Plutonus

Approach AI-native, deliberately directed

Premise 01 / Responsibility

AI-native, deliberately directed

More reach. Clear responsibility.

AI systems are Plutonus’s primary production environment. They extend what one studio can research, design, build, test and produce. I set the direction, define the constraints, review the work and remain responsible for the result.

The tools may change. That division of responsibility should not.

How it works 02 / Reach + responsibility

AI adds reach. Direction stays human.

01 / Reach

More within reach

AI makes it practical for one studio to research, prototype, build and test across a wider range of work.

02 / Control

Decisions stay human

One person remains responsible for the goals, constraints, task scope, acceptance and outcome.

Control points 03 / What stays deliberate

Generated work is not treated as automatically correct.

Five points keep the work directed, reviewed and accepted by one person.

  1. 01

    Direction

    Set the goal, relevant context and constraints before the work begins.

  2. 02

    Boundaries

    Give each task a clear scope: what it may change and what it must leave alone.

  3. 03

    Review

    Inspect decisions and changes; correct drift before it becomes the product.

  4. 04

    Evidence

    Test or validate in proportion to risk, complexity and expected life.

  5. 05

    Decision

    One human decides what is accepted, what is delivered and what still needs work.

Across the studio 04 / Different work

One approach. Different kinds of work.

Software, websites, internal tools and procedural visual work require different methods. In each case, one person sets the direction, reviews the work and owns the result.

  • 01

    Software & product builds

  • 02

    Websites & digital experiences

  • 03

    Internal tools & automation

  • 04

    Procedural / AI-native motion & visuals

Tools in use 05 / Current examples

The tools change with the work.

The studio uses general AI systems, code-focused agents, specialist models and its own tools. The mix changes with the project.

Current examples
  • ChatGPT

    research, planning, review and synthesis

  • Codex

    implementation and testing inside the codebase

  • Specialist and local models

    focused production tasks

  • Studio-built tools

    controlled access to development context, focused change review and media production

When existing tools get in the way, the studio may build the missing piece.

Responsibility 06 / Outcome

AI participates in the work. It does not own the outcome.

I decide what is accepted, what must be corrected, what evidence is enough and what is delivered.