[Anthropic Reveals Claude Has Contributed to Over a Quarter of Internal AI R&D Work] According to Bloomberg, Anthropic disclosed in its latest report that its large model assistant, Claude, has "led" approximately 26% of the company's AI R&D work, a rapid increase from nearly zero at the beginning of the year, and has participated in about 90% of employees' daily R&D collaborations. Anthropic also stated that at any given time, there are over 30,000 intelligent agents on its platform performing research and engineering tasks. The company plans to introduce multiple independent third-party evaluation organizations to be "embedded" into its internal systems, granting them access to data and processes similar to that of employees, in order to track AI capability evolution and safety risks. Their calculations indicate that approximately 6%–12% of computational resources in various types of research are allocated to safety monitoring and evaluation.