Vision
Why AI changes the nature of work, and how we approach the transformation.
The Objective
Build an organization where AI is the first-class resource — not an optional tool. Where every role is designed assuming AI exists, and its potential is evaluated for every task. Where humans direct and systems execute.
This is not an initiative. It's a structural shift in how work gets done.
Why Now
Traditional SaaS: ~US$125K revenue/employee. AI-native startups: US$2M–$3.5M. Outliers in a bubble — but the directional signal is consistent. AI-native models produce asymmetric output per person.
AI-native SaaS trades at ~25× revenue vs 2.5–7× for traditional SaaS. Markets are pricing a structural advantage into AI-native models.
Teams operating in autonomous production mode — specs and results instead of sprints and standups — are multiplying. Each model generation accelerates them further. The gap between early movers and the rest is compounding.
The labor market is already responding: job-finding rates for 22–25 year-olds in AI-exposed occupations dropped ~14% after late 2022, while less-exposed occupations held steady. Not mass unemployment — a quiet repricing of which skills get hired.
Where Value Migrates
Technology doesn't eliminate value. It reassigns it to the scarcest layer.
- •Content, code, and analysis production
- •Implementation of known solutions
- •Iteration speed
- •Information gathering and synthesis
- •Problem selection — choosing what to solve
- •Strategic framing — defining constraints and positioning
- •Proprietary data — the assets models don't have
- •Distribution leverage — the ability to reach the market
- •Brand authority — accumulated trust
- •Risk ownership — bearing the consequences
Scarcity used to be labor. Now it's taste, trust, signal, distribution, decision rights.
AI reduces the cost of answers. Money moves to those who define the questions.
This pattern is consistent across previous technology shifts. The specific categories will evolve, but the direction of migration is well-established.
What This Means
Work becomes more interesting. The repetitive, predictable, mechanical tasks are the ones AI takes over. What remains is judgment, creativity, strategy, human relationships.
Skills become more valuable. Someone who knows how to direct AI, who can specify what they want, evaluate what they get, and build systems that work, is significantly more valuable in the job market.
Impact multiplies. With AI integrated into workflows, one person can do what used to take a team. Individual contribution has a disproportionate effect.
The Human Role
Five functions remain irreplaceable in an AI-native organization:
Direction
Choosing which problems to solve. Defining positioning, constraints, risk tolerance.
Judgment
When to trust data vs. ignore it. Ethical boundaries. Tradeoffs between brand, revenue, and long-term trust.
Taste
What rings true. What's original vs. derivative. What's aligned with identity.
Relationship
Earning trust. Managing tension. Reading subtext.
Accountability
Owning it when results fail. Making irreversible decisions.
The human role becomes architect, editor, risk manager. Not operator.
Accountability evolves with the model. When humans no longer read the code or review every artifact, accountability shifts from "I checked every line" to "I designed the system that validates the result." An engineer is accountable for their agents the way a manager is accountable for their team — not by doing the work themselves, but by defining the constraints, scenarios, and escalation thresholds that guarantee quality. Demanding both 10x velocity and accountability at the same inspection point as before is incompatible. Accountability moves to the process: spec design, scenario quality, failure detection systems.
Let's Be Honest
This is a reskilling. The nature of work is changing. Existing skills don't disappear — they become the foundation. But an expert who refuses AI finds themselves competing against someone who has mastered it. Adaptation is a condition of the market, not a policy choice.
This isn't about factory work. The most AI-exposed workers earn 47% more than unexposed workers and are nearly 4× more likely to hold graduate degrees. Knowledge work is the primary target.
Results first. Evaluation is primarily based on what is built. Not just on enthusiasm, the number of prompts, or attitude toward change.
Status doesn't diminish — it gets redefined. If a professional identity is tied to a task that AI can now do, that doesn't mean the person is worth less. It means that task wasn't worthy of what they can actually contribute.
The discomfort of change is temporary. The cost of standing still compounds.
The Approach
The operational principle is simple: replace "human produces" with "human defines specs → system produces."
The approach requires every person to look at their work honestly and ask: "If AI had existed when this role was designed, would it have been designed the same way?"
The Target End State
The following describes the operating model we are building toward. The path will adapt; the direction is deliberate.
An organization where autonomous AI agents make routine operational decisions and execute without constant oversight. Where humans define objectives, constraints, and escalation thresholds — not steps. Where people spend their time on judgment, strategy, and relationships — not on repetitive execution. Where output per person is significantly higher because agents handle both mechanical work and routine decisions. Where a new employee, from their first day, works with AI agents as a first-class resource — not as an option.
The organizational structure transforms: the old pyramid of executors gives way to a diamond of thinkers with an AI core. Fewer humans, high caliber, high context, high authority. The role evolution patterns describe the structural forces — convergence, specialization, elevation, absorption, emergence — that drive this transition.
Guiding Principles
- Automate before hiring. Systems before manual processes.
- Evaluate results, not activity. What is being produced? What is being optimized? What is being changed?
- Transformation is a strategic necessity. The market dynamics described above are external. Our response to them is a choice — one we've made deliberately.
← Back to home · The reference framework · The AI Lab · Glossary