Abundant resources but unable to self-detect? Benchmark partner questions Anthropic's push to restrict model distillation

Odaily Planet Daily News Benchmark partner Chetan Puttagunta posted on platform X, expressing confusion over Anthropic's public call for stricter regulation of AI model distillation. He stated that Anthropic is currently a company valued at approximately $1 trillion with substantial technical and financial resources. At such a scale, what they describe as a "large-scale distillation attack" should theoretically be relatively easy to identify and trace. If Anthropic chooses to restrict related activities, the real cost may not be a lack of technical capability, but rather the forfeiture of some API revenue. "The only cost that seems to be required is reducing revenue from related API businesses."

Previously, Anthropic and other AI companies have been closely focusing on the issue of model distillation. Model distillation typically refers to using the output of a large model (teacher model) to train another model (student model) in order to reduce costs and improve efficiency. Some AI companies are concerned that competitors might excessively call APIs to obtain model outputs for training their own models, thereby bypassing the original research and development investment. Puttagunta's perspective suggests that the controversy surrounding model distillation in the AI industry essentially involves a balance between commercial interests, open competition, and intellectual property protection. For leading AI companies, finding the equilibrium between protecting core technology and maintaining an open ecosystem will become a crucial issue in future industry competition.