Anthropic thinks they're Apple. They're actually hypocrites.

Sup y'all. Anthropic's number one hater reporting for duty because they just made a big big mistake. This is something I'm very excited about because it happens to be a weird niche I know way, way, way too much about. The niche is iMessage. This might sound crazy. Why are we talking about iMessage and iPhones when we're supposed to be talking about anthropic and the weird rate limit stuff that they just did? Well, I already did the video on the rate limit stuff, and I think that was more excusable than most of their [ __ ] but also their com sucked. That's not what we're talking about, though. We're talking about this. I bought a Mac Mini so I could have blue bubbles when texting Claude and it started roasting me. Try the iMessage plugin for Claude Code today with /plugin install iMessage at Claude plugins official. Kenneth here is an anthropic employee and he just advertised a product that actively breaks Apple's terms of service. To be clear, terms of service are not the law. They are the recommendations the company has for how you should use the thing. And if you don't follow them, they have the right to ban you as a user because it's the terms of you using the service. Anthropic can absolutely choose to ignore them. And they can also choose to not enforce them as they have been very well known for doing to many other players in the space. As somebody who has been very carefully, delicately working around Anthropic's terms of service in order to prevent my users from getting banned and our account from getting banned as T3 chat and now T3 code. I have had to work very hard to keep Anthropic happy even though their terms of service are not just unclear but they are also kind of [ __ ] And the same thing has happened to Open Code and Jet Brains and many many other groups. The hypocrisy is insane but it also shows a fundamental failure deep inside of anthropic that I really want to talk about. Do you know who isn't a hypocrite though? Today's sponsor. You know what OpenAI, Cursor, Replet, and T3 Code all have in common? I'll give you a hint. It's not that we use AI. So, we're all using the same O solution, Work OS. I could tell you all about their cool features. I really like the admin panel that makes it literally just sending a link in order to onboard other companies that have their own SSO platforms, but I'd rather just tell you guys about a fun thing that happened on Twitter. I announced that Work OS is one of our platinum sponsors. And normally when I announce things about sponsors, it gets no attention at all because who's interacting with a sponsored post on Twitter? This post got almost a thousand likes and a ton of people leaving their thoughts and feedback because work OS is so well-loved by my community. Actually, cool. We started using work OS based on your recommendation when we started developing our app. Safe to say we started right. Work OS is actually so good. It's got the best dev experience for setting up SSO. Work OS authent integration took me 2 hours versus 2 weeks with other providers. Their SSO docs actually work copy paste and their dashboard doesn't require a CS degree to navigate. Worth every penny when you're shipping under tight deadlines. And by the way, this person doesn't even follow me. Stop wasting time and losing customers when you could be shipping at soy. Oo. Oh, sorry. Didn't see you there. Theo from the future again. I filmed this video when this all happened a few days ago. But since then, there's been a lot of news about Cloud Code. This video isn't meant to be about the Cloud Code source leaks, nor is it meant to be about the hypocrisy of how they keep DMCAing people over all of that. But I do want to quickly highlight one more very funny hypocritical thing that Anthropic has done recently. In the source code leaks that many of us have looked into now, we couldn't help but search for open code and found that they are using open code as a reference for things like the autoscroll behavior or as a reference for how to size windows in the app. I just thought that was fun. Once again, anthropic being the greatest hypocrites of all time. Back to the original video. So, let's break this one down a bit. Anthropic are hypocrites is a bold statement to make, but I've never seen something so egregiously hypocritical in such a tight succession in this space. Literally a week and a half ago, open code was forced to comply with legal requests from Enthropic, removing the plugin that allows you to use the cloud code subscription inside of Open Code. So what this plugin did, the thing that they had to remove was allow users to hit the specific endpoint that served cloud code requests using their claude code off in a harness that was different from the user interface that Anthropic expects. So if we were to diagram this out, Anthropic has a server and this server has let's just say the endpoint is / API/cloud to make it clear it's/CCloud code. So there's this API and normally this API isn't hit directly because it really shouldn't be. If you want to hit the API directly, they have a separate one for that. They have API slash I don't even know what the endpoint is. We'll just call it usagebased. So there's the normal API endpoint that you hit when you're paying money per request to generate responses from anthropic. But there's a separate endpoint that is hit by tools like cloud code where it uses your subscription instead. So you likely have a user interface in front of these. It