News for hackers. Stuff that matters.

A comment on Letting Claude Play Text Adventures. Parent comment

Re:Letting Claude Play Text Adventures by moyix on Wednesday January 21, 2026 @05:11PM (#46712331)
Also, unlike OpenAI, Anthropic's prompt caching is explicit (you set up to 4 cache "breakpoints"), meaning if you don't implement caching then you don't benefit from it.
[ Parent | on HN ]
Re:Letting Claude Play Text Adventures by netcraft on Wednesday January 21, 2026 @05:47PM (#46712746)
thats a very generous way of putting it. Anthropic's prompt caching is actively hostile and very difficult to implement properly.
[ Parent | on HN ]