I wouldn't be too fixated on the specific numbers in that post.
replyAnthropic was extremely capacity constrained at that point. They still are but not to that extent.
I'd note that OpenAI offers 24 hour caching. I'd be surprised if Anthropic hasn't optimised their caching for Claude code too.
SemiAnalysis recently posted that their actual Opus usage works out at $0.99 because of caching.
The principles remain though.
Recently I started getting messages from Clause Code (on a plan). "You're restoring an old session are you sure you don't want to compress the context? This will use a substantial amount of your usage quota"
replySo it seems they do care.
That's exactly what I said. They do care when FLOPs are involved. Restoring an old session with 900k tokens will require a lot of FLOPs to reprocess the 900k token.
replyMeanwhile, they don't really care if you use hundreds of millions of cached input tokens, which doesn't consume any FLOP.