Why we're launching the Frontier Biodefense Fellowship

Unclear whether the final text exceeds the Forum's 10% threshold.]This is the companion post to our Frontier Biodefense Fellowship announcement. The short post covers the practical details. This one explains why we think a biodefense fellowship focused on defense-in-depth is particularly impactful now. What follows is a synthesis of recent thinking by people who have been building this agenda for years, particularly Coefficient Giving's biosecurity team and researchers & practitioners in the network around them.We are explaining our reasoning in four parts. Global catastrophic biological risk (GCBR) is mostly engineered, and possibly growing because of AI. Prevention at the source is not enough. The defense-in-depth agenda is tractable, and severely neglected as a response. And we are in, or quickly heading into, a transitional period where work in this field is unusually high-leverage.Biological risk may be growingGCBR is primarily not about natural pandemics because naturally arising diseases optimise for evolutionary fitness, not for killing everyone. That humanity has already existed for hundreds of thousands of years bounds how high a natural extinction risk we should expect. Andrew Snyder-Beattie puts more than 99% of biological x-risk on engineered rather than natural threats. State weapons programmes have explored parts of this design space already, with tools much less capable than what exists today.The likely reason why historically biological attacks have been rare is that the people who wanted to cause them mostly lacked the technical capability. AI is changing this fast. Snyder-Beattie estimates in the same resource as above that now most of the risk from engineered pandemics comes from non-state actors because AI closes their capability gap faster than anyone else's.Concrete examples for bio uplift are accumulating. Evo 2, a DNA foundation mod