Meta’s internal tool, Metamate, now combines OpenAI’s GPT-4 with Meta’s own Llama model to improve its coding capabilities for developers.
This integration allows Metamate to pull responses from the most suitable model, depending on the query.
While Llama is Meta’s flagship model, GPT-4 is used to address its limitations, particularly for more complex queries.
Metamate, which launched in early 2024, is comparable to other AI coding tools like GitHub Copilot, though it’s best for basic coding tasks.
Meta’s reliance on GPT-4 alongside Llama highlights the challenges even large models face.
Additionally, Meta’s philanthropic arm, the Chan Zuckerberg Initiative (CZI), uses OpenAI’s technology to develop a customized educational tool.
Despite its success, Metamate’s role in automation has led to further layoffs, as AI is seen as reducing the need for human workers.
