We are building at the edge of a civilizational shift.
AI is no longer a tool you choose to use — it is infrastructure you must learn to govern. Every enterprise deploying AI agents today is navigating without instruments: no real-time visibility, no cost accountability, no control layer between intention and outcome.
That gap is not an inconvenience. It is a systemic risk. And it is exactly the problem ThinkNEO was built to close.
“AI without governance is just automation waiting to fail at scale.”
I started ThinkNEO because I believe the companies that will win the next decade are not the ones with the most models — they are the ones who understand what those models are doing, why, at what cost, and with what consequences.
Governance is not the enemy of innovation. It is the foundation that makes innovation sustainable. From Hong Kong, we are building the control plane for the age of AI — not for the labs, but for the enterprises, governments, and institutions that have to operate in the real world, where reliability, auditability, and cost discipline are not optional features. They are the product.
I have never been employed. Always the founder. Every chapter below was built from scratch — in a different country, in a different industry, often before the market existed.
Built an independent television business in Rio de Janeiro — purchasing airtime, producing original programming, selling sponsorship. No employer, no safety net. The internet ended that chapter; it opened every one that followed. Invited speaker on Internet & Technology, Universidade Candido Mendes (2006).
Built and scaled Brazil's first major OpenSimulator grid from 11 to 309 regions through corporate partnerships — meeting rooms, training, and virtual events for Brazilian companies years before the metaverse hype. The technical and community foundation that, 14 years later, became ThinkSim.
Founded the first legal hemp and cannabis company in South America following Uruguay's landmark legalization under President Mujica. Registered the company when the government system had no category for it — and helped shape the country's hemp framework. Grew to operations across Asia, Europe, and Latin America. Manufacturing partnership with the China Hemp Research Center.
Building the control plane that sits between enterprises and their AI providers — applying governance, observability, guardrails, and cost enforcement at runtime. Simultaneously building ThinkSim, the first AI-native virtual world, where government AI agents issue verifiable certificates with SHA-256 signatures and QR codes. Member, NVIDIA Inception Program.
Organizations are deploying agents at unprecedented speed — but few have instrumented what those agents cost in real time. The CFO who ignores AI spend today will be answering for it tomorrow. Here is what the control layer looks like, and why it cannot wait.
In cloud infrastructure, we learned that raw compute without orchestration is chaos. The same lesson is arriving for AI — faster than most people expect. The organizations that build the discipline layer early will have a structural advantage that compounds over time.
Government AI deployments carry a weight that enterprise ones do not. Every decision made by an agent affects a real person's life. Our work with USP/ICMC on Agents4Gov pushed us to think harder about what auditability truly means — and what it demands from the infrastructure layer.