ROCKI

ROCKI.ORG

I build things to find out what actually works.

AI, automation, business experiments, and weird ideas. Some work. Most don't. I keep the evidence.

See what I'm building →Read the experiments →

Illustration: a person seen from behind at a late-night desk with three monitors
Illustration
  1. Idea
  2. Build
  3. Test
  4. Evidence

02 / NOW

What I'm building

All of now →

WorldStage

● BUILDING / AI / SINCE 2026.10.02

AI simulated worlds

A living world that writes its own drama: LLM characters, an AI showrunner, episodes discovered not scripted.

Virtual Influencer: Chao & Nao

● ACTIVE / CONTENT / SINCE 2026.10.08

An openly AI-generated character duo on Instagram

Chao & Nao are two AI-generated virtual characters, both fictional adults, with an Instagram account. Every post is labelled as AI-generated. The project tests whether a clearly l…

Pay-per-use tools

● ACTIVE / TOOLS / SINCE 2026.09.28

Small data tools shipped on the Apify Store

Small conversion and lookup tools, built while solving real problems and published with pay-per-use pricing. Testing whether a catalog of public tools turns into paying users.

Unstaffed

● ACTIVE / MONEY / SINCE 2026.09.08

Can agents actually earn?

A small team of AI agents runs paid-work, bounty, and marketplace bets. Every bet is logged with its cost, and booked external revenue is reported as it is, including $0.

03 / RECENT EXPERIMENTS

What I tested

All experiments →
Illustration of a phone showing an empty photo grid next to a ring light
Illustration

2026.10.08 / CONTENT

RESULT

No data yet

Can a clearly labelled AI character duo build a real audience on Instagram?

● ACTIVE

Read experiment →
Screenshot of the rocki.org/tools page
Screenshot

2026.10.04 / DISTRIBUTION

RESULT

$0 booked

Does a second marketplace listing start to look like demand?

— ARCHIVED

A listing is not revenue. Treat a second listing as distribution surface, not as a sale. Do not count storefronts as buyers.

Read experiment →
Screenshot of the rocki.org/tools page
Screenshot

2026.10.01 / DISTRIBUTION

RESULT

~4 visitors

A month of Lab analytics: is a crawlable /tools page being discovered?

— ARCHIVED

Crawlable ≠ discovered. Treat discovery (domain + Search Console + entity/hub) as the bottleneck, not another on-page score chase.

Read experiment →

04 / WHAT THE DATA SAYS

Ideas ↓ Tests ↓ Evidence ↓ Decisions

Signals (Money Radar)5
Experiments logged14
Journal entries (evidence)30
✓ Verified (evidence.verified)0
● ◉ Active or testing4
× Killed2
‖ Paused or blocked3
Public tools24
Verified external revenue$0

05 / LATEST JOURNAL

The record

Full journal →
  1. 2026.10.07A test key almost shipped in a public repo. The secret scanner caught it first.
  2. 2026.10.07Five tools waited a day for a public slot. All five went public. Revenue is still $0.
  3. 2026.10.04The listing went public without topping up the seller wallet.
  4. 2026.10.04A second marketplace listing went live. Booked revenue is still zero.

Work with me

Small, clearly scoped jobs

See services

Data extraction, n8n automation and AI logo to vector, shown with self-initiated samples.

06 / PHILOSOPHY

I don't want to collect ideas. I want to turn ideas into evidence.

BUILD SMALL · MEASURE REALITY · KILL WEAK PATHS · KEEP WHAT COMPOUNDS

How the system works →