OpenClaw Killer - Claude Computer Use

By Beau Johnson · · 16 min · 2.2K views

About this video

We do 6 live bootcamps every week in Shipping Skool! Full courses on OpenClaw and Claude Code! Join Here ⬇️ https://www.shippingskool.com/ 🔗 GET CONTENT MACHINE: https://www.shopclawmart.com/listings/content-machine-0c67b3b3 Breaking down Anthropic's just-announced Claude Computer Use feature that lets Claude control your Mac. Full comparison to OpenClaw - what each does better, security concerns, and why this actually validates everything the OpenClaw community has been building. Plus the press is noticing. Get the weekly AI builder newsletter 📕 https://substack.com/@buildnpublic Follow Me On X - https://x.com/BeauJohnson89

Transcript

Read the full transcript

0:00Anthropic just announced that Claude can now control your entire computer. And in this video, we're going to break that down and then compare it to Open Claw. It can take over your mouse, it can take over your keyboard, click around, open apps, fill in spreadsheets, basically everything that you would do on the computer, now Claude can do. And I have a lot of thoughts about this because you see all the Open Claw videos, is Open Claw dead? I think we could pump the brakes a little bit, but let's go over what Anthropic just shipped literally about 4 hours ago. You know, I run 13 AI agents on the Mac Mini behind me and I've been living in this world for 45 days now and I'm going to break down what this actually is and how it compares to Open Claw. And what it means for builders like you. So, it's actually great news for everyone in the Open Claw community, just validating kind of what

0:46we've been working with with building out AI agents. I believe that Anthropic's just shipping features that makes it more available for the non-tech user. And, you know, Anthropic announced two things that really make sense for a lot of people getting into this whole space. Claude Co-work, which is basically Claude working alongside you in your desktop, and then Claude Dispatch, which lets you assign tasks to Claude from your phone. The big headline is the computer use. Claude can now point and click and scroll and type on your screen like a basic human would. It can open apps and navigate websites. It fills out forms and such. So, it does stuff like you would normally would do and sit there and do yourself. So, good time saver, I think. But, the catch is this is a research preview. It's only available for Claude Pro subscribers at

1:3220 bucks a month and Claude Max subscribers at 100 to 200 a month. So, it's macOS native right now, no Windows. That's kind of to be expected. There's no Linux, either. It's just Mac. So, I'm going to walk you through how it actually works because there's some interesting stuff happening under the hood. When you give Claude a task, it first tries to use what they call connectors. So, if you say schedule a meeting for tomorrow at 2:00 p.m., Claude checks if it has a Google Calendar connector. If it does, it uses it directly. It's clean, it's fast, it's efficient. But, if there's no connector for what you need, that's when Claude fails back to actually controlling your screen. It moves around the mouse, it clicks buttons, it types into fields. It's literally watching your screen and figuring out where to click. I I think

2:18that's pretty cool. They built in a permission system as well. Claude always asks before accessing your new app and you can stop it at any time. Some sensitive apps are turned off by default, so it's not just running wild on your machine. They clearly thought about the safety angle. Now, the dispatch thing is really interesting. You can message Claude from your phone and assign tasks while you're away from your desk. So, imagine you're driving to work, headed to Taco Bell, and you tell Claude, "Hey, check my email, summarize anything important, and draft replies for the urgent ones." Now, and Claude goes and does that on your Mac while you're not even sitting in front of it. That's pretty cool. One continuous conversation between your phone and your desktop. I mean, I think we kind of already know where this is going. Real quick, before we keep going, if you're

3:03watching this and you want to actually build with some of these tools, not just watch videos about them, then you're going to want to check out our community down below, Shipping School. We have a full Claude code course, a full Open Claw course, and four live bootcamps every single week where we actually help you get set up from scratch. Like actually set this thing up, not just watch a tutorial and figure it out by yourself. And we also provide one-on-one coaching, so you could book a call with me. We could share screens and I can help you get Claude code or Open Claw running on your machine. That's it, no fluff. I built this community because watching YouTube's only gets you so far. We launched it just 3 days ago and we have over 55 members. You need people around you who are actually building,

3:48people who hold you accountable, and coaches who can help you when you get stuck. I'll put the link in the description down below. Get in now before the price goes up. They gave an example that you could tell Claude to check your emails every morning, run your test suite, and pull up a pull request if anything passes. That's a daily workflow basically running on its own, which is pretty cool when you think about it, you know. Here's where I need to get real with y'all and, you know, put this plain and simple. And because, you know, there's just some serious security concerns that here that no one is talking about. There's no press release talking about it. Anthropic themselves said this and it's still early. They said it may make mistakes. They're scanning prompt injection attacks, which if you don't know what that is, it's basically when a malicious website or document tricks the AI into

4:34doing something you didn't ask it to do. And when the AI has control of your mouse and keyboard, not good. That's just a bad chatbot response. That's someone potentially controlling your computer through Claude. They recommend not using it with sensitive data. Some apps are disabled by default for a reason. And look, you know, I'm not trying to scare anyone, I swear, but giving away your full machine to AI control of your computer is a real security risk. You need to understand what you're opting into, which is why I feel so comfortable just running Open Claw on a fresh, clean machine with nothing on there. It just gives me a peace of mind for logins and banking information. Could you imagine? So, this is a research preview that Anthropic just dropped and it's research preview for a reason. It's not production ready.

