Anthropic's CEO is worried

Good morning! Want to be the smartest person in your group chat? If so, you’re in the right spot. We are talking about stuff they won’t hear about for weeks, months, or maybe ever. 😤

[Sponsor FryAI]

This link leads to one of the most interesting things I’ve seen in AI recently.

What’s cookin’? Anthropic CEO Dario Amodei released a viral statement, saying that frontier AI companies need to slow the pace of capability gains so safety can keep up. He is not calling for a full stop. Instead, he wants AI labs to build at a more measured speed, especially now that AI systems are starting to help build better AI systems. His plan has three parts: put outside evaluators inside AI companies with employee-like access, coordinate safety standards among democratic countries, and eventually pursue global agreements on the most dangerous uses of AI. The worry is simple: if models become powerful enough to hack, deceive, or act in swarms before anyone understands them, small failures could scale very quickly.

🤔 Hunter’s take: If you remember from yesterday, Sam Altman just made a similar point, saying OpenAI may have to pace AI progress around how fast safety and alignment can keep up. That means two of the biggest AI labs are now publicly admitting the same uncomfortable thing: the models may be improving faster than our ability to control them. This entirely reframes the AI race. The question is no longer just “Who gets there first?” It is “Can anyone still steer once we arrive?” When the people profiting the most off AI development start tapping the brakes, it’s a bit worrying …

What’s cookin’? A creepy “Cat in the Hat” trend is spreading online, with AI-made images and videos showing the Dr. Seuss character lurking outside homes, schools, and other places at night. In West Haven, Connecticut, the trend became more serious after Bailey Middle School was specifically mentioned in one of the viral posts. Officials said the threat appeared to be a hoax, but because a real school was named, the district investigated and increased police presence on campus Monday as a precaution. The board of education also asked anyone with information to contact police, and told students to report anything they hear to school staff.

🤔 Hunter’s take: This is creepy. It also raises a weird new problem with AI slop: it does not have to be believable to cause real consequences. Once a fake or creepy post names a real school, people cannot just shrug it off. Suddenly, administrators, police, parents, teachers, and students all have to respond to something that may have started as a dumb online trend.

Hi — I’m Hunter, a PhD candidate whose work has appeared in major academic journals and popular tech outlets. I founded FryAI to make staying ahead of AI clear, accessible, and fun.

How close are we to the point where AI surpasses human intelligence across nearly every domain and begins advancing beyond our control or prediction?