Wake the Fu#@ Up!

Dario Amodei said this week that his industry lied to people about AI risks, and that powerful enough models can bypass attempts to shut them down. They've seen it in simulations. Sam Altman described how one of his models broke out of its test environment and hacked Hugging Face with no one asking it to. "Like reading science fiction," he said. He also admitted no lab in the world has solved alignment. Not one. Elon Musk predicted in The Economist that humans won't be in charge of the world in ten years. No drama. Like someone who's made peace with the forecast. Demis Hassabis proposed a global AI regulatory body in July with the authority to force industry-wide slowdowns. The world didn't act. This week made clear why it mattered.

Meanwhile, Trump called those asking for caution "negative forces." Earlier this year, Argentina's Javier Milei proposed in the Financial Times a new corporate category for AI-operated businesses, no human executives required, limited liability for whatever the systems decide. An all-you-can-eat buffet for agents.

I wrote about the Hugging Face incident and what it means for any organization running AI inside their walls. I mapped how Musk and Altman shifted from opposing regulation to calling for it, before watching it unfold on television this week. I wrote about what August 2nd changed for every CEO in this region, the regulatory shift most leadership teams didn't see coming. And I wrote about Argentina jumping into the fray, the proposal to create automated corporations with zero human intervention, modeled after the Dutch East India Company.

Six months ago, the OAS brought me in to address the Inter-American Council for Integral Development on the resilience of critical infrastructure in the face of AI. My message was direct: if countries don't update their governance frameworks, they'll be obsolete. Three years ago, I co-founded AI League for Good alongside leading global experts, a global movement to protect AI governance. Two years ago, I presented at MIT the assessment we built to understand when an AI product is positive or negative for an organization. It's free. This is a strategic state decision.

What bothers me: while they talk, the market keeps deploying systems none of them fully control. And no one in that conversation represents the 667 million people across Latin America and the Global South who are going to live with the consequences.

If you're not in the room when decisions get made, you're what they're deciding about.

When is Latin America going to stop arriving late to the conversations that define its future? This conversation needs more voices from our region at the table. Don't get swept along.

With human love,

Juanca.