could be cloud.ai or cloud code or T3 code or open code or even codeex as open source. So let's say you have Claude code in front. I'm using Claude code. I pay for the subscription. I am the user over here. So I'm paying some amount of money. Let's just say I'm paying the 200 bucks a month in order to use Claude code which goes and hits the API for me. And it sends a response. And then I see that response as a user. To a lot of users, the thing they are paying for is Claude Code. And the API behind it just so happens to be the thing that fulfills the promise. But you're not really paying for cloud code. Cloud code is free. You can download it and play around with it. And you can even have it hit the API for usagebased stuff instead if you want. You can not have it hit the special endpoint and instead just have it hit the traditional usagebased endpoint instead and have it bill accordingly. So the subscription is effectively letting the user steer where the request goes. Cloud code is the same quad code. When the user is subscribed, instead of going to these endpoints and billing you per request, it instead goes to these endpoints where they have the subscription pricing where you get usage based on the tier that you pay for. So, let's be realistic here. What are users paying for? If you are subscribed to Cloud Code, drop a one in the chat if you're paying for Claude Code and drop a two in the chat if you're paying for the usage of the models. And it just happens to be through Claude Code. So one's in chat if you're paying because of cloud code and two's in chat if you're paying because of the usage inside of cloud code. Interesting. A lot of people are saying they pay because they like cloud code, but it seems like a roughly 50/50 split between the two. Personally, I am exclusively paying because of this. I think Claude Code is a pretty [ __ ] harness at this point. I'll do a whole dedicated video about harnesses and how they work in the near future, but I am definitely not paying my subscription because I like Claude Code so much. I'm paying because the model is pretty dang good at coding and is really dang good at UI. And unlike Gemini, it doesn't just have seizures and lose track of what it's doing all the time. So, I'm paying for this, but I don't like this. And since in the end, what you're paying for, literally speaking, is this because Cloud Code is free. It's just a binary you're downloading. You're paying for a subscription for this for Cloud Code, the endpoint. There's a lot of people, myself included, that want to use that separately. So if what I'm paying for is the usage, what if I want to put open code in the middle? Or what if I want to put VS code in the middle? Or what if I want to put pi or open claw or some other thing here? Anthropic does not want you to do that. It is anthropic's belief and arguably their right to have this special endpoint that they charge money for, which is their special claude code endpoint, and choose to not let you use it through other things. And that's what Anthropic wants. They don't want you to use this endpoint as a source of inference. They want you to use claude code to get addicted to cla code specifically and to go all in on it and its ecosystem so that you never leave. They're trying to sell this. But what they're charging for is this. And as such, the people who understand the economics here, they just want this. They just want the endpoint. I want to be able to use my cloud code sub with my openclaw because I like how cloud bottles talk more. I like working with them more. And I don't use my cloud code sub as much because I'm mostly using the GPT models for coding nowadays. I would love to use my cloud code sub for my openclaw instance, but I'm scared I'll get banned if I do, which is why I do it anyways cuz I'm just waiting to get that banned. It's going to be such good content when I eventually get banned regardless. Not everybody can just make content when they get banned. They're all very scared of it. We even have a ton of users and more importantly a shitload of misinformation that we're dealing with with T3 code. I would love to implement the clawed models by having them go through our own harness or pi or open code or something like that, but if we did that, we would have to charge way more. We'd actually have to charge because right now T3 Code is entirely free. You have to bring your own sub. And I want people to be able to do that. I know I can't be as cheap as Anthropic is because I don't get the discounts that they do. Anthropic charges you $200 and they give you $5,000 of compute for that per month as long as you go out of your way to max it out. Most people don't get close to that, but you can go way over that. We couldn't use your sub because we don't have the right to touch your access key, your OOTH token that goes through your subscription is a no-go zone. I'm not allowed to touch that as a developer. You're allowed to use it in cloud code and in cloud.AI, you're not allowed to use it outside of those things. They've been very explicit about this. They have this authentication and credentials use section in the legal compliance section of Claude Codes docs, which funny enough is the section I spend the most time in by far. This page changes every week or two and they've redefined it a little bit clearer here but still not very cloud code authenticates with anthropic servers using OOTH tokens or API keys. These authentication methods serve different purposes. Oath authentication which is used at