Blog
Notes from the team building DecisionOS.
On encoded judgment, reasoning infrastructure, and the future of how drug discovery decisions get made.
BioBox MCP: Institutional Memory for Agents — Why Skills Aren't Enough for Drug Discovery Decisions
A skills file lets an agent approximate your scientific framework. Across a head to head evaluation on target prioritization, indication selection, multi-modal questions and gap analysis, that approximation drifted silently, while a reasoning system executed the framework and recorded which version ran.
Perspective
Points of view on how the highest-stakes decisions in drug discovery should be made.
What is decision provenance and why does drug discovery need it now
A practical definition of what it means to make scientific decisions traceable, and why it becomes critical the moment AI enters the loop.
Encoding scientific judgment into AI systems
Pilots pass every benchmark yet rarely change a decision that matters. The durable opportunity is to encode scientific judgment into a shared reasoning substrate, so the reasoning behind every program decision compounds instead of being rebuilt from scratch.
Fundamentals
The foundational principles of decision-making and problem-solving.
BioBox MCP: Institutional Memory for Agents — Why Skills Aren't Enough for Drug Discovery Decisions
A skills file lets an agent approximate your scientific framework. Across a head to head evaluation on target prioritization, indication selection, multi-modal questions and gap analysis, that approximation drifted silently, while a reasoning system executed the framework and recorded which version ran.
What exactly is "tacit knowledge"?
Expert judgment is the compressed residue of every failure a scientist has seen. It only becomes visible at the point of decision, and it is easier to extract through disagreement than explanation.
What is a multi-modal data foundation and why does the schema matter more than the storage?
Most pharma organizations have data. Very few have a data foundation. The difference is whether the schema reflects how scientists think about biology.