5:20They're being upfront about that and I respect Anthropic for being honest, but just know what you're getting into. Here's the part I've been wanting to talk about the whole video is the comparison to Open Claw. Because if you've been watching this channel, you know I've been running the entire content operation on Open Claw for 45 days now, 13 agents, 24/7 running on a Mac Mini sitting on a desk. And when I saw this announcement, I had to make a video. I wasn't threatened and neither should you if you have a great Open Claw setup going. But, I was a little bit excited cuz I wanted to know what was under the hood here and I'm going to explain why. So, what does Claude computer use do that Open Claw doesn't? That is a fair question. The big one is the visual screen control. Claude can literally see your screen and click around on it. Open Claw doesn't do that.

6:05Open Claw works through APIs, through command line tools, through skills and plugins. It doesn't take over your mouse. I would say that's a genuine difference and it is pretty cool tech. Also, the second thing is the native macOS integration is pretty smooth. You don't need to set up a bunch of stuff. It just works within the Claude app. Open Claw has more setup involved, you know, you're configuring agents, you're installing specific skills, you're setting up cron jobs. It's a builder's tool, not a consumer app. The third thing is the mobile dispatch. I don't know what to say about it. It's, you know, you're being able to message Claude from your phone and have it work on your Mac is is genuinely useful, but I think there's more to it. You know, Open Claw already has Telegram and Discord and WhatsApp and over 15 other integrations for chatting with your AI agents. But, this phone-to-desktop

6:50workflow with Dispatch is a different vibe. I'll give it that. The fourth thing is the built-in permission system. It's pretty nice. You know, ask before accessing apps, you know, stop anytime, sensitive data apps disabled by default. I like that. Open Claw has safety measures, too, but they're more hands-on. Okay, so that's what Claude computer use has going for it. Now, I'm going to tell you what Open Claw does that Claude computer use doesn't even come close to touching. Number one is Open Claw works on every operating system. Linux, Windows, Mac, you name it. Claude computer use, Mac only. That's a huge limitation. Half of the developers I know are on Linux. A huge community is running Open Claw on $5 VPS servers on Raspberry Pis. You can't do that with Claude computer use. Second,

7:36this is the biggest one, is Open Claw supports every model. Claude, GPD, Gemini, Llama, MiniMax, open models, whatever you want. Claude computer use only uses Claude. I love Claude, though. I mean, I'm saying I love Claude, but not everyone does. So, you're locked into one model from the company. I get it. With Open Claw, you know, I'm running Opus for my main agent, MiniMax for my worker agents, and I pick the right model for the right job. I like the flexibility. The third thing is the skills and plugins ecosystem. Uh, Claw Hub. They have hundreds of thousands of skills that community members have built. Weather, YouTube analytics, GitHub automation. You want social media posting? You got it. You need a skill for SEO? Boom, it's there. Speech-to-text? Okay, you got it. You install them with one command and they just work. Claude computer use has

8:22connectors for like Google Calendar and Slack. Wow, that's a fraction of what Open Claw can do. Fourth, and this is where it gets real. 24/7 autonomous operation. My Wu-Tang pipeline runs every single day without me touching it. They're just in there doing moves like crazy and I don't even do anything. I just wake up to everything finished. Ghostface is our intel agent running at 5:00 a.m. gathering all the intel everywhere, all the time, all at once. The RZA writes the outlines for these videos. Inspectah Deck handles everything on X. U-God generating the thumbnails to be amazing. Method Man chopping up clips into short form. The GZA does the newsletters Monday and Thursday, mainly, but we just set it to every day cuz why not? So, all this happens automatically through cron jobs

9:08while I sleep. Claude computer use needs you to tell it what to do. Mm. So, it's more reactive. Open Claw is proactive. My agents have memory. They remember what happened yesterday. They remember I burnt my birthday. This is amazing. They learn from performance data. They get better over time. They improve on their own output based on what worked last week. Claude computer use doesn't do any of that. The fifth thing is multi-channel integration. I can get messages through Telegram or Discord or WhatsApp or Signal or MSN Messenger for all that matter. My agents can reach me anywhere. Claude computer use lives in the Claude app and that's it. Sounds depressing. So, the sixth one is there's no subscription wall. Open Claw is open source and free. You bring your own API keys and you're good to go. Claude computer use requires you to have a pro or max subscription, 20 to $200 a month

