Judgment Is Uniquely Human
Human judgment should scale with machine intelligence.
Philosophy
Human Agency Is the Frontier.
Human agency must advance with intelligence. This is the central premise connecting our mission, our vision, and everything we build.
Foundational belief
Greater machine capability does not automatically increase human agency.
AI is an extraordinary advance in human capability. As it becomes more capable, autonomous, and embedded in everyday life, the ability of people and organizations to understand, direct, and constrain its influence matters more and more.
That ability has to be deliberately supported. Rather than competing to build more powerful intelligence, we build technologies that give people greater agency in their relationships with intelligence.
- Mission
- To build technologies that ensure balance between human agency and artificial intelligence, empowering people and organizations with greater judgment, clarity, and autonomy.
- Vision
- A future where artificial intelligence advances humanity without diminishing our freedom to think, choose, and determine our own direction.
The imbalance
The ability to build intelligence and the ability to shape its effects are not distributed equally.
Power has historically been concentrated through governments, institutions, capital, infrastructure, and control over information. Advanced AI introduces another increasingly consequential concentration of power.
We respect the achievements of the organizations building frontier AI. Our response is not to oppose AI advancement. It is to build complementary technologies that give individuals and organizations more meaningful influence over how AI affects them.
The future doesn’t need less intelligence. It needs more human agency.
The human–AI relationshipHuman in the loop is supervision.
Human in the loop is supervision.
Human defining the loop is agency.
We don’t advocate constant human intervention. We advocate meaningful human authority over intentions, boundaries, and outcomes.
- Person
- AI
- The work loop
- Hover a stage to see it run
- Stage 1 · 2025 and prior
Human as Operator
Humans initiate tasks. AI assists with execution. The human stays responsible for most decisions.
- Stage 2 · 2026
Human as Supervisor
AI performs complex tasks autonomously. Humans monitor outputs and intervene: the human-in-the-loop model.
- Stage 3 · 2027 and beyond
Human as Architect
Humans define the intentions, principles, priorities, and boundaries, and AI acts within them. When intent changes, the loop adjusts in real time.
Supervision asks
Did the AI do the right thing?
Agency asks
Who gets to define what the right thing is?
The gap
Execution is scaling faster than judgment.
Autonomous systems act continuously. The decisions meant to steer them still travel at the speed of meetings, memos, and release cycles.
- Benefit leadership is aiming for
- Benefit the agent delivers
- Benefit lost to the gap
Stage 2 · Judgment on a delay Off course: 4 months
Delivered
Lost to the gap
Net negative: the gap outweighs everything the agent delivered.
Stage 3 · Judgment in real time Off course: 0 months
Delivered
Lost to the gap
Net positive: nearly all of the intended benefit, delivered.
With judgment on a delay, this agent spends 4 of its 8 months working in a direction nobody wants anymore, and the benefit lost to that gap outweighs everything it delivers. Busy, fast, and net negative. That’s why human agency has to live in definition, alignment, and governance, not at a checkpoint. AI should serve the current intentions of the people it acts for.
Human agency
Judgment. Clarity. Autonomy.
Not categories for organizing products. Together, they form our working model of meaningful human agency.
01 / Human intent
Judgment
Preserve human intent as intelligence becomes autonomous.
Who determines what should happen? Why judgment matters
02 / Understanding
Clarity
Make increasingly complex information environments understandable.
Can I understand what’s happening? Why clarity matters
03 / Meaningful choice
Autonomy
Give people greater control over their boundaries and choices.
What can I choose or refuse? Why autonomy matters
Judgment without clarity risks decisions based on incomplete or misleading understanding.
Clarity without autonomy lets people understand situations they may lack the ability to influence.
Autonomy without judgment creates choices without enough direction.
Values
Five values.
Clarity Creates Power
Clarity is the foundation of independent thought.
Autonomy Is Essential
Privacy is a condition of freedom.
Technology Must Serve Humanity
Progress should be measured by human empowerment.
Challenge the Default
Just because technology makes something possible doesn’t mean society must accept it.
As intelligence advances, we give people more power to understand, decide, and shape their own futures.
A better balance
We’re defined by what we enable, not what we oppose.
We build ventures and technologies that make human intentions more influential, information more understandable, and personal boundaries more enforceable.