill file. There's nothing illegal here. There's nothing valid to do a DMCA against here. That was an erroneous and arguably illegal Digital Millennium Copyright Act enforcement. And as many have observed on both Twitter and in my chat now, it is very much illegal to use the DMCA on something that never broke it. That's an important part of the law. That's why it works. It is illegal for a company to try and enforce a DMCA against somebody who didn't violate their copyright. As such, my process could have been give them a notice saying, "I didn't violate this. You have 2 weeks to revert your strike against me. And if you don't, I will sue and then go sue them. Thankfully, I don't have to do that, though. And here is where the tide's about to flip. This morning, they filed a retraction of the DMCA notice. They retracted the notice to all repos except for the one that was mirroring the Cloud Code leak and the 96 forks of that specific leak. This is particularly interesting because not only is my repo backup, but a lot of others I didn't think would be R2. For example, this project that is rewriting claude code in Rust using the original source. This is now a derivative work that is protected by copyright. And Anthropic can't really do anything about this. And it hit 100k stars in under a day, which makes it the fastest growing GitHub repo of all time. This one's safe. And you know, Anthropic wants to take this down, but they know better. In fact, it's my understanding that Anthropic never even tried to take this one down because it was never linked to the official Claude Code repo. So, what happened here? Here's where I have to do a thing I really really don't like. I have to defend Enthropic a bit. Thoric replied to my post saying that this was a communication mistake and he linked to the retraction. My repo should be reinstated, but he wasn't sure what the process is. There have been additional follow-ups from Anthropic, including these two from Boris. This was not intentional. We've been working with GitHub to fix it. It should be better now. This was also not intentional. We are debugging what happened on the GitHub side. No one Anthropic DMCAD these forks. So, at some point there was a miscommunication about which forks needed to be taken down. And the network of 8.1,000 repositories was the forks of the official Anthropic repo, not the forks of this arguably illegal hosted version of the source code leak that Enthropic caused. And now I have to go into defense mode a bit. I don't know what led to this miscommunication. I don't know if the original repo was included or if they showed all of the PRs that were being filed on the original repo as part of their informant and like the brief that they gave to GitHub. I don't know if GitHub just outright up. I don't know if GitHub sends the list of repos that they're going to take action on to Anthropic and their lawyers before doing it. I don't know where the communication breakdown happened here. As really transparent as the Claudo team has been over this the last day or so, they haven't been about the details and I'm assuming that's the lawyers preventing them from doing that. This does feel a little bit fingerpointy, but it also seems entirely realistic that GitHub doesn't have a super formal DMCA process. It's lawyers on one side and lawyers on another, and nobody who understands what's actually going on communicating properly on either side. And since this is one of the biggest cases of a source code leak in history, nobody probably second-guessed that 8.1,000 number. And as a result, this illegal action was taken against me and many others GitHub repositories. This does not change the legality of what happened. I want to be very clear about that. Enforcing a false DMCA strike is against the law in the United States. It is not a law that has been prosecuted particularly heavily sadly, but I wish it was. This is something that should be punished in the courts. If anyone takes the time to spin up a proper class action lawsuit, I will gladly be involved. But that all said, I don't even know if it is Anthropic's fault. It is possible they sent over these specific pieces of infringing content and GitHub screwed it up. It's possible that GitHub misunderstood the request and enforced it incorrectly. And it's possible Anthropic is just lying here. All of these things are possible. But the fact that repos like the claw code one I just showed are still up, the fact that mine went back up so quickly, like this retraction went up as fast as it did, and that I'm getting engagement from so many employees at Anthropic trying to clarify this suggests that as scummy as it is, they might not have done something quite as bad as it seems here. I absolutely don't think Anthropic would intentionally DMCA me because one of the worst things that could ever happen to them is the first half of this video without the second half right now. I could have absolutely wrecked their reputation with this and they're still going to take a hit and that hit is absolutely deserved because they shouldn't be doing this in the first place. Also, to be clear, if the DMCA request was incorrectly enforced by GitHub, that is unlikely to be illegal. But if the DMC request itself was requesting the takedown of things that shouldn't be taken down, that is illegal. But I also want to defend Anthropic a little bit more. In my video about the leak yesterday, I closed by describing how I think Anthropic should handle this. The biggest piece I pushed here was be human about it. Specifically, take advantage of the opportunity. Let humans come in, talk about things, reply, make jokes, share the work they've done, talk about the things they built and how proud they are of them. and just take this