The Pulse: is GitHub still best for AI-native development?

This article examines GitHub's recent reliability problems, highlighting a drop to 'one nine' availability (99.9%) due to infrastructure overload from AI agents like Claude Code. It details three major incidents involving database saturation, failover telemetry gaps, and Redis configuration issues, as acknowledged by GitHub's CTO. The piece contrasts GitHub's struggles with a startup, Pierre Computer, which claims better handling of AI-native development workloads. It explores whether GitHub's legacy systems are straining under new AI-driven demands, making it relevant for IT professionals evaluating version control platforms for modern development.