the free pro and max plans is intended exclusively for cloud code and cloud.AI AI using OOTH tokens that are obtained through those methods and through those accounts in any other product tool or service including the Asian SDK is not permitted and constitutes a violation of the consumer terms of service. This is absolute [ __ ] Legal ease is meant to scare you and they do enforce this. If you use your subscription in other places like OpenClaw, there's a high chance you'll be banned. But what if I'm using this in a tool that directly wraps cloud code like T3 code where we require you to have cloud code and we're effectively just calling your existing cloud code install through our UI. Historically they've said you can use the Asian SDK for this but they've backpedal on even that and now they're saying it's not permitted. We are safe in T3 code because we are forced to use their [ __ ] CLI and integration layer that is closed source for no goddamn reason. Fun fact, Asian SDK is open source for Python but closed source for TypeScript because they're [ __ ] It is what it is. But they don't want us to do that because they don't want us to be able to compete with them with a better product that uses the same subscription. They're taking advantage of the fact that they can charge 25x less to force their product onto people. And it sucks. Do you know who else makes it nearly impossible to use their APIs the way that users want to in order to lock people in? Our friends over at Apple. I'm in an interesting spot here because I find myself equal parts stuck defending Apple and stuck being the only one [ __ ] on them for the things they're actually doing wrong. Just to clear the air before we go any further. Batterygate was Apple being prouser, not anti-user. You have to fundamentally misunderstand the case if you think otherwise. Apple does not do planned obsolescence. They actually have a massive market incentive to keep their devices working for long term because most iPhones don't get thrown away or sold when they are done being used. they get handed off to somebody like a kid or a friend or a young cousin or somebody who doesn't have a phone yet or a good phone. Apple needs to keep their old phones working well. So, if you think otherwise, you just don't [ __ ] get it. Their stuff's not overpriced anymore. It hasn't been for a while. Basically, since Apple silicon, it hasn't. Apple actually represents one of the best values in the market, generally speaking. Just a few weeks ago, I had somebody claiming that I wasted all my money buying the 5K Studio Display XDR for $3,000. I said, "Okay, if you find a better value display with similar specs, specifically the resolution and the refresh rate, as well as ideally a mini LED with similar dimming zones, I will buy you one if you find it for cheaper." I forgot to update y'all chat. He actually hit me up. I'll show you guys what he sent. Never mind. He deleted his messages. He sent me a 4K 60 Hz display with a much worse panel. Are you joking? I said, "I need this resolution and this refresh rate and ideally this panel technology." and he hit zero for three. Not a single goddamn speck. So yeah, all the people who think Apple's overpriced are just [ __ ] brain dead and don't know what specs even mean anymore. All the people who think Apple has planned obsolescence are beyond stupid. I have my thoughts here. I have my opinions here. Just unlike other people's, my opinions are correct. All of that said, I'm also Apple's number one hater. There's a reason I've never gotten a deal with them. They've never sent me anything. Apple hates me more than they like me. I know this for a fact. I have way too much internal information. Apple's way of running their app store is borderline [ __ ] extortion and it should probably be forced to do different based on the laws. It's horrible. The fact that most people don't realize when they buy their phone that Apple chooses what apps can and can't exist for them is absurd and they should be tried and fleeced in the courts for that as Epic has been helping with a lot. You can watch my other coverage of that if you haven't. I don't want to talk about all of that. I just want to make sure it's clear that I both defend Apple when they do things right and people misunderstand. and I will [ __ ] tear them to shreds when they do things wrong, which is what makes the iMessage thing interesting. It's weird getting to talk about this in this capacity because I've gotten into a lot of fights about this. One of these was the original creator of the Pebble who at the time was building a project called Beeper, which was a method app solution to give you one app for using chat across all the different apps that have chat. So you could have one app that had Discord, Slack, Twitter, LinkedIn, iMessage, WhatsApp, Telegram, all of those. These chat aggregate apps are the ultimate [ __ ] power pit. Everybody thinks they're a good idea and then one feature is missing from one chat and you give up and go back and then you stop using the thing. I don't know anybody who stayed on a chat aggregator for more than like a week. They just they don't work. They're not good. Nobody does it. That's just a fact. And every single time this happens, the founders convince themselves it's not their fault. It's Zoom's fault for not exposing the API. It's Apple's fault for being so evil with iMessage. It's always one of those things every single time. I have a hot take here. That founder was specifically trying to say Apple was breaking the