Is Anthropic's acquisition of Bun alone still not enough to infer their tech stack? What more obvious signals would be needed?

Also, honestly, given the speed constraints of large models, it makes almost no difference what language an agent is written in. The small performance differences between programming languages do not even begin to matter compared with network latency, let alone the speed at which a large model streams tokens.

Why do you think they acquired Bun. To look at it?

Also you needed a leak to know that CC, that is out for 1 year, is in TS?