Meta tells applied AI engineers to restrict coding assistant use
Meta just told its Applied AI engineers to restrict use of coding assistants like Claude Code and Codex.
They're worried that working with these tools could inadvertently cause distillation of proprietary capabilities from Anthropic's and OpenAI's models.
It's one of the first times an AI company has restricted employee access to rival AI coding assistants to prevent model distillation.
Model distillation and cloning risks
Model distillation is when a smaller AI learns by copying a bigger, smarter one.
While this makes AIs faster and more flexible, it also means rivals could end up cloning powerful features without ever seeing the original code.
Meta's move shows just how seriously tech companies are taking the need to guard their innovations as AI tools get more popular in the workplace.