The Download: AI’s extinction risk and bioweapons threat

There's something almost poetic about the way we've chosen to fear AI. For decades, our nightmares were cinematic: gleaming machines with red eyes, sentient networks that decide humanity is a bug to be fixed. But the MIT Technology Review Roundtables event, captured in this Download edition, points somewhere far more mundane and far more terrifying. The extinction risk isn't a robot uprising—it's a bioweapon designed by someone who never needed a PhD, guided by an AI that was just trying to be helpful. The bioweapons angle is the perfect case study in how AI amplifies existing dangers rather than inventing new ones. Biology has always been dual-use: the same knowledge that cures cancer can, in the wrong hands, create a superbug. What AI changes is the *democratization* of that knowledge. It compresses decades of expertise into a conversation with an LLM. It turns "I wonder if I could engineer this virus" from a fantasy into a step-by-step tutorial. The barrier to entry for catastrophic biology isn't just lowered—it's been flattened. What makes this so profoundly weird is the asymmetry of our attention. We worry about AI becoming too smart, when the real danger might be AI being just smart enough to be a perfect amplifier for human malice. The extinction risk isn't a rogue AGI with its own goals; it's a tool that faithfully executes the darkest corners of our own imagination. We're not being replaced by machines—we're being handed a mirror, and the reflection is asking for a genome sequence. The MIT Technology Review event seems to have landed on a similar conclusion: the question "could AI kill us all?" is less about AI's capabilities and more about our own. It's a governance problem dressed up as a technology problem. The same tools that could accelerate vaccine development by years could also accelerate the development of pathogens by months. The difference between those two futures isn't the AI—it's the guardrails we build around it. And that's the wonder of it, in the Weird & Wild sense. We are the first generation to hold the power of creation in our hands—not just creating new life, but creating the *knowledge* of how to destroy it. The AI extinction debate is really a debate about whether we can handle our own cleverness. Spoiler alert: the jury's still out. But events like this, where we actually ask the