Skip to content
GENAIRALISTGENAIRALISTby Andrejs Karpovs

Manifesto

01 — 2026

A position for people who intend to participate in what comes next.

We will not wait to become obsolete.

The GenAIralist is a stance more than a title: keep moving, keep connecting, use AI without surrendering judgment, and act while the choice is still yours.

Begin with the shift

The shift

Five desks are becoming one.

For years, the career advice was simple: narrow your lane, deepen it, defend it. AI compresses much of the execution that once made a narrow specialty valuable, and the work that remains recombines into wider, more connected roles.

AI compresses the repeatable.

Human capacities compound.

  • First drafts— compressed by AI
  • Direction— stays human, compounds
  • Routine analysis— compressed by AI
  • Documentation— compressed by AI
  • Judgment— stays human, compounds
  • Dashboards— compressed by AI
  • Synthesis— stays human, compounds
  • Research passes— compressed by AI
  • Trust— stays human, compounds
  • Accountability— stays human, compounds

The role gets wider as the tools get stronger.

Waiting defends a résumé that is already aging. The move is to expand — across disciplines, through ambiguity, into the work where judgment decides the outcome.

And you, reader

You’ve done the courses.

You can prompt. You’ve built a GPT or two.

And nothing about your position has changed.

AI fluency is already table stakes.

Once fluency becomes expected, it will differentiate you about as much as “proficient in Excel.”

What differentiates you is what fluency lets you reach: domains that used to be closed to you.

06

Six principles. One evolving practice.

Together they form a practical stance for working across domains, using AI with intent, and keeping judgment in human hands.

01

Move first.

Agency over dependency

We don’t wait for institutions, governments, or employers to solve this for us. Universal basic income may come and retraining programs may help, but neither is a plan we control. What we control is what we learn, build, and ship — so that is where we start.

02

Connect the dots.

Breadth is the new depth

The deepest expertise still earns its keep. What changed is where the leverage sits: a question that once crossed five desks — research, strategy, design, engineering, operations — can now be read, prototyped, and shipped by one person who connects those domains. Range, connected with intent, becomes synthesis.

03

Keep judgment human.

Human skills are a durable edge

The faster the tools get, the more the outcome depends on what surrounds them: judgment under ambiguity, leadership across functions, trust that took years to earn, the sense for which question to ask. None of it ships with the model.

04

Build with it.

AI is a power tool

We treat AI the way a builder treats power tools: it changes what one person can make in a day, and it rewards the hand that already knows what good looks like. We use it to prototype, build, and orchestrate — and we keep the decisions about what is worth building.

05

Stay in motion.

Evolution is a practice

Range is built like fitness: repeatedly, in public, against resistance. A new domain every season, connections made before they are needed, the recurring discomfort of being a beginner. The fellowship exists to make that pace sustainable.

06

Choose meaningful work.

Purpose requires participation

For us, work carries identity, contribution, and meaning — more than a paycheck. The point of handing AI the repeatable parts is to spend more of a career on problems that deserve a person’s full attention. That is the work we intend to keep.

The stack

Four capacities. An order to build in.

An editorial sequence rather than a measured ranking: start with AI nativeness, then build range, judgment, and a position the market can understand.

The position holds only in practice.

The fellowship is where the practice starts: a small group of leaders and builders doing real work outside their home disciplines, with honest feedback about what’s weak.

Join the waitlist →