Building useful
AI systems.
Ideas become working experiments—combining models, software, evidence, and human judgment.
MODEL / 0.82HUMAN / READYLive systems / 03
Currently in the Lab
Active experiments change as questions sharpen, assumptions break, and better architectures emerge.
Research fields
Five areas. One system view.
The lab moves across domains, but the method stays consistent: make the intelligence visible, testable, and accountable.
Research fields
Explore by area
AI Agents
Systems that reason, use tools, maintain state, and verify outcomes.
Healthcare AI
Evidence-centered intelligence designed for accountable human validation.
Human + AI
Experiences that amplify agency rather than automate ownership.
Applied AI
Structured intelligence for practical, physical-world problems.
AI Infrastructure
Model routing, providers, state, observability, and evaluation.
Operating principles
How H2 builds
The model is only one component of the system.
Human Agency First
AI should expand human capability without silently taking control.
Evidence Over Assertion
Important outputs should make their supporting evidence visible.
AI Proposes. Humans Control.
Human approval remains available when consequences matter.
Verify Outcomes
Completing an action does not guarantee the intended result.
AI Is Part of the System
Useful products combine models with software, rules, tools, data, and people.
Learning system
The AI Builder Journey
Progressing from using models to designing complete, accountable systems.
systems
Model + software + tools + data + people + verification
Field notes
Latest Lab Notes
Sep 18, 2026 · 6 min
Why AI Agents Need Verification, Not Just Actions
An action is only useful when the intended outcome can be confirmed.
Sep 04, 2026 · 7 min
Human-in-the-Loop Is More Than an Approval Button
Designing meaningful control before, during, and after an AI recommendation.
Aug 21, 2026 · 5 min
Why Deterministic Logic Still Matters in AI Systems
The strongest AI applications know which decisions should not be probabilistic.
Why H2?
H2 AI Labs is where ideas become AI experiments, prototypes, and useful systems.
I learn AI by building systems around it—combining software, data, tools, evidence, human judgment, action, and verification.
More about H2