GitHub continues to experience frequent outages and performance issues, undermining its role as a cornerstone of software development. The platform’s 99.8% uptime claim is laughable given its history of unaddressed incidents, with users reporting daily disruptions. Efron Licht argues this reflects a broader pattern of decay in big tech infrastructure, where reliability is sacrificed for AI-driven features. The crisis isn’t just technical—it’s cultural, as GitHub’s design choices prioritize flashy tools over user trust.

Specific technical failures highlight the severity. GitHub’s front-end is bloated and slow, often crashing on Firefox and Safari despite millions of users. Its pull request system consumes excessive RAM, and Actions workflows are poorly documented and memory-leaking. These issues stem from Microsoft’s push for ‘agentic’ workflows, which have overwhelmed the platform. GitHub’s lack of transparency—no public bug list or clear status updates—fuels frustration. Users report settings resetting unexpectedly, forcing them into risky default configurations. The platform’s AI buttons now number four per repository, directly tied to Microsoft’s aggressive adoption strategy.

The consequences extend beyond inconvenience. For developers, GitHub’s instability risks undermining confidence in open-source collaboration. The platform’s prioritization of AI over core functionality mirrors a broader trend in tech where innovation is equated with disruption. Microsoft’s subsidized AI tools may have accelerated this by incentivizing excessive usage. Licht warns that without a return to fundamentals, GitHub could become a ‘crime against software’—a system designed for experimentation but failing to serve its primary purpose. The article concludes that reliability must be rebuilt, not reframed as a secondary concern.