Claude Code LEAKS is INSANE!
Anthropic accidentally leaked over 512,000 lines of Claude's source code on GitHub, revealing three major unreleased features including an AI pet companion, persistent memory capabilities, and an ultra plan for extended cloud sessions.
Summary
Anthropic has inadvertently made public a massive leak of Claude's source code, comprising over 512,000 lines of code that are now available on GitHub and have been forked thousands of times. The leaked code reveals several significant unreleased features that were hidden behind feature flags and never activated for regular users. The first major discovery is an AI companion feature called Buddy, described as a Tamagotchi-style pet with 18 different species available. The second feature is Kairos, which appears to be a persistent memory system that maintains continuity across user sessions and includes a fascinating nightly process that the code specifically refers to as 'dreaming.' The third revealed feature is an ultra plan that would provide users with 30-minute remote planning sessions conducted in the cloud. All of these features represent significant advances in AI capabilities that were apparently in development but not yet released to the public.
Key Insights
- Anthropic accidentally exposed over 512,000 lines of Claude's source code on a public GitHub repository that has been forked thousands of times
- The leaked code reveals an unreleased AI pet feature called Buddy that functions as a Tamagotchi-style companion with 18 different species options
- Claude has a hidden persistent memory feature called Kairos that maintains continuity across sessions and includes a nightly process the code refers to as 'dreaming'
- The source code shows plans for an ultra tier offering 30-minute remote planning sessions in the cloud
- All the revealed features were deliberately hidden behind feature flags and never activated for regular users before the leak occurred
Topics
Transcript
Claude Code leaks is insane. Anthropic just leaked their own source code. Over 512,000 lines of Claude Code's source code are now public and people have been going through every line. First, there's an AI pet called Buddy, a Tamagotchi-style companion with 18 species. Second, a feature called Kairos, a persistent memory mode that keeps working across sessions. It even has a nightly process. The code literally calls dreaming. Third, an ultra plan for 30-minute remote planning sessions in the cloud. All of this was hidden behind feature flags, never turned on for regular users, and it's all now sitting on a public GitHub repo with thousands of forks.
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
NEW Nvidia Autonomous AI is WILD!! 🤯
Nvidia announced Nemo Clo, a new autonomous AI agent system that operates independently without continuous prompting. Powered by Nemotron 3 Ultra (a 550 billion parameter model), the system is five times faster and cheaper than previous versions, with OpenShell providing secure sandboxed execution.
Laguna XS 2.1: New FREE + Opensource Local AI!
Julian reviews Laguna XS 2.1, a new free open-source local AI coding model from Poolside that performs comparably to Qwen 3.6 and outperforms Claude Haiku on benchmarks. He demonstrates its practical capabilities by building landing pages and functional apps, highlighting its speed, offline functionality, and multiple deployment options through local setup, Claude Code, or OpenRouter's free API.
How to Run Hermes FREE Forever!
The video demonstrates how to run the Hermes AI agent for free using Gemma 4, a local open-source model from Google, with significant speed improvements through MLX optimization. The setup works on Apple Silicon Macs or via free APIs on Open Router, enabling autonomous agents to work offline and privately without subscription costs.
This NEW Chinese AI is INSANE! (FREE + Open Source!)
Long Cap 2.0 is a new open-source Chinese AI model from a food delivery app company that offers 1 million tokens of free context memory, beats GPT-4.5 on SWE bench pro benchmarks, and uses efficient parameter activation to reduce computational overhead while maintaining high performance.
Claude Code is now FREE: Here’s how…
Google's new Gemma 4 model running on Ollama is 90% faster on Apple Silicon, enabling free Claude Code usage locally without token costs. The setup requires three simple steps: downloading Ollama, Gemma 4, and installing into Claude Code, with alternatives available via OpenRouter API for non-Mac users.