X-twitter Bookmark

The Long Tail

Creator Information

Name:

Dylan Roberts

Role:

Sane Stayer

Social Media/Website:

Visit

World Summary

Vision

By 2035, an AI co-scientist collapsed the cost of designing a therapy so far that diseases with a few hundred patients — or five — finally get treated, and a public institution makes sure the cure reaches people whose suffering would never have moved a market. The hope: medicine stopped abandoning the people too few to be profitable.

Transformative technology

An AI co-scientist for therapeutic design

Proposes candidate therapies, designs gene constructs, runs in-silico toxicology, and drafts trial protocols. The AI proposes; humans test, approve, and treat. Collapses design cost so a disease with five patients becomes a finite, fundable problem rather than an economic impossibility.

What kind of AI exists in your world?

Governed, bounded Tool AI — a co-scientist, not an autonomous physician. Reasons over biology at superhuman breadth but is leashed where a mistake reaches a human body. Every co-scientist must clear human wet-lab validation, independent safety review, and a regulator before touching a patient. Keeps an open notebook so a pharmacologist can see why it favored a candidate and overrule it. Core models open-weight, developed by a public-benefit consortium.

New or Reformed Institution

The Orphan Cures Commons (OCC). Thanks to the OCC, the world now treats rare-disease therapy as shared infrastructure, not a market. The OCC pools genetic and clinical data with patient consent; runs the AI co-scientist as open infrastructure any clinical team can access; manufactures small-batch therapies at cost; funds through a levy on blockbuster-drug revenues. It is forbidden from patenting any therapy out of reach of the patients whose data trained it.

Transformed Sector

Rare-disease drug development
Before: drug development was blockbuster-or-nothing. Fewer than 5% of rare diseases had an approved treatment; a diagnosis of an ultra-rare condition was effectively a diagnosis that no one was coming. After: design costs collapsed and the small market stopped being disqualified. The share of rare diseases with at least one approved therapy is climbing toward one-in-three; time from gene ID to testable candidate is month. For the first time since diagnosis, people and families feel hope.

A Major Crisis Overcome

The 2029 Patient-Data Revolt: rare-disease families had pooled their children’s genomes into platforms. They leased the data to profitable models, who found cures for large diseases, but never rare ones. In spring 2029, a coalition of data pools across forty countries staged a synchronized data strike, freezing registry access. Three Phase II trials for rare paediatric cancers halted within a week. Governments conceded: rare-disease needed infrastructure to be treated at all. The OCC was born.