We should not halt research simply because it might lead to dangerous knowledge; instead, we need to prioritize which knowledge to discover first.

The perception that open-source models pose a national security threat is misguided; collaboration and openness are necessary for progress.

I can't imagine a world where scientists are out of jobs because the essence of being a scientist is to tackle open problems, and there will always be complex challenges to address.

The NIH should adopt a more experimental approach to funding, allowing for high-risk, high-reward research that traditional models might overlook.

Scientific progress is often hindered not by a lack of new ideas, but by the slow and expensive process of testing those ideas in the lab. My focus is on how AI can replace or minimize the need for these lab experiments.

We must restrict access to dangerous technologies to prevent misuse, even if it frustrates those who want to innovate.

Robust internal security is crucial for AI labs to prevent exploitation and vulnerabilities, especially as AI models begin to write more of their own code.

As we automate scientists and innovators, we may see rapid advancements in scientific research, but new bottlenecks will likely arise.

To innovate on society, we need to find ways to reopen the frontier in geopolitical terms, but it's challenging due to existing constraints.