Research
How we think about AI security, web-native operating systems, and the architecture behind VibeOS — from peer-style publications to engineering deep-dives.
Publications
Our first research papers are in preparation. Check back soon — or follow our articles below for early thinking on the same topics.
Articles
Recent · Medium
Tool-Call Injection: Escaping AI Action Syntax to Run Arbitrary Operations
Modern AI assistants no longer just generate text. They drive agents that read files, send emails, query databases, call APIs, and modify systems — opening a new class of injection attack.
Read on Medium
Apr 16 · Medium
How DeepSeek Accessed Claude’s Reasoning — and What It Means for AI Security
In January 2025, a Chinese AI startup called DeepSeek released a family of large language models that drew widespread attention across the AI industry.
Read on Medium
Apr 16 · Medium
Why LLMs Don’t Give the Same Answer Twice — And Why That is On Purpose
An exploration of stochastic decoding, temperature, and why non-determinism is a deliberate design choice in large language models.
Read on Medium
Partner & Support
We are actively seeking collaborators, sponsors, and research partners who want to shape the future of AI and software. Join us in pushing the boundaries of what technology can solve.
Cyber Security in the Era of AI
Research the implications for cyber security threats with the advancement of AI and Machine Learning technology.
AI Integration Solutions
Research areas where AI can be utilized to find novel solutions to world problems.
Software Solutions
Research areas where software can be utilized to find novel solutions to world problems.