Rob Reid
About
Entrepreneur, author, host of the After On podcast; interviewer
Cast within
No topic-region cast yet — this appears once Rob Reid's compiled claims are aligned into a topic region's argument tree.
Claims by Rob Reid (20 of 79)
Terrorist nuclear attack would be low-grade devices
If a terrorist group ever obtained a nuclear weapon it would not be an arsenal of hydrogen bombs but one or two relatively low-grade devices deployed rapidly before capture, illustrating that catastrophic-weapon threats from non-state actors are constrained relative to state arsenals.
Weapon determines deadliness of lone wolves
The deadliness of lone-wolf mass murderers scales with the weapon they wield — mass stabbers in China kill far fewer than mass shooters in the US, who in turn kill far fewer than the occasional suicidal commercial pilot — so arming such people with pandemics must be prevented.
AI classifiers are porous defenses
Frontier AI models use biological 'classifiers' — digital bouncers that filter dangerous prompts and responses — but every model's system card acknowledges these can be dodged by sophisticated parties, and essentially everyone in biology has stories of extracting things they shouldn't from frontier models.
Secure DNA offers a cheap universal screening upgrade
Secure DNA — built by three Turing Prize winners and a Nobel laureate, run by a Swiss foundation on a $500,000 annual budget — is a fully encrypted screening tool designed to catch split orders, lets whitelisted parties anonymously submit new sequences of concern, is free to adopt in about an afternoon, but must be made mandatory and universal to work against split orders.
Cloud labs abstract away tacit knowledge
Tacit knowledge — the manual dexterity, taste, and nuanced lab skills accumulated over 15 years of higher education that are hard to write down — is currently a major barrier to who can cause biological mayhem, and cloud labs threaten to drastically reduce or even abstract away that barrier, collapsing the gap between having a bad idea and implementing it.
Pandemic-proof PPE is engineering-feasible and cheap
Pandemic-proof personal protective equipment (P4E) is within our engineering capabilities — as proven by high job-safety records in BSL4 labs — and a not-for-profit group believes it can be developed for well under a billion dollars, then domestically manufactured and stockpiled so that, given early biosurveillance warning, frontline workers can operate in full confidence during even a grim pandemic.
Deep VISION's three escalating bad ideas
Deep VISION planned three escalating dangerous activities: virus hunting (extracting ~10,000 undiscovered viruses from remote locations into leaky labs), characterization (experiments to determine which are pandemic-grade), and publication of the list and full genomes of the most dangerous viruses to the entire world.
All labs leak at unknown rates
Every category of laboratory, up to the highest biosecurity level, demonstrably leaks pathogens, and we do not know the rate at which they leak because there is no uniform reporting system, making an isolated bat cave a safer place for a pandemic-grade pathogen than a human-staffed lab.
Multiple simultaneous engineered pandemics could topple civilization
Whereas COVID emerged from a single point and spread at roughly 4.5 miles per hour giving two months of warning, a malevolent actor releasing seven engineered pandemics simultaneously from 20 airports could overwhelm diagnosis, cause combined fatality rates far beyond COVID, and collapse society when frontline workers refuse to risk their lives, shutting down food, law enforcement, and electricity.
Chaos agents could bluff pandemic deterrents
In a post-Deep VISION world, 'chaos agents' like rogue states, cartels, or warlords could credibly claim to possess engineered pandemics as a deterrent more powerful than a nuclear weapon, because published genomes would make the threat unverifiable and impossible to dismiss, causing geopolitical destabilization, panic, and market crashes.
Very few entities could have conceived Deep VISION
In 2021, extraordinarily few entities could have conceived and implemented an idea with Deep VISION's destructive potential — not terrorists, criminal gangs, cartels, or even a rogue state like Iran, leaving only the nine nuclear states and fewer than ten well-resourced biology entities capable of marshalling a $125 million budget, expensive labs, and partnerships across a dozen countries.
Deep VISION was defanged then formally killed
Following quiet bipartisan pressure — including from senators Lindsey Graham, James Risch, and Rand Paul, plus figures like Chelsea Clinton, Daniel Schmachtenberger, Tristan Harris, and the organization Helena — Deep VISION was first effectively defanged and then formally killed in September 2023, an unexpected public demise signaling that this sort of work should no longer be done.
USAID's PREDICT found 1,200 novel viruses
USAID previously funded a virus-hunting program called PREDICT that discovered roughly 1,200 novel mammalian viruses, and the Wuhan Institute of Virology has long collected coronaviruses, though no virus-hunting program approached Deep VISION's scale.
New knowledge and open science have limits
While 'new knowledge is good' and 'open science' (sharing raw data, code, lab notebooks, and protocols) are near-universal and valuable scientific precepts that enable replication, parallel exploration, and prevent information monopolies, open science is not a suicide pact and should not extend to publishing pandemic-grade genomes to every potential terrorist on earth.
Mutual assured destruction was a public good
During the Cold War only two people were empowered to risk tens or hundreds of millions of deaths, and their judicious use of that power — combined with total transparency, since everyone knew the power existed — probably prevented a conventional World War III through mutual assured destruction, making it paradoxically a public good.
Two iron laws of bio risk in the AI age
Bio risk in the age of AI must be governed by two iron laws: first, de-privatize the apocalypse by getting the number of people empowered to make existentially risky decisions as close to two as possible; and second, recognize that the passage of time is the master variable driving the expansion of danger.
My Notes
Loading notes...