law by not exposing their API to third party users who want it. It was his belief that Apple was doing a vertical monopoly by having their messaging app with an API only they could use on a phone that they were selling. I have to fall on the side I don't like falling on which is defending Apple here. The law should not control what products you build. Apple did not want to make an API for messaging. Apple wanted to make an app that did SMS and MMS but also when you were online would route it through their own separate trusted servers from their own trusted devices that they sell. That is the product Apple wanted to make. They should not be compelled by the law to do a different product when this is the product they want and their users are totally happy with it. If Apple was to make it so other apps couldn't work on the phone, then this would be very different. If, let's say, Apple, I don't know, banned WhatsApp or refused to let Telegram ship updates because they wanted their thing to win, that should be illegal. Let's say Apple has their own subscription service for a thing that has competitors like, I don't know, Apple Music. I think it should probably be pretty scrutinized that they charge such a high percentage to their competitors that they don't have to eat. They don't have to eat the 30% hit even though Spotify does. So if you choose Spotify over Apple, they still get 30% of the money they would if you picked Apple Music instead. That's really bad. That should probably regulate some amount. Those things I don't like. And if Apple put in a special API in their phone that only worked for Apple Music and didn't work for competitors, if they had a new feature like AirPods had a cool thing that only Apple Music could do that they didn't allow for competitors to implement, that should absolutely be illegal. Competitors should be allowed to compete on the platforms with their own fair alternative product. If people don't like Apple Music, I should be able to make an alternative and sell it to you. I should not be able to force Apple to give me their API to have access to all of Apple Music. I should not be able to legally force them to do that and provide a product they don't want to. And guess what? I feel the exact same way about Anthropic. Anthropic's behavior here should not be illegal. Anthropic is absolutely within their rights to offer this service in the way they do to decide they want to charge you this amount of money they do to give you something better than what you're paying for as long as you use it through their services. The same way Apple wants you to use their servers through their iMessage app on their phones. Anthro wants you to use their subscription service on their servers through their user interfaces. These are the exact same thing. So why am I being so nice to Apple and so mean to Anthropic? It's cuz Ananthropic are hypocritical bastards. That's why it is so [ __ ] funny to me that they just did to DAX the exact same thing that this should happen to them here because it is absolutely against Apple's terms of service to use their APIs this way. And unlike Enthropic, Apple's very explicit about this in their terms. They've made it incredibly clear. And thank you to Community Notes, by the way, for highlighting the multiple different policy violations Anthropics plugin is guilty of with Apple's terms of service. iMessage is intended for communicating with family and friends and is not for conducting commercial activities or disseminating unwanted messages. And when you're setting up a Mac, you agree to these terms which say very explicitly you may not and you agree not to or enable others to copy, decompile, reverse engineer, disassemble, attempt to derive source code of decrypt, modify or create derivative works of the Apple software or any services provided by the Apple software or any part thereof which you have to do because you are reverse engineering iMessage to do this. And one last one, you also agree that you will not do anything that interferes with or disrupts the service, including accessing the service through any automated means like scripts or web crawlers or any servers or networks connected to the service or any policies requirements. You get the idea. You're not allowed to access their servers in ways that are not through their applications as a user. They have been explicitly clear about this in many places in their terms. Meanwhile, we have poor Matt PCO trying desperately to get anybody at Ananthropic to answer the question of whether or not he can use this for an open-source project that he is building, trying to make the Claude code SDK and UX for developers better. He's not trying to make money on this. He's not trying to sell something. He's just trying to make a wrapper for quad code locally that users can use their existing subscription with to do different things and to do better things. He has been trying now for almost a month to get an answer out of anthropic. And credit to Thoric for trying, we basically just get sorryries. And after he said sorry enough times, Matt got fed up and just said ETA. Like Matt PCO is the very definition of a good faith actor. You can talk all the [ __ ] you want to on me. It's understandable that anthropic doesn't answer my questions. I get it. But not answering Matt in his many times over desperate attempts to get an honest answer about this and to have a real back and forth about it. Nobody knows what the [ __ ] the policy is, and I know for a fact the Anthropic employees don't