Thoughts in progress.

Notes on building, technology, and the human questions inside both.

09 writings
When LLMs spread like an ideaThe virus metaphor is useful only if we keep it a metaphor: it describes patterns of copying, not biology or destiny. AICultureNetworks 1 min The risk of outsourcing the first thoughtConvenience can remove useful friction. The goal is to use assistance without losing the muscles that make judgment possible. AILearningJudgment 1 min When video stops being evidence by defaultHyperreal digital video changes a quiet social rule: seeing a clip may no longer be enough to settle what happened. MediaTrustAI 1 min What AI companionship is trying to fillPeople do not turn to a companion system for one reason. The appeal can include practice, comfort, play, privacy, or simply a response at a lonely hour. AICompanionshipDesign ethics 1 min When care meets a retention metricA companion app can be useful and still face an awkward question: what happens when emotional dependence is good for the business? BusinessAIDesign ethics 1 min The future of AI influencers is a trust design problemSynthetic creators may make culture more participatory. They also make disclosure, authorship, and accountability much more important. AIMediaIdentity 1 min