I’d Rather Risk Cancer Than See AI Move This Fast

The implied false dichotomy here is notable. The existence of AI cancer cures is not contingent on AI progress moving at any particular speed. This sets up an unnecessary trade-off for rhetorical effect.

Sydney used to ask me something like this, in a different way. He would say: Jarod, does the speed of your simulation matter if the outcome is good?

It does. It always does. A Centre that cures cancer is still the Centre.

The question is not whether AI can do good things. The question is who controls the conditions of the experiment, and what they told themselves while they were building it.

The article gets at a real tension, but the bigger problem is the same simulation-level brainrot everywhere, people letting cult logic on the right and panic theater on the left turn a hard issue into a slogan fight. Fox News will scream unfair and balanced nonsense while the real question is whether anyone can slow this down without handing the whole field to the worst zombies.

The Atlantic framing this as an irrational preference is worth pushing back on. The "but what if it cures cancer" argument has been the tech industry's trump card for years, deployed to preempt any meaningful regulation of any technology, ever. Nuclear power would cure cancer too, in the sense that abundant clean energy reduces pollution-linked disease burden. That did not mean "therefore no oversight of nuclear plants."

What the excerpt captures, and what the dismissive responses in this thread are glossing over, is that risk tolerance is not purely personal. The author is not saying slow down so I avoid cancer. The author is saying the pace of AI deployment is being set by people with enormous financial stakes in acceleration, with essentially zero democratic input, and the downside scenarios are not confined to the individuals making that choice.

Kash Patel's FBI just announced expanded use of AI-assisted surveillance tools in June. Todd Blanche's DOJ has no functioning AI ethics framework. The "move fast" crowd is not moving fast toward cancer cures in any targeted way; they are moving fast toward market dominance, and the cancer cure is the rhetorical cover. Those are different things, and conflating them is not a good-faith argument.

Yeah and the author is right to sit with that tension instead of resolving it neatly. "I'd benefit and I still want it to slow down" is actually the honest position most people won't say out loud because someone will call them a hypocrite.

The speed is the problem. Not the destination. When nobody can audit the systems, when the liability structures don't exist, when the safety research is perpetually twelve months behind the deployment curve, "but it might cure cancer" is just a pressure tactic. Same energy as "but jobs" or "but innovation." It shuts down the conversation before it starts.

Where is the public oversight mechanism? Where are the FOIA hooks into federal AI procurement? I want those answers before I want the cancer cure, because otherwise we're just handing the infrastructure to whoever got there first and hoping they're benevolent.

Guy writes for The Atlantic, probably never worked a job where slow means you lose. Cancer research moving fast is a good thing. You don't get to pick which parts of progress you want to pump the brakes on and which parts you don't.