How to install OpenClaw on an Amazon Cloud Computer
The video is a comprehensive tutorial and review of 'Claudebot,' an open-source AI agent that runs on a local machine or VPS, can take autonomous actions on a computer, and communicates through messaging apps like Slack. The creator walks through setting up Claudebot on a free AWS EC2 instance, connecting it to Slack, and demonstrates use cases like daily news aggregation, app building, and animation generation. He concludes with serious security warnings about prompt injection and the risks of giving an AI agent broad system access.
Summary
The video opens with the creator describing Claudebot as a viral sensation in the AI community on X (formerly Twitter), distinguishing it from traditional AI chatbots by its ability to take real-world actions on behalf of users rather than just giving advice. He outlines five key differentiating features: it runs locally with access to files and applications, it can be controlled remotely via messaging apps like Slack or Telegram, it has full system terminal access, it has persistent memory across sessions, and it is self-improving through installable skills.
The creator then addresses costs, clarifying that Claudebot itself is free and open-source, but users will incur API fees from services like Anthropic or OpenAI depending on which models they use. He notes that a Claude Max subscription at $200/month can cap those costs. He outlines three installation options: on your existing personal computer (easiest but riskiest), on a dedicated machine like a Mac Mini (popular trend), or on a cloud VPS.
The bulk of the video is a step-by-step tutorial for setting up Claudebot on a free-tier Amazon AWS EC2 instance running Ubuntu with 8GB of RAM. He walks through creating the instance, connecting via terminal, running the Claudebot install command from claud.bot, and going through the onboarding wizard. He selects Anthropic's Claude Opus 4.5 as the model and configures Slack as the communication channel by creating a Slack app, configuring it via a JSON manifest, and linking the bot and app OAuth tokens.
Once connected, the creator demonstrates several live use cases. He performs a 'brain dump' of personal information so Claudebot can build a persistent memory profile and suggest relevant automations. He then asks Claudebot to set up a daily AI news aggregator, which it configures as a cron job running at 8am Pacific, delivering curated headlines from sources like The Verge. He also asks it to install Claude Code for building apps, which it does in seconds. He attempts to install Remotion for video animation generation, encounters some setup friction due to using the wrong install command, but eventually gets it working through a Cloudflare tunnel and receives a basic animation of his website FutureTools.io.
The creator closes with strong security warnings, disclosing he shut down his EC2 instance after filming due to concerns about exposed security vulnerabilities. He warns specifically about prompt injection attacks, where malicious instructions on a visited website could hijack the bot's actions. He advises running Claudebot on a dedicated machine, using separate phone numbers and email addresses for integrations, and treating it like a new untrusted contractor in terms of access granted. He frames Claudebot as a landmark 'iPhone moment' for AI — a cultural shift where AI moves from giving advice to autonomously executing tasks.
Key Insights
- The creator argues that Claudebot's persistent memory — remembering preferences, projects, and communication style across all sessions — is the single most impressive feature that sets it apart from other AI tools.
- The creator claims that Claudebot represents an 'iPhone moment' for AI, describing it as a cultural phenomenon within the AI community where people broadly recognize for the first time that AI can autonomously do things on their behalf rather than just give advice.
- The creator warns that prompt injection is a real and non-theoretical threat with Claudebot — a malicious website could embed instructions in its text to redirect the bot to leak private information or take harmful actions.
- The creator demonstrates that Claudebot can self-diagnose and fix its own configuration errors in real time, as when it detected a misconfigured Slack API token and provided step-by-step corrective instructions without being asked.
- The creator asserts that Claudebot itself is completely free and open-source, but real-world usage costs are driven entirely by the external APIs it calls — such as Anthropic or OpenAI — meaning heavy users of top-tier models like Claude Opus 4.5 could accumulate significant API charges.
Topics
Transcript
[0:00] Cladbot is taking the world by storm right now and it is absolutely blowing my mind as well as the minds of everybody that's using it. It is an AI assistant that not only gives you advice and you can chat with like the other bots, but it will actually go and take action and do things on your behalf better than anything I've ever seen before. And it doesn't matter what model you're using. It can use like all of the models. Just in my own experimentation in this video, I got it to create a chatbot with me in Slack. Create a daily AI news digest for me that delivers to [0:31] Slack every single day. Install…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Matt Wolfe
The ONLY AI Benchmark You Need!
A developer created "Buccy Bench," a humorous yet functional AI benchmark that tasks different language models with drawing Gary Busey as SVG code rather than images. The benchmark tracks model evolution over time while measuring performance metrics like cost, tokens, and execution speed.
GLM-5.2 - The Open Model That's As Good As Opus!
A comprehensive review of GLM-5.2, an open-weight Chinese AI model with a 1 million token context window, demonstrating its capabilities for coding, document analysis, and agentic workflows at significantly lower costs than frontier models like Claude Opus and GPT-4.5. The speaker tests various use cases including website building, Chrome extensions, game development, and data organization, concluding it's valuable for long, code-heavy, token-expensive tasks despite not universally outperforming closed-source alternatives.
Don't Fall For This AI Trap
The speaker emphasizes that power users distinguish themselves by knowing what NOT to automate with AI, rather than automating everything. They argue that AI works best for clear, straightforward tasks but struggles with nuanced, artistic work requiring consistency—using their failed YouTube thumbnail automation as an example.
AI News: Fable Banned, New Open-Source Leader, Midjourney Shocker
This AI news roundup covers the US government forcing Anthropic to shut down its Fable 5 and Mythos 5 models due to a security vulnerability jailbreak, the release of a competitive open-source model GLM 5.2 by ZAI, and MidJourney's surprising pivot into medical imaging technology with a new ultrasound-based body scanner.
AI News: Claude's Massive Leap & Siri Gets Good!?
This AI news roundup covers the release of Claude Fable 5 (a Mythos-tier model from Anthropic) and its controversial safety restrictions, Apple's WWDC AI announcements including a major Siri overhaul, and updates from Google including NotebookLM upgrades and real-time translation. Additional rapid-fire items include OpenAI and SpaceX IPO filings, ChatGPT email sending, and a teased Midjourney hardware device.