Thoughts on AI, Code, and Culture
Deep dives into how we build GitSniff, our philosophy on usage-based pricing, and the future of software engineering.
Claude API Limitations: The Seed Parameter Gap and Pricing Reality
Claude excels at reasoning, but missing reproducibility controls and premium pricing create real trade-offs for production systems.
The Future of Code Review: Predictions for 2026
Agentic AI, real-time collaboration, and automatic remediation. Here is where we think code review is heading in 2026.
OpenRouter Integration: Access to 400+ AI Models
Different models excel at different tasks. With OpenRouter, you can evaluate and route across a huge catalog without rewriting integrations.
Why Developer Experience is Our North Star
The best security tool is one that developers actually use. Here is how we design for adoption, not just compliance.
Beyond AI: Integrating Security Scanners into Your Workflow
AI catches logic bugs. Security scanners catch CVEs. Together, they form a comprehensive safety net for your codebase.
Quality Without Councils: The Review Pipeline That Scales
We moved away from multi-model councils. A tighter review pipeline delivers higher signal with less overhead.
Smaller PRs, Faster Shipping: The 200-Line Habit
Big refactors feel productive, but small, frequent pull requests keep teams shipping with less risk.
Leveraging Gemini 3 Pro for Million-Token Context Audits
Static analysis misses the forest for the trees. With Gemini 3 Pro and ~1M-token context, we can reason about architecture-level security risks at scale.
Why we moved from Per-Seat to Usage-Based Pricing
The SaaS industry is broken. Companies are paying for shelfware. Here is why we decided to charge for code analyzed, not for developer headcount.