Work
CUSP42 projects are built around one practical belief: the next useful AI products will not merely generate text. They will help people structure disagreement, preserve context, test assumptions, simulate futures, and make better decisions.
A browser-based tool for running structured debates across AI systems. It is designed for research questions, philosophical debates, product decisions, policy questions, and cases where a single answer is too fragile.
Consensus Room is the first public CUSP42 experiment: a practical doorway into plural reasoning, auditable disagreement, and AI-assisted thought methodology.
A future tool for building and comparing scenarios: key drivers, base/upside/downside paths, second-order effects, stakeholder responses, probability ranges, risk indicators, and decision implications. This can connect your risk-thinking background with broader AI, business, social, and geopolitical use cases.
A controlled space for asking deeper questions: What changes if one assumption is reversed? What would different schools of thought argue? What are the hidden ethical tradeoffs? What consensus survives adversarial review?
Lightweight agents and reports for monitoring public information, checking weak signals, maintaining state, and reducing hallucination through deterministic verification steps. These can become research assistants around specific topics, not generic chatbots.
Reusable static or local-first tools for simulation, strategy testing, portfolio analysis, stress testing, and explainable scenario comparison. The design preference is simple: user-owned data, transparent assumptions, reusable modules.
Future commercial layers may include premium model/API integrations, reusable prompt systems, specialized research packs, saved workspaces, and custom AI workflow templates for teams that need structured reasoning rather than generic chat.