AI Product Management
People Don't Fear AI. They Fear Losing Agency.
AI products earn trust by giving people control, evidence, and a way to recover. How to turn those requirements into product decisions.
[ Essays & experiments ] // AI product leadership
The work starts after the demo.
AI product leadership: turning model capability into dependable, trusted products. Writing on AI product management, enterprise AI, agents and evaluation.
Read the writing→Essays010
Topics008
Last signal2026.08.23
StatusLive
01
Featured
Editor's pick- 001Write More Persuasively with ChatGPTEnterprise AI2026.08.23 · 2′
- 002Delete the Word 'AI'AI Product Management2026.06.30 · 2′
- 003Most Things You Could Build Aren't Worth BuildingAI Agents2026.05.20 · 3′
- 004Scaling Qualitative Research with AIexperiments2025.12.09 · 3′
- 005You Need TasteNotes2025.12.08 · 4′
- 006LLM-Expect: Testing LLM Functions in Pythonexperiments2025.11.24 · 3′
- 007GRPO, Step by StepNotes2025.11.15 · 8′
- 008AI Bias: Where It Enters and How to Test ItNotes2025.11.15 · 4′
Enterprise AI
Write More Persuasively with ChatGPT
AI Product Management
Delete the Word 'AI'
AI Agents
Most Things You Could Build Aren't Worth Building
experiments
Scaling Qualitative Research with AI
Notes
You Need Taste
experiments
LLM-Expect: Testing LLM Functions in Python
Notes
GRPO, Step by Step
Notes
AI Bias: Where It Enters and How to Test It
03 Search
Explore the essays.
Search by meaning to find relevant passages. Open the essays here, or continue your question in Claude or ChatGPT.
03 Explore the essays
Make systems explainable.
If people cannot understand why they got a particular output, they cannot trust it. Expose the key factors, constraints, and signals that shaped the result, including any limitations or biases in the underlying data or model.
Reduce the surface area for surprise.
AI should behave in stable, expected ways. Avoid unnecessary randomness, hidden states, or behavior that changes without clear cause.
Favor clarity over capability.
A simpler model or workflow that’s understandable is often better than a more capable one that’s opaque or fragile.
Constrain systems to well-defined boundaries.
Clear limits create predictable behavior. Define what the system should do, and just as importantly, what it should never do.
Evaluate continuously, not only at launch.
Behavior shifts over time. Evaluation should be continuous, using regression tests, structured prompts, and real user feedback. Assess for bias, uneven performance across groups, and drift that could introduce harmful patterns.
Design features around real human expectations.
Technical correctness isn’t enough. The system’s behavior must match what users think it will do.
Make failures observable, recoverable, and safe.
Assume failures will occur. Design workflows where errors are visible, explanations are meaningful, and recovery is simple. Treat biased or uneven behavior as a system failure that is discoverable, correctable, and never silent.
Prioritize meaningful user control.
Users should be able to steer, pause, override, or refine the system. Autonomy without control is not helpful.
Build with careful iteration, not leaps of faith.
Start with restricted capabilities, test assumptions early, expand carefully, and monitor the system as it grows.
Document decisions, not just code or models.
A human-centered system is as much about why choices were made as how they were implemented. Write down rationale, trade-offs, evaluation results, and constraints.