- AI command interpreter with Claude/GPT-4 support
- Sandboxed execution engine with security controls
- Hardware profiling system (CPU/GPU/RAM detection)
Phase 1 is 60% complete already. The open source community is amazing.
Just posted 10 new issues with bounties ($25-100): github.com/cortexlinux/cortex/issues
Huge thanks to dhvil and Sahilbhatane for the excellent work.
Looking for more contributors - especially for package manager integration and CLI interface.
We're building Linux that understands natural language commands.
Example: "Install Oracle 23 AI with GPU acceleration" → System detects your hardware, installs correct CUDA version, configures everything, validates it works. Done in 5 minutes.
Problem: Developers waste 30% of their time fighting package managers and reading documentation.
Solution: Embed AI at the OS level. The system understands intent and optimizes for your hardware.
Open source, seeking contributors. We have 3 initial issues ready for PRs.
Update (12 hours in):
We now have working code! Three PRs merged:
- AI command interpreter with Claude/GPT-4 support - Sandboxed execution engine with security controls - Hardware profiling system (CPU/GPU/RAM detection)
Phase 1 is 60% complete already. The open source community is amazing.
Just posted 10 new issues with bounties ($25-100): github.com/cortexlinux/cortex/issues
Huge thanks to dhvil and Sahilbhatane for the excellent work.
Looking for more contributors - especially for package manager integration and CLI interface.
We're building Linux that understands natural language commands.
Example: "Install Oracle 23 AI with GPU acceleration" → System detects your hardware, installs correct CUDA version, configures everything, validates it works. Done in 5 minutes.
Problem: Developers waste 30% of their time fighting package managers and reading documentation.
Solution: Embed AI at the OS level. The system understands intent and optimizes for your hardware.
Open source, seeking contributors. We have 3 initial issues ready for PRs.
Tech stack: Ubuntu 24.04, Claude/GPT-4 for intent parsing, Firejail sandboxing, Python orchestration.
Feedback welcome - what would you want an AI-powered OS to handle?
> The Problem
> Installing complex software on Linux is broken
Agreed.
> The solution
> Cortex Linux embeds AI at the operating system level.
That is not the solution. Determinism is the goal; introducing AI will create more problems than it fixes.
Valid concern about determinism. Let me clarify:
The AI layer translates intent → deterministic commands. It doesn't execute randomly.
Example workflow: 1. User: "install docker with nvidia support" 2. AI generates: ["apt update", "apt install docker.io", "apt install nvidia-docker2"] 3. Sandbox validates commands against whitelist 4. Execute deterministically 5. Verify installation
The AI is the interface layer, not the execution layer. Commands are reproducible and auditable.
We're also building dry-run mode so users can review before execution.
Think of it like: AI = smart man page reader, not autonomous agent making random decisions.
Open to feedback on making this more deterministic. What would make you more comfortable with the approach?