Simulating Pharma Frontiers

WHY

We needed to know whether the simulation engine we'd built for AI risk could carry a different kind of frontier — one driven by social demand rather than technical capability.

WHAT IT IS

A rebuild of our stakeholder simulation with its scenario layer replaced by behavioral pharmaceuticals — post-GLP-1 compounds that modulate focus, mood, appetite, compliance. Same five-team structure, same voting rounds, same progress curve. Same engine, different risk domain.

DESIGN DECISION

We kept the mechanics intact and only swapped the subject matter. The question wasn't whether the engine worked — it was whether it could carry a risk whose dynamics live downstream of the lab, in adoption, backlash, and norm-shift. If the structure held, we had a portable tool for any domain where demand outruns governance.

ONE OBSERVATION

The thing we were trying to model — social demand getting ahead of regulation and strategy — turned out to be harder to render than we expected. The AI version had a clean driver: capability goes up, stakeholders react. Here the driver is messier. A drug works. People want it. Employers offer it. Insurers price it. Schools absorb it. Each step is individually reasonable and collectively reshapes the ground before anyone has written a rule. We kept trying to express that through curve thresholds and headline feedback, and kept finding the curve wanted to behave like a capability line when the real dynamic was closer to a contagion — uneven, sector-specific, with feedback loops the original engine didn't quite have slots for.

Designed: August 2025