Replit is no longer working with Anthropic, CEO says (techcrunch.com)

Replit, the popular AI coding platform, has announced it is ending its partnership with Anthropic, as revealed by CEO Amjad Masad at TechCrunch Disrupt. The decision, attributed to factors like cost and performance comparisons, comes shortly after Replit's collaboration with OpenAI was established in August. Masad praised OpenAI's ChatGPT for its superior performance according to internal benchmarks and public metrics, particularly for tasks involving vibe-coded applications. This shift reflects a strategic pivot towards partnerships that better align with Replit’s growth trajectory and revenue objectives. While Masad's comments suggest a clean break, some insiders dispute the assertion, indicating that Replit may still be utilizing Anthropic's Claude models in certain capacities. This highlights ongoing complexities within the AI partnership landscape. Additionally, Masad emphasized the growing significance of enterprise contracts, which have expanded from just 2% to roughly 15% of Replit's revenue within a year. He also called attention to the importance of fostering a safe AI environment, advocating for rigorous safety measures, orchestration, and error recovery to ensure effective AI deployment. This nuanced landscape underscores the evolving dynamics in AI partnerships and the critical need for safety in AI applications.