Show HN: Orangensaft – A mini Python-like language with LLM eval in lang runtime (github.com)

Orangensaft is a newly introduced experimental programming language designed to blend deterministic scripting with the capabilities of large language models (LLMs). This hobby project offers a mini-Python experience where LLM evaluations are integral to the language syntax. For example, code snippets enclosed between `$ ... $` allow users to seamlessly integrate LLM calls that perform tasks like determining the longest name in a list or transforming items in an array into uppercase. This innovative approach highlights the potential for dynamic, AI-assisted programming, enabling developers to create code that interacts with AI in real-time. The significance of Orangensaft lies in its unique architecture that encourages a dual coding approach—allowing developers to choose between probabilistic LLM output and traditional deterministic functions. This flexibility not only enhances the capabilities of programming languages but also opens up new avenues for AI integration. Moreover, the language includes features for handling dataframes and CSV files effectively, allowing for more nuanced data manipulation with context-aware prompts, yet still providing deterministic options for exact computations. Overall, Orangensaft could spark new ideas in AI-driven programming paradigms, despite being labeled a "toy" project, suggesting a rich potential for future development in AI and machine learning integration.