Anthropic backtracks on policy that 'sabotaged' researchers' work
Anthropic is reversing a policy that secretly degraded performance for researchers using its Claude Fable 5 model to build competing AI systems, the company told Wired. The restrictions, which were no…