Rosenfeld Review Podcast Designing Assistant Technology with Chris Noessel
Jan 21, 2026
In this engaging conversation, Chris Noessel, a design practitioner and author known for his insights into interaction and assistant technology, dives into the complexities of AI assistants versus agents. He explains how assistants help us, while agents act on our behalf, and warns of 'cognitive debt' from over-relying on them. He highlights practical uses of AI, discusses its impact on skills, and shares pop culture references to illustrate his points. Chris advocates for responsible design as crucial in shaping how AI can truly enhance our intelligence.
AI Snips
Chapters
Books
Transcript
Episode notes
Assistants Versus Agents
- An assistant helps a user do something while an agent does the thing for them.
- Locus of attention determines whether a feature acts assistively or agentively in practice.
Mode Changes With User Attention
- The same feature can be assistive or agentive depending on whether the user attends to it.
- Predictive text is assistive when users use suggestions and agentive when it autocorrects without attention.
Inventory And Maps As Assistive AI
- Chris describes using AI for large-scale inventory prediction at IBM to guide human decisions.
- He contrasts that with ETA regression in mapping as a subtle but powerful assistant example.
