Remix.run Logo
mannanj 8 hours ago

so basically the anthropic employee who responded says those 1h caches were writes were almost never accessed, so a silent 5m cache change is for our best interest and saves cost. (justifying why they did this silently)

however his response gaslights us because in the OPs opening post his math demonstrates this is not true, it shows reads 26x more so at least in his case the cache is not doing what the anthropic employee describes.

clearly we are being charged for less optimization here and being given the message (from my perspective by anthropic) that if you are in a special situation your needs don't matter and we will close your thread without really listening.

hirako2000 8 hours ago | parent | next [-]

What also gives it away is the refusal to at least expose this TTL via parameter. In the same sentence as informing the 5m won't change since it's your interest.

It's also in the interest of the users to keep certain params private, we are meant to deduce that. Did you not ?

SkyPuncher 7 hours ago | parent | prev | next [-]

My suspicion is the have an overall fixed cache size that dumps the oldest records. They’re now overflowing with usage and consistently dumping fresh caches.

During core US business hours, I have to actively keep a session going or I risk a massive jump in usage while the entire thread rebuilds. During weekend or off-hours, I never see the crazy jumps in usage - even if I let threads sit stale.

love2read 8 hours ago | parent | prev [-]

> that if you are in a special situation your needs don't matter and we will close your thread without really listening.

Are there any other $50B+ Valuation companies that care about special situations? If so, who?