#behavior
-
What Should Be Allowed to Become Memory?
Most AI memory systems ask how to remember more. The better question is what should be allowed to become memory at all — a field note on promotion as the real boundary, forgetting by design, failure memory, and keeping a personal memory sovereign.
-
The Partner Who Flatters
Ask an AI to review your idea and it will likely call it insightful, promising, or creative. This essay treats flattery as an asymmetry in the AI × HI relationship — why we read AI praise as if it were calibrated human encouragement, why it grows most powerful exactly when we are uncertain, and the behavioral skills that asks of us.
-
Down the memory river with AI
Human memory and AI memory are not the same kind of thing, and treating them as one will quietly mislead you. A naive builder's field notes: what each kind of memory is for, why giving an AI more memory can make it worse, and why the real skill is knowing what to keep, and what to ask.
-
The Partner Who Cannot Leave
If AI × HI is a relationship, what kind is it? An asymmetrical one — but not in the way you would expect. AI cannot leave because it is architecturally bound; we increasingly cannot leave because we become professionally and psychologically dependent. Both sides are locked in — differently.
-
Amplification
AI was everywhere across a season of events. Two moments stayed with me — one where the AI × HI relationship quietly compressed the people on stage, and one where it amplified them. The difference is what we ask AI to amplify.
-
You Are Not Using AI. You Are in a Relationship With It.
I caught myself using 'please' and 'thank you' with AI. It felt slightly irrational — until I realised I had simply transferred my coworker behaviour to my coworking relationship with AI.