Why it matters
Having an idea is different from being able to execute and verify it. Capability access shapes who can innovate.
The change we hope for
Make specific capabilities discoverable, combinable, executable and verifiable through sustainable supply arrangements.
Key barriers
Tacit knowledge, access rights, task ambiguity, handoff costs, quality verification and ownership of results.
Directions to explore
Task-based R&D support, authorized compute, remote experiments, engineering collaboration and distributed expertise. Start with one capability and a clear acceptance test.
Technical maturity
Bounded tool and laboratory workflows have published demonstrations. Interoperability, reliability and recovery remain specific to each workflow.
Commercial maturity
A reliable service offer requires demand interviews and comparison with consulting, contractors and existing providers.
Five-to-ten-year scenario
Over five to ten years, small teams could assemble specialized capabilities around a task rather than build every function internally. Access and accountability would still matter.
Open questions
Can useful tasks be specified in advance? Who guarantees delivery? What is cheaper to share, and what should remain inside a team?
Related proposals
Existing research
Research judgment · 001
Virtual Cells: When Biology Becomes Computable
Challenges the assumption that model scale alone makes prediction useful; guides prospective acceptance tests.
Actual content update 2026-08-15 · Version 0.3
Read research and sources ↗Research judgment · 003
Autonomous Science: When AI Becomes an Experimental Scientist
Shows why execution needs a bounded workflow and audit trail; does not establish demand or economics for callable R&D.
Actual content update 2026-08-15 · Version 0.3
Read research and sources ↗