Questions I'm Exploring
The questions driving the work
Frameworks and systems are answers in progress. These are the questions underneath them - the ones worth a decade of attention.
- 01
Security Governance
Can AI become genuinely accountable?
Accountability requires someone who can explain, defend, and own a decision. What has to be true of an AI system - its reasoning, its records, its oversight - before that accountability is real rather than theater?
- 02
Future Systems
Why do well-designed systems still fail under complexity?
Most failures aren't from bad components but from interactions no one designed. How do you architect for the failure modes that only emerge at scale?
- 03
Context Intelligence
Can context become infrastructure?
Organizations treat context as disposable - trapped in chat logs, lost when people leave. What changes when we treat the knowledge behind decisions as durable, governed, queryable infrastructure?
- 04
Founding Architect
How do you scale a founder's judgment without diluting it?
A founding team's intuition is an organization's edge and its bottleneck. Can that judgment be made into a system others can apply - without flattening it into rules that miss the point?
- 05
Enterprise AI
Can machine reasoning be made legible to the people it serves?
If a decision-maker can't inspect why a system concluded what it did, they can't truly own the outcome. What interface makes reasoning inspectable without drowning the user in detail?
- 06
Decision Architecture
Can organizations preserve judgment as they scale AI?
Every layer of automation is a chance to outsource a decision no one will later understand. Is there an architecture that lets organizations move faster without quietly losing the judgment that made them good?
- 07
Context Intelligence
Who should own an organization's context?
When context lives in a vendor's model, the organization rents its own memory. What does context sovereignty look like, and what does it cost to keep it?
- 08
Future Systems
What survives when every tool changes?
The model, the framework, the vendor - all of it will be replaced. What is the durable layer worth designing carefully because it has to outlast a decade of churn?
Have a question worth exploring?
The best questions come from people building hard things. If one of these resonates - or you have a better one - I'd like to hear it.