Ten weeks after Groq shut down its models, GitHub still finds them in 314k files (tiltmeter.dev)

In a recent field report, it was revealed that 314,000 files on GitHub still reference the Groq AI models, which were shut down ten weeks ago. The findings underscore the pervasive issue of retired APIs remaining in active production code. Tiltmeter, a tool designed to adapt code to API changes, has already opened 64 pull requests aimed at addressing these deprecated connections in various repositories, including notable projects like LibreChat and DB-GPT. The challenges highlighted include not only the direct replacement of retired models but also the complexities of changing request structures needed to ensure functionality with new models. This situation is significant for the AI and machine learning community as it underscores the importance of actively managing API dependencies in a fast-evolving landscape. The ongoing reliance on outdated API calls can lead to system failures and mishandled data, especially when the models are no longer functionally supported. Tiltmeter's approach of automatically detecting API changes and facilitating code adaptations before breakdowns occur could set a new standard for handling such transitions, particularly for smaller vendors and those providing undocumented endpoints.