SF Protest · Saturday, July 11, 2026

Freeze
AI Development

Before it kills us all.

Free slushies for everyone who shows up with a sign.
Cherry red. Blue raspberry. One simple demand: Freeze All The GPUs.

❄ Join the Protest · 7/11

Bring a sign · Get a slushie · 12 PM – 4 PM · 500 Howard St

The Protest

Saturday, July 11

12:00 PM – 4:00 PM PT

Downtown San Francisco

500 Howard Street →

Free Slushies 🍧 One per person — just show up with a sign.
Cherry & blue raspberry · Vegan & sugar-free options · While supplies last

Join us on 7/11

Why Freeze?

Slow down before we lose what we can't get back.

Protect Human Creativity

Writing, music, art, code — the things humans make for each other are being flattened into training data. A freeze gives us time to draw the lines.

Protect Human Jobs

Frontier AI is replacing entire professions before any of us got to vote on it. A freeze gives workers and policymakers time to catch up.

Stop Automated Decisions

Hiring, lending, sentencing, healthcare — being decided by systems no one can audit. A freeze lets us demand accountability before it's everywhere.

Speak Up for Humanity

The people building this technology have told us, repeatedly, that it could kill us. The rest of us deserve a say in whether they keep going.

Our Single Demand

Freeze All The GPUs.

No new frontier training runs. No more scaling toward systems no one can control. Stop the compute, stop the race.

Dario Amodei

CEO of Anthropic

"I wish we had 5 to 10 years."

"The thing that keeps me up at night is the incredible market race between top AI companies."

"This is the trap: AI is so powerful, such a glittering prize, that it is very difficult for human civilization to impose any restraints on it at all."

The architects of these systems know the race is reckless. They've warned us about the existential risks themselves, repeatedly. But each argues they can't slow down because they need to beat the others.

We're not asking anyone to stop alone. Just to say: if the others freeze, I will too.

In February 2026, Anthropic dropped its commitment to pause development if its own AI became too dangerous. Holden Karnofsky, former CEO of Open Philanthropy and now at Anthropic, admitted there was "enormous pressure" to downplay risks, because triggering a pause would hurt the company. Their solution? Remove the commitment entirely.

That's why we're setting up shop on July 11th — and why we're bringing slushies. A freeze is the simplest, most legible thing a CEO can commit to: if everyone stops, we stop. They claim they're building safely for the benefit of humanity. It's time they prove it.

Join the Freeze

Are you against these companies or AI?

No. We are not protesting the technology; we are protesting the reckless race to build it. We are here to give AI lab CEOs the public mandate they need to freeze development and start coordinating.

Why "freeze" instead of "pause" or "stop"?

A freeze is conditional, mutual, and reversible — if the others freeze, we freeze too. It's the smallest commitment that creates the conditions for international coordination. (And yes, the slushies are on theme.)

Is this protest nonviolent?

Yes. This is a peaceful gathering of concerned citizens, families, workers, and researchers. We are not here to disrupt, and we welcome employees of these companies to join us — they get a slushie too.

What about China?

China would also need to freeze. Any agreement only works if it includes all major AI labs globally. The MIRI Technical Governance Team published a detailed proposal for an international agreement centered on a coalition led by the US and China, with verification mechanisms including AI chip tracking. Public commitments from Western lab CEOs are a first step toward the kind of international coordination that makes this possible.

What would a freeze look like in practice?

No new training runs of larger or more general frontier models. The teams currently working on improving the capabilities of these models would move to narrow AI applications or alignment research instead. Current models stay available. For a detailed technical proposal, see MIRI's international agreement paper, which operationalizes limits through compute thresholds (FLOP caps) and restricts research that advances toward superintelligence while preserving access to current AI.

Are the slushies really free?

Yes — one slushie per person. Just show up with a sign. Cherry and blue raspberry. Vegan and sugar-free options available. While supplies last. Get there early; we are not running a chain of convenience stores.

What kind of sign?

Anything that gets the point across. "Freeze All The GPUs," "Slow Down AI," "Humans First" — bring your own. Cardboard, paper, poster board, whatever. The sign is the ticket.

Have questions?

Reach us at contact@freeze-ai.help

Want to help organize?