Nothing works and everyone is euphoric

# Summary Despite AI tools promising greater productivity and higher-quality software, real-world experiences show software is deteriorating across the board—banking apps require multiple logins, applications crash and steal focus unexpectedly, and car infotainment systems are riddled with bugs that affect driving safety. The author suggests that while AI models are powerful tools, teams with access to cutting-edge LLMs and generous token budgets are still producing increasingly fragile, buggy software, indicating the problem lies not with the technology itself but with how it's being used or integrated into complex modern systems.