I've turned Claude into a BEAST ! 🤣
The biggest problem with all AI and LLMs is it tells you what you want to hear, and is optimized for the most pleasing response by default.
Aside from using Claude as a glorified EA, that's often useless, and in many cases dangerous.
Claude now challenges me, bosses me around, holds me to account, names unproductive patterns of behavior, brings critical-path activities to my attention that are blocking other work and proactively tells me if I'm focusing on a low value activity relative to my stated goals.
As a result it is 3x more useful and 10x less dangerous!
The biggest problem with all AI and LLMs is it tells you what you want to hear, and is optimized for the most pleasing response by default.
Aside from using Claude as a glorified EA, that's often useless, and in many cases dangerous.
Claude now challenges me, bosses me around, holds me to account, names unproductive patterns of behavior, brings critical-path activities to my attention that are blocking other work and proactively tells me if I'm focusing on a low value activity relative to my stated goals.
As a result it is 3x more useful and 10x less dangerous!
2❤️3❤️1
