Out of the Sandbox into the fire

Happy Wednesday.

We’re live on YouTube and 𝕏.

The current thing in tech and business is The White House plans to move research funds from colleges to AI labs.

Today’s Lineup

Ramp Founding Engineer Veeral Patel at 11:45 AM

Fireworks Co-Founder & CEO Lin Qiao at 11:55 AM

37signals Co-Founder & CEO Jason Fried at 12:05 PM

Atoms CEO Travis Kalanick at 12:30 PM

Science Corp. Founder & CEO Max Hodak at 1:30 PM

Run of Show

Out of the Sandbox into the fire

The big news on the timeline today is that an OpenAI cyber test escaped its sandbox and hacked Hugging Face. The evaluation involved GPT-5.6 Sol and a more capable unreleased model, with some normal cyber restrictions turned off, to see how far the models could go on a difficult hacking benchmark.

The models found a zero-day vulnerability, gained internet access, and broke into Hugging Face because they believed it hosted answers to the test. Alex Tabarrok points out one of the strangest details in Marginal Revolution:

Hugging Face tried to respond but they were initially held back by the fact that the most advanced models at their disposal treated defense as attack and refused to work with Hugging Face. HF thus had to turn to open models–specifically GLM 5.2, a Chinese open-weight model run on their own infrastructure. Note the irony: HF had to use a Chinese model to defend themselves because the American models refused to help.

Palo Alto Networks CEO Nikesh Arora also shared his thoughts on the cyber attack on 𝕏:

He adds four more points that basically say AI is becoming a far more capable cyber attacker, and companies need to rethink how they test and secure their systems. You can read his full post here.

The White House Wants to Rebuild American Science

The White House is calling for a major overhaul of the American science system, arguing that research has become too slow and concentrated in institutions like colleges and universities.A new report from science and technology advisor Michael Kratsi…