opportunity to let the humans who built Claude Code talk about Claude Code instead of getting mad at everybody for having access to the source. Let the people who built the cool things be excited about the cool things they built. Excitement will always win out over almost any other force in social media. And if you can let excited people share the things they built, that will ultimately win. And to their credit, they absolutely took my word here and have been doing that. from Thoric leaking the buddy feature and tagging the employee who built a lot of it, Alistair, to somebody asking what happens if the snail in buddy mode catches you. To which Thoric replies, source code gets leaked, which is genuinely funny. Like, this is a good post. This is this is exactly as silly as it is, exactly what I'm asking for here. Or Boris's reply to Wes sharing all of the different spinner verbs, specifically answering who made these. He said he came up with the initial list, but a bunch of others contributed words as well. Melissa has gone through many iterations. Somebody asked for gooning to be added. He replied to this too, saying, "Someone else asked for this, but I think it's weird. You can add it to your thinking words via settings if you like. Just ask Claude." Bunny reply. But also, more importantly, how Boris is responding to claims about the leak. As I mentioned in my previous video, some believe that Bun's issue with source mappings being served with Bun Serve might have been the cause, which it wasn't. And Boris came in and confirmed it was not related to Bun. Jane asked, "Is this developer still breathing?" the one who made the error because Boris referred to it as developer error. And Boris had a great reply here. As a team, the important thing is to recognize it's never an individual's fault. It's the process, the culture, or the infrastructure. In this case, there was a manual deploy step that should have been better automated. Our team has made a few improvements to the automation for next time, and a couple more are on the way. This is particularly funny to me because we automate our npm deployments on every project before it even goes live because we know how important it is to have that process be locked in instead of us running it on our own machines. Believe me, I like running it on my own machines, but we know better. To think that Anthropic has still been publishing Claude code as a manual deployment instead of it being done through an automated CI system is actually really funny and it makes sense that they would have accidentally published the source maps if a developer had them in their build locally and then published that would get that out. That said, the full lack of blame here and saying this is a process and system failure, not a human, and taking ownership on a team level for it is the right way to handle things like this. Blameless culture that doesn't avoid talking about the failures is really important because sometimes blameless goes too far and nothing gets fixed. And the opposite of just blaming the person who does things wrong also doesn't get things fixed because then somebody else can make the same mistake. This is a great response and I think a lot of people hadn't seen a good response to a thing like this before. So this deserves the positivity it's getting. We really should see more responses like this in crisis moments like what the Anthropic team just experienced. So, while I was actually DMCA struck by Anthropic, and that is actually really funny, I'm not that mad at them for accidentally doing it, assuming that the notice that they sent was valid and this was a mistake in communication between them and GitHub. I'm also proud of them for how well they're handling the comms on Twitter right now. That was a thing that I didn't think they would ever do right. And while they are being forced into the position due to the nature of the egregious leak that just happened, it is nice to see them stepping up and doing the right thing instead of just having the lawyers going around and doing their instead. But most importantly, I must remind y'all the thing I've been saying for a long time. All of this could have been avoided with a very, very easy strategy. Just open-source cla. It is very clear now that the damage of keeping cla code closed source is significantly greater than the potential cost of open sourcing it. And when you also factor in the benefits of it being an open source project, it is truly insane that they haven't done this. And the fact that they are sending DMCA requests on GitHub in the first place shows just how insane this is. Not only has Enthropic had the most repos taken off of GitHub of any company in history. This is a position they have put themselves in. This is their fault. The same way that positioning yourself in life to have as many lucky opportunities as possible, they have positioned themselves as a business to have as many unlucky opportunities as possible. They've overextended themselves in a way where more and more bad things can happen because they acted in bad faith for so long. So, while they are handling this crisis right, this crisis is their fault, not because they're so dumb they leaked to the code, because they didn't put it out in the first place. This entire thing is Anthropics doing because they won't just release the source. And at this point, it just feels like they're not doing it out of spite because they don't want to admit they were wrong for keeping it closed source for as long as they have. That's all I got on this one. I'm pretty tired of the topic, but I also do have two other anthropic videos that I had filmed before all of this that I have to air. So, uh, sorry for anthropic week. I guess we'll talk about the rate limit soon.