Context
Alongside his work in engineering and system integration, Robert is co-founding AnyMDL, where he serves as CTO, focused on building a control layer for AI execution in environments where outputs carry real-world consequences. He is building the venture with Robert Fisher II, who leads the business and market side of the company. The work involves system-level control over knowledge authorization, execution policy, and traceable decision paths before an AI action is allowed to occur.
Problem
Most AI systems generate output first and validate later. That model can work in low-consequence settings, but it breaks down when outputs become contractual, regulated, licensed, or operationally critical. Once an output is produced, it cannot be un-generated. In many environments, there is still no neutral control layer determining whether execution should occur before it happens.
Approach
Robert is designing AnyMDL and FolderOS as a governed execution layer that sits above AI systems, while Robert Fisher II focuses on how the system is positioned and delivered.
Result
The architecture is advancing toward governed execution rather than open-ended generation. Elements of the control model are now being formalized through patent filings.
It matters because high-consequence AI systems need authorization before action, not validation after the fact.