3 ms·
Tool calls (particularly fetching for context) eats the context window heavily. I explicitly send MCP calls to sub agents because they are so “wordy”.
by master_crab 6mo ago
Tool calls (particularly fetching for context) eats the context window heavily. I explicitly send MCP calls to sub agents because they are so “wordy”.
- bensyverson 6mo agoEveryone who has not hit this bug thinks it’s user error… It’s not. It happened to me a few days ago, and the speed at which I tore through my 5 hour usage cap was easily 10x faster than normal. Also: sub agents do not get you free usage. They just protect your main context window.
- piva00 6mo agoDon't they consume less of the token quota in case the subagents are running cheaper models like Sonnet and Haiku compared to Opus?
- bensyverson 6mo agoCorrect—I just wouldn't want folks to mistakenly think that the context fill % corresponds 1:1 with session token use.
- master_crab 6mo agoYes, sorry. I meant it more as a descriptor of how many tokens it consumes. You are still stuck burning money.
- dmd 6mo agoI'm on Max. This morning, just to test, before doing anything else whatsoever, I was at 0%, and I typed 'test one two three' into CC. That put me at 12%. I have no MCPs except the built in claude-in-chrome. This is clearly a bug.
- cyanydeez 6mo agoReadimg through this thread, it seems likely is a KV cache "bug". Theyre likely doing too many evictions of the LLM cache so the context is being reloaded to often. Its a "bug" because its probably an intended effect of capturing the costs of compute but surfacing a fact that they oversold compute to a situations where they cant keep the KV cache hot and now its thrashing.
- bensyverson 6mo agoCaching helps them too, so I hope they fix it