9:53depending on your use case, and you still have usage limits. With OpenClaw, I spend maybe $3 a day on API calls and get way more done than Claude computer use could do. Seventh one is the agent infrastructure is just way more mature. Memory systems, tool calling, scheduling, sub-agents, context management. OpenClaw has been building this for a while and it shows. Claude computer use is brand new and it feels brand new, but it's still cool. That's why we're talking about it. So, here's my honest take and I want you to hear me on this. Claude computer use is Anthropic building a walled garden version of what OpenClaw already does, but with screen control added on top. This is just 1.0. It can get much more robust. It's a consumer play. It's designed for the person who wants to download an app and have AI do stuff on

10:39the computer without configuring anything. And that's totally fine. Different avatar, different market. And there's a huge market for it. But OpenClaw is the builders builders. It's the builders platform. It's the power user platform. It's for people like us, you and me. We want full control. We want to choose our dang models. We want to build custom workflows. Who want agents running 24/7 without baby sitting them. I think we can all agree on that. And you know what? I don't even think there's any competitors to OpenClaw. I think they're complementary and probably can make OpenClaw better. Here's the thing, is Claude computer use could potentially be used through OpenClaw as another tool in the toolbox. If you're struggling to keep up with content, well, I'm about to save you about 40 days worth of work. something called

11:25Content Machine. It's 10 AI agents that run on the OpenClaw orchestration and they handle everything. Scripts, thumbnails, X posts, blogs, outreach, clips, newsletters, all of it. So, I went from 1,000 subscribers to 4,000 subscribers on YouTube in 7 days using this exact system. Every single morning I wake up [music] and the content's already done. I spend maybe 15, 20 minutes reviewing and approving them and I move on with my day. It works for any niche, fitness, finance, real estate, marketing, whatever you are building and it is 100% completely customizable to your use case. So, you get the mission control dashboard, all of the cron jobs,

12:10everything I've built over the last 40 days helping me gain more and more people to subscribe and join the community. So, you plug in your own thing and it molds it to you. It learns how you talk and it writes [music] so it doesn't sound like AI slop. $97 one time. It's not a subscription. I'll put the link down below and you'll thank me later. You know, if you imagine having OpenClaw orchestrate Claude's computer use capabilities alongside GPT's reasoning and Gemini's multimodal abilities, that's a pretty big power play. That's where basically this is all heading. You know, the press coverage on this has been pretty crazy. It just dropped a few hours ago. CNET literally said, I'm not making this up. Go check out their website. They said that Anthropic is boosting Claude's AI capabilities in an apparent effort to

12:55compete with platforms like OpenClaw. Silicon Angle speculated that Claude's computer use capabilities might be enabled through OpenClaw under the hood. 9 to 5 Mac, Engadget and The New Stack, they all covered this today. And on Reddit, already going crazy. The Claude AI subreddit community is already pointing out the differences. One of the top comment says, OpenClaw runs on every OS including Linux, supports all models with far less restrictions including local ones. People see through the hype. They understand that the single vendor lock-in solution isn't the same as an open platform. Here's the thing that gets me the most. The most excited about all of this is when Anthropic, a company valued at billions of dollars, decides to build their own version of what

13:40OpenClaw already does. That's validation. That means the open source community was right. The vision of AI agents running on your computer, handling tasks, operating autonomously, it's not a niche anymore. Anthropic just told the entire world, this is the future. You know, Jensen Huang said at the GTC last week that every company needs an OpenClaw strategy. Just the biggest company in the world, you know. He compared OpenClaw to Linux and now Anthropic is building their own version. This is exactly what happened with Linux. Companies kept trying to build proprietary alternatives and Linux kept winning because open source wins in the long run. I believe the same thing is going to happen here. So, if you're watching this and you're running OpenClaw, I would not panic at all. This is good news. If you're watching this and you're curious about getting started

14:25with AI, I'd honestly say try both and try to formulate your own opinion. You know, try Claude computer use if you're on a Mac and you want to do something simple. See how it works. But when you're ready to go deeper, when you're ready to build real workflows that run on their own, come back and check out some OpenClaw stuff. And if you need help setting any of this up, we have a community of 126 members. We launched this community 12 days ago and there's already people building full-on working apps, building out agentic workflows. There's people that have no technical experience and never even touched a computer have OpenClaw running from day one. So, and that's what we do inside of Shipping School. We help you make progress. We do six live calls every single week and we focus on community first and human to human interaction. No

15:10AI avatar bots talking to you. We do have courses in there as well, but we're really focused on building out the bootcamps. We have, like I said, 126 members. The community is incredible. Lots of great resources in there. So, I'll put the link to that in the description down below. And I just say this is the greatest time to be alive to be a builder. The whole AI agent space is absolutely exploding and we're right in the middle. Can I get an amen? I mean, praise God for this time. You have an idea, now's the time to build it. Check out the links below if you're interested in the community and we'll see in the next one. Be blessed.

Build this with us, live

Shipping Skool is a free community with 4 live build sessions a week. Bring your idea.

Join Shipping Skool — Free