#463 — Privatizing the Apocalypse
A detailed discussion about Deep Vision, a canceled $125 million USAID program that would have collected 10,000 unknown viruses from remote locations, characterized their deadliness, and published their genomes publicly. The speakers describe their successful whistleblowing campaign that led to the program's termination.
Summary
The conversation focuses on Deep Vision, a USAID program that the speakers viewed as extremely dangerous and successfully campaigned to shut down. Deep Vision had three components: virus hunting (collecting ~10,000 unknown viruses from remote locations like bat caves in developing countries), characterization (testing to determine which were most dangerous), and publication (making the genomes of deadly viruses publicly available). The speaker argues this was dangerous because laboratories leak, vaccine development from hypothetical threats is ineffective, and publishing deadly virus genomes would give 30,000 people worldwide the ability to recreate pandemic-grade pathogens. The whistleblowing effort involved multiple high-profile individuals including Tristan Harris, Daniel Schmachtenberger, and Chelsea Clinton, as well as bipartisan political pressure from senators like Lindsey Graham, James Risch, and Rand Paul. The program was eventually defanged and formally killed in September 2023. The speakers argue that while COVID was relatively benign compared to potential threats, Deep Vision could have enabled much worse scenarios involving multiple simultaneous pandemics. They emphasize that this dangerous idea came from well-intentioned people, making it more concerning as AI and other technologies will soon enable many more people to conceive and implement similarly destructive projects.
Key Insights
- The Deep Vision program would have collected 10,000 unknown viruses from remote locations and published their genomes publicly, potentially giving 30,000 people worldwide the ability to recreate deadly pathogens
- The speaker argues that isolated bat caves are safer repositories for dangerous pathogens than laboratories, since all laboratory security levels demonstrably leak at unknown rates
- Creating vaccines for hypothetical pandemic threats discovered through virus hunting is not practically useful because safety and efficacy cannot be tested without actual outbreaks
- The successful campaign to kill Deep Vision involved bipartisan political pressure and a coalition including Tristan Harris, Daniel Schmachtenberger, Chelsea Clinton, and multiple senators
- The speaker contends that in 2021, fewer than 10 entities worldwide had the resources to implement something as potentially destructive as Deep Vision, but this number will dramatically increase as technology advances
Topics
Full transcript available for MurmurCast members
Sign Up to Access