DREWSKY / FIELD LAB First principles · LLMs · Agent-native tools
Drewsky / field lab

SMALL EXPERIMENTS / OPEN QUESTIONS

Think from the roots.
Build from there.

A field lab for first-principles thinking, LLMs, and tools that extend what we can make. Start with a question small enough to test. Leave behind a method someone else can try.

Try the first exercise ↗

ONE METHOD / MANY LANGUAGES

First principles

Observe → Question → Test → Revise

A small experiment in perspective.

A CONCEPT ATLAS

Give an idea a structure.

Philosophy and biomimicry become design questions we can test.

Explore the Gītā, Stoic, aviation, and swarm lenses in my ontological harness work. Compare how each frames a decision, then ask what evidence would challenge the shared structure.

Explore the harness ideas ↗

TWO WORKING LABS

From a strange idea
to a useful question.

Explore the rules. Then bring the discipline back to everyday work.

EXERCISE / 001

Change one thing.

An invented example. A question you can test with a model you already use.

01 / PREDICT

Write the expectation.

Give a model three fictional task notes. Ask it to separate a stated due date from a missing one. Predict where it might make an assumption.

02 / COMPARE

Change one instruction.

Add: “If a date is missing, write unknown.” Keep the invented notes unchanged and compare the outputs. Repeat before drawing a conclusion.

03 / EXPLAIN

Keep the uncertainty.

Did the changed instruction help? Which errors remain? Record the model, input, output, and limits. One result does not establish a general rule.

BRING IT BACK TO WORK

One recurring handoff, made clearer.

A scoped workflow pilot for independent restaurants and small service businesses. Start with the tools you have and a result the owner can inspect.

Work with Drew ↗

PART OF A LARGER JOURNEY

Build. Play. Keep a record.

Drew’s journal connects these experiments with human needs, philosophy, biomimicry, and thinking with machines.

Read Drew, in Progress ↗