Like Amazon, But For Ebola
When we imagine the machines turning on us, we reach for the cinematic. The red-eyed killer robot. The plasma rifle in the 40-watt range. Skynet. Terminators marching through the rubble. It's a comforting fantasy, because it's a fight ... and a fight is something humans understand.
But an AI that actually wanted us gone wouldn't bother with the theatrics. It would do something quieter and infinitely more efficient: it would order up a pathogen for which we have zero immunity, order it synthesized in a lab, and wipe the slate clean. No drama. No final stand. Just a sequence of nucleotides and a tracking number.
The more immediate threat isn't the machine's intentions ... it's ours. The knowledge that once gated bioweapons, built over years of doctoral training and a well-funded lab, is eroding. You no longer need a virology degree and a state sponsor. You need malice, patience, and a subscription to a sufficiently capable AI model. The same tools accelerating cancer research and collapsing drug-discovery timelines are the tools that hand a motivated lunatic a recipe he could never have written himself.
Here's the part that should make you sit up: the people who built these machines agree. This week, the chief executives of OpenAI, Anthropic, and Google DeepMind"Sam Altman, Dario Amodei, and Demis Hassabis, men who agree on almost nothing and compete for the same oxygen, put their names to a letter begging Congress to act.
Joined by Microsoft's Mustafa Suleyman and Meta's Alexandr Wang, they're asking lawmakers to require that companies selling synthetic DNA and RNA screen their orders, verify their customers, and block the combinations that could end us. When the arsonists are ready to burn the entire global economy down with AGI show up at the City Council meeting asking for stricter fire codes, it's worth listening.
The ask is almost insultingly modest. Make the people synthesizing the literal building blocks of life check whether what they're building is a weapon. That's it. As one former Trump AI adviser put it, it seems like a reasonable thing for society to insist upon. It does. It is the regulatory equivalent of asking a gun shop to glance at an ID.
So naturally, it will go nowhere.
Trump already torched the Biden-era screening framework and promised a replacement that the White House has yet to write. Bills will languish in committee, going precisely nowhere. And the administration's tech brain trust now runs through David Sacks, lately promoted from AI czar to co-chair of the President's science advisory council, where the gospel remains the same: regulation is the enemy, friction is sin, and anything that slows the velocity of innovation is a heresy to be crushed.
A man whose entire public theology is deregulation is not about to bless a screening mandate, however sane, however bipartisan, however urgent.
And so we arrive at the punchline of the age. The men building the machine are warning us about the machine. The government's job is to listen.
So the models will get better at biology. The tools to sequence DNA, RNA, and viruses will get cheaper and better, Moore's law for plagues.
No one will regulate it. No one will stop it. It's Amazon, but for Ebola.
Two-day shipping. No questions asked.
