> so from my perspective I don't see why we should be skeptical that frontier LLMs guided by experienced devs can translate whole libraries like that

There's a laundry list of reasons to be skeptical about LLM results in general. The biggest one is that LLMs are excellent at creating output that looks right but is nevertheless still wrong.

In this particular case, there's also a specific reason: This is obviously a marketing stunt regardless of whether the claimed results hold up. Anthropic has a long history of just lying and making shit up, so any results are subject to extra scrutiny.