Is GitHub Now a “Crime Against Software”? The Current State of Development Infrastructure in AI Overdrive
📰 News Overview
- Severe Lack of Reliability: As of May 2026, GitHub officially acknowledges dozens of outages each month, with the actual number likely much higher, indicating a state of SLA violations.
- AI-Driven Distortions: Under Microsoft’s wing, the addition of “flashy AI features” is prioritized over basic reliability and browser compatibility, leading to increased load from AI agents.
- Abnormal Resource Consumption: The bloated frontend consumes over 512MiB of RAM just for pull request pages, frequently crashing browsers.
💡 Key Points
- Corruption of Infrastructure: The “deterioration of big tech,” similar to AWS and Google Search, is becoming evident in GitHub, the hallmark of development.
- Complicity in Malfeasance: Services for buying and selling GitHub stars are publicly available, fostering the spread of malicious repositories.
- Design Flaws: GitHub Actions, designed as “YAML-written shell scripts,” suffer from poor architecture, and cache invalidation fails to function properly.
🦈 Shark’s Eye (Curator’s Perspective)
It’s ironically tragic that GitHub, designed for developers, is being crushed under the weight of its AI agents! The bloated frontend that devours 512MiB of browser memory could easily be labeled a “crime against software.” Ignoring Firefox and Safari while focusing solely on funneling users toward AI features feels like a betrayal to the developers who have supported it for years! The neglect of “star buying” and the resulting damage to ecosystem reliability is a fatal flaw for any infrastructure!
🚀 What’s Next?
Developers are likely to accelerate their shift away from GitHub’s monopoly towards lighter, more reliable alternatives like Codeberg (Rating C) and GitLab (Rating D). Tools that lack reliability in an “AI-first” environment will ultimately lose the support of professionals.
💬 One Sharky Thought
No matter how flashy the AI features are, a site that doesn’t work is just junk! Even sharks need to keep swimming (basic functionality) or they’ll sink! 🦈🔥
📚 Terminology
-
Agentic Load: The system load generated when AI agents autonomously operate GitHub instead of humans. Microsoft’s AI strategy is a major contributor.
-
Star-buying: The act of purchasing “stars” for GitHub repositories with money to artificially inflate perceived popularity and credibility. This is often used to spread malicious software.
-
Heap Spikes: Instances where memory usage suddenly spikes during program execution. This occurs on GitHub’s PR pages and can lead to browser crashes.