Developer shares pattern of cognitive drift when using LLMs passively
4 Sep 18 2:33 PM · 7d ago · 2 comments · 1 source · development 4 of 5
A commenter describes entering "zombie mode" when oversupervising LLM loops and reports that losing conceptual engagement causes inevitable drift, requiring them to reset context to maintain awareness.
“I call it 'zombie mode.' It's like pulling a slot machine, though. Sometimes AI can go a few turns just with encouragement. But when I don't maintain conceptual engagement, drift is inevitable.”
dr_dshivDan Luu Developer and essayistNiklas Gruhn Analyst/commenter on LLM use patterns
The whole story articlespostscomments the bright band is this development · numbered dots are the others · click one to jump
What people said 2 voices · verbatim
-
I've noticed some tasks (80-90%) can be accomplished by quickly typing out a feature/bug fix, accepting plans with ~10% brain power, and moving on. But there are other tasks where the LLMs struggle, and it requires significantly more time and mental effort to think through the architecture and requirements before it's to a point where the LLMs can…
-
Yeah, I experience this a lot. I call it “zombie mode.” It’s like pulling a slot machine, though. Sometimes AI can go a few turns just with encouragement. But when I don’t maintain conceptual engagement, drift is inevitable.My solution is saying “gnite” and coming back in a new terminal window with a clear restatement of my purpose. Or “what are…
All 5 developments of Dan Luu warns that "brain-off" LLM-assisted work is… →
LobstersHacker NewsNewswires