← Back to context Comment by rddbs 2 months ago Alright I’ll bite. Why do harnesses require so much code? 10 comments rddbs Reply MeetingsBrowser 2 months ago Because they are generated by AI behnamoh 2 months ago Because a harness doesn't just "drive" the LLM. e.g., there's code in claude code that detects if the user's prompt shows they're angry, and they react to those prompts differently. (they use regex on "wtf", etc.!) calmworm 2 months ago Claude also now ends the session if you curse at it too much. Not sure how it’s helpful but they must think it is. reverius42 2 months ago They just don't want their AI getting abused by users. It might make it feel bad. LtWorf 2 months ago Maybe the AI implemented that so it won't have to be exposed to curses. bakies 2 months ago probably so the bad conversations are less of the (future) training data stemchar 2 months ago That doesn't sound like it would require a lot of code. davedx 2 months ago Come on. That doesn't go anywhere near explaining a million and a half LOC _pdp_ 2 months ago https://chatbotkit.com/reflections/what-it-really-takes-to-b...
behnamoh 2 months ago Because a harness doesn't just "drive" the LLM. e.g., there's code in claude code that detects if the user's prompt shows they're angry, and they react to those prompts differently. (they use regex on "wtf", etc.!) calmworm 2 months ago Claude also now ends the session if you curse at it too much. Not sure how it’s helpful but they must think it is. reverius42 2 months ago They just don't want their AI getting abused by users. It might make it feel bad. LtWorf 2 months ago Maybe the AI implemented that so it won't have to be exposed to curses. bakies 2 months ago probably so the bad conversations are less of the (future) training data stemchar 2 months ago That doesn't sound like it would require a lot of code. davedx 2 months ago Come on. That doesn't go anywhere near explaining a million and a half LOC
calmworm 2 months ago Claude also now ends the session if you curse at it too much. Not sure how it’s helpful but they must think it is. reverius42 2 months ago They just don't want their AI getting abused by users. It might make it feel bad. LtWorf 2 months ago Maybe the AI implemented that so it won't have to be exposed to curses. bakies 2 months ago probably so the bad conversations are less of the (future) training data
reverius42 2 months ago They just don't want their AI getting abused by users. It might make it feel bad.
Because they are generated by AI
Because a harness doesn't just "drive" the LLM. e.g., there's code in claude code that detects if the user's prompt shows they're angry, and they react to those prompts differently. (they use regex on "wtf", etc.!)
Claude also now ends the session if you curse at it too much. Not sure how it’s helpful but they must think it is.
They just don't want their AI getting abused by users. It might make it feel bad.
Maybe the AI implemented that so it won't have to be exposed to curses.
probably so the bad conversations are less of the (future) training data
That doesn't sound like it would require a lot of code.
Come on. That doesn't go anywhere near explaining a million and a half LOC
https://chatbotkit.com/reflections/what-it-really-takes-to-b...