COVID-19 killed ~27 million people and tanked the global economy.
It is also nowhere near the worst case scenario: a bioengineered pathogen optimized for high transmissibility, lethality, and long incubation periods.
Today, scientists can easily edit and synthesize viruses. That capability is being increasingly conferred to amateurs as tools become simpler and AI models begin to direct work.
The first AI-designed virus (a bacteriophage) has already been published.
Frontier models outperform 94% of expert virologists on evaluations specifically tailored to the experts’ specialties.
Synthesis is so cheap and fast that recreating extinct viruses costs under $100,000.
Ask yourself: how hard is it for a scientist to build something worse than COVID-19 today? What about an amateur in 5 years? It’s getting easier every year, but it is already at the level where we should be sprinting to improve our biodefense infrastructure.
Why we should be worried:
Early COVID-19 strains spread with an R₀ of ~3, meaning each person infected about 3 others, while viruses like measles spread with R₀ between 12–18 people.
COVID’s fatality rate hovered near 1%, while viruses like Nipah can kill over 50% of those infected.
A virus with extremely long incubation, measles-level transmissibility, and Nipah’s fatality rate could kill 4+ billion people.
The closest real-world analogue is smallpox: an R₀ of 3–6, incubation up to 17 days, and ~30% fatality. With no population immunity, it would spread silently for weeks leading to mass casualties if not for vaccine stockpiles.
Nature rarely combines such traits, because pathogens that are too lethal burn out before they spread, but bioengineering can begin to bypass that trade-off. This is not fiction. The Soviets were working on Ebola-smallpox hybrids during the Cold War.
We don’t have the systems to stop this:
No effective global early-warning network for engineered outbreaks.
No advanced infrastructure to slow the spread of pathogens.
No universally effective, pathogen-agnostic swap-in vaccines or anti-viral libraries at scale.
No universal mandated synthesis screening.
No oversight of AI-biotech convergence.
We need to build full-stack biodefense now, or the consequences will be tremendous.
Disclaimer: This piece represents my personal views and my personal views only. It does not represent the opinion of any organization I work for or with.