#Programming

Anthropic's Claude models generate so much code that the company created its own version control system, Scion, to prevent infrastructure collapse.

🚨 **AI Code Works, But Is It Safe?** The Veracode 2026 report shows that 44% of AI-generated code contains vulnerabilities. 📉 **Key Facts:** • Code is syntactically correct in 99% of cases, but secure in only 56%. • Model size and specialization have almost no impact on security. • Python is safer than Java (63% vs 30%). • GPT-5.5 leads with 68%, but most models fail the tests. ⚠️ **Conclusion:** Generation speed does not replace control. All AI code requires strict verification before implementation. #CyberSecurity #AI #Veracode #GenAI

Linus Torvalds has firmly responded to AI critics in the Linux project: he offered them either to leave or to fork the kernel. The OS creator recognized AI as a useful tool, noting that code quality has improved thanks to neural networks, while skeptics simply haven't tried using them. 🤖💻