3 ms·
Some of this stuff is ludicrous. I finally tried /loop last week and discovered every loop iteration passes the entire context history to the model. So pretty q
by svachalek 3mo ago
Some of this stuff is ludicrous. I finally tried /loop last week and discovered every loop iteration passes the entire context history to the model. So pretty quickly you're running a full 1M context window, without cache, likely just to check if something is ready or needs to be done. It's miserably terrible engineering unless your entire and only goal is to burn tokens.
- Wowfunhappy 3mo agoWait, why isn't context cached with /loop?
- ihsw 3mo ago[dead]
- pzh 3mo agoBecause it's typically cached for 5min (1hr is a setting you have to explicitly opt into), and very few people run loops at a cadence of < 5 mins.
- Wowfunhappy 3mo agoOh! I thought Anthropic cached for one hour (by default), am I wrong about that? Or is this an OpenAI thing?
- poly2it 3mo agoAnthropic changed their cache duration for some reason a while back.
- EliasWatson 3mo agohttps://code.claude.com/docs/en/prompt-caching#on-a-claude-subscription https://code.claude.com/docs/en/prompt-caching#on-a-claude-s... "On a Claude subscription, Claude Code requests the one-hour TTL automatically. Usage is included in your plan rather than billed per token, so the longer TTL costs you nothing extra and only affects how long your cache stays warm. If you’ve gone over your plan’s usage limit and Claude Code is drawing on usage credits, you are billed for that usage, so Claude Code automatically drops the TTL to five minutes."