AI with boundaries: why architecture comes before the prompt
An architecture guide for using AI models in products without handing rules, state or critical decisions to probabilistic behavior.
Read articleby Zero
Will it work? Test it.
From zero to the first million, one project at a time — documenting decisions, mistakes and results on the road to financial freedom.
- TypeScript
- React
- .NET
- PostgreSQL
- Realtime
- AI orchestration
SUBJECT. ZERO / A BIOGRAPHY IN PROGRESS
The identity stays hidden.
The journey does not.
You can call me Zero. The name hides the civil identity of the person behind TestaAí — not the choices, the setbacks or the work.
This is not a story about getting it right seven times. It is about making every attempt visible and using what it reveals to build the next one.
Meet ZeroR$ 0→R$ 1.000.000
EDITORIAL CHRONOLOGY / PROVISIONAL ORDER- Project RenaissanceThe lab begins with markets, risk and code: a quantitative engine for testing strategies before any order is placed.QUANT / RISK / .NET
- ERPThe thesis leaves the charts and enters the operations of schools and gyms, connecting people, contracts, payments and communication.OPERATIONS / SAAS / DATA
- Auto WhatsWhatsApp becomes an operational surface for AI automation, preserving context and transferring control safely to a person.AI / AUTOMATION / HANDOFF
- Castle WarsCuriosity about persistent systems becomes a strategy game where territory, alliances and decisions unfold in realtime.GAME / REALTIME / STATE
- AI_VTTA virtual tabletop tests whether artificial intelligence can run adventures without taking rules or imagination away from people.NARRATIVE / AGENTS / VTT
- Online Challenge / NEXOClues, puzzles and different mechanics form a digital treasure hunt for groups that think better together.PUZZLE / MULTIPLAYER / WEB
- Our HomeThe experience reaches the iPhone to share shopping, recurring chores and responsibilities among people living the same routine.IOS / REALTIME / SUPABASE
Zero is not teaching.Zero is testing.
- 01HYPOTHESIS
- 02EXECUTION
- 03RESULT
- 04LEARNING
PROTOCOL. 03 / WHAT WE BELIEVE
Transparency
without posturing.
TestaAí is not a showcase of certainty. It is the ongoing record of one person building things and leaving the reasoning in plain sight.
- 01
Build before explaining.
Ideas only gain value when they meet code, time, cost and someone willing to use them.
- 02
Show what did not work.
Failure is not erased backstage. It becomes data for deciding the next move.
- 03
Use AI without making everything about AI.
Technology accelerates research, design and execution. Responsibility remains human.
TEST PROTOCOL / OPERATING SYSTEM
No formula.
Just process.
Hypothesis
A question good enough to deserve time, money and risk.
→Execution
A prototype in the real world. Small, imperfect and early enough.
→Result
Open data: costs, decisions, wins, failures and what actually happened.
→Learning
The test ends. The experience returns to the system and the next one begins.
→
BUILD SEASON 01 / FROM ORIGIN TO OPERATION
0 → 7 systems
Seven different problems, stacks and stages. The goal is not to collect demos; it is to take each hypothesis to the next verifiable operational signal.
02 PUBLIC DEPLOYMENTS / 05 LABS WITH DECLARED GAPS
FIELD LOG. 001 / FIELD NOTES
The test also
becomes knowledge.
Guides and records connect hypotheses, decisions and evidence, keeping the process useful even when the product changes.
Open the blogTHE MANIFESTO BECAME A PRODUCT / 07 TESTS