either. They're intentionally keeping it vague because they don't know where they want to draw the line yet. They're giving themselves the freedom to arbitrarily kick out whoever they want whenever they want, and at the same time are doing [ __ ] that should and hopefully will actively get them banned from Apple's development ecosystem. Epic Games did less than this, and the result was that every single Mac developer at Epic had their computers locked by Apple. Apple should not have done that. And thankfully the courts reversed it because Apple should not have [ __ ] done that. But god damn if anthropic isn't hypocritical as hell here. The fact that the same company is now selling a solution to working around Apple's policy with how you're supposed to access iMessage. the same week they're sending [ __ ] requests in legal threats to open code to take down their plugin that does the same thing in a way that is way more reasonable and makes way more sense is just such an absurd level of hypocrisy that I feel like I'm going [ __ ] mad even talking about it. It's insane. I've never seen anything like this where it's like I didn't even have to work to make the analogy work. I just described the two things and chat was immediately like, "Oh yeah, that they're the same thing. That makes sense." How the [ __ ] is Apple the better faith actor? In what world is Apple and their absurd way of running things, their locked down closed walls [ __ ] feel more transparent and open than what Anthropic is doing here? What the [ __ ] And unlike Apple who made it trivial to vote with your wallet and if you don't want to use iMessage, you can install a different app, Anthropic has made it really hard to do anything with their stuff. You can use a different app with their models. You just have to pay 50 times more money. And you can't even use other models with their tool. If you do happen to be one of the people who likes Cloud Code the CLI, they are pretty strictly trying to keep you from using it with other models. That's also why it's closed source. With iMessage, it's the same exact app if you use SMS instead. Yeah, this is just absurd. The fact that I can highlight a bad policy that is enforced in weird ways between Apple and Enthropic and it makes Apple look good should show you just how anti-user Anthropics [ __ ] is. And if anybody tells you that Anthropic is like Apple, the way that they love their walled gardens and try to lock people in the way they do, they are right to an extent. But they are more Apple than even Apple is here. They are so goddamn closed. They are so goddamn locked down. And they are so unwilling to be transparent at all about these things. It's insane. So, if you want to vote with your wallet here, you have to do it a bit different. Instead of installing a different chat app on your phone, you have to install a different harness. Thankfully, you can use basically any harness and your subscription with OpenAI will work. That's why OpenAI acquired Stipe, the creator of OpenClaw. They don't own OpenClaw. They just employ the creator of it. They let you bring your OpenAI sub to Open Code or Pi or OpenClaw. They don't care. They let you use it for whatever because their goal is to give you inference for a discount and give you different services to use that inference. In Enthropic's goal is purely to lock you in. Funny enough, even GitHub has been similar here where they're letting you use the GitHub co-pilot subscription. and across other things too. So, GitHub is doing this right. Kilo is doing this right. Open code is doing this right with their Zen subscriptions. OpenAI is doing this right with the codec subscriptions. Seems like pretty much everybody is letting you use the inference in other places except for our friends at Anthropic. And that's why we are where we are today. Anthropic is poking the bear. They are doing the same thing that they are accusing others of doing and sending legal threats for. And while I am not in support of Apple having the ability to lock a company out of their computers because they don't like something they did, I am also not in support of Anthropic doing the same thing, which is pretty much exactly what they did here. So, [ __ ] them. If they're going to play hard, they should get [ __ ] played hard back. And if Apple decides to [ __ ] over Anthropic, I will compliment them as they do it. Even though I don't like the ability for companies to do those things, they are the hypocrites here. Anthropic deserves to be burned. I actually can't believe they announced this. I I I am blown away that the legal team spent more time harassing Dax than they spent realizing that this could [ __ ] them. And I do genuinely hope this [ __ ] them. I think that's all I have to say on this one. I'm just happy to see the tides turn more and more when I complain about Anthropic. The comments aren't, "Oh, you're an open AI paid shell." Even though they've never paid me anything, people are realizing just how awful Anthropic has been, how awful they are to interact with. And beyond poor Thoric on Twitter trying his best to clarify things. It just feels impossible to know what's going on there. It feels like they look down on us for liking their product, and I will always hate when I feel that way. So, stick with companies that are more transparent and care more about their users, like Apple. Yes, Apple. God, how did we get here, guys? Yeah, I got nothing else on this one. I hope this was a helpful video. [ __ ] anthropic. And until next time, peace nerds.