OpenAI has introduced a new update called function calling, which allows large language models like GPT-4 to use external code, databases and other APIs to accomplish tasks. Previously, GPT models had limited capabilities and relied on third-party libraries to implement tool use, which was often slow and inconsistent. With function calling, GPT can call out to external resources when necessary, making it significantly more powerful and replacing some of the functionality of open-source libraries. However, this development raises questions for companies building at the infrastructure layer, as their work could be quickly obviated by new OpenAI features.
Mira el artículo original en inglés aquí o también traducido al español aquí
