10% chance AI kills us?

Happy FRY-day! Let’s end the week right. 👊

[Sponsor FryAI]

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

What’s cookin’? A former Anthropic and OpenAI researcher, Jacob Coxon, quit his job and warned that top AI companies are moving too fast with technology they may not be able to control. His concern is not just that AI will get smarter, but that future systems could become powerful enough to hack, automate research, gain resources, and even help design better versions of themselves — an idea called recursive self-improvement. Other AI safety researchers echoed his concern, with one Anthropic alignment lead saying he believes there is a “more than 10%” chance AI could kill all humans within the next decade.

🤔 Hunter’s take: The scary part is not one dramatic prediction, but that some of the people closest to the technology are basically saying, “We don’t fully know how to stop this if it goes wrong.” Maybe the worst-case fears are too extreme, but when the builders themselves are asking for brakes, the rest of us should probably take notice.

What’s cookin’? Apple just announced its newest iPhones, including the iPhone 18 Pro and its first foldable model, the iPhone Duo. Alongside the hardware upgrades, Apple is adding new AI-focused camera features designed to help people trust what they see. The biggest one is Apple Reference Image, which works like a digital negative for iPhone photos. When someone takes a picture on the iPhone 18 Pro, the camera captures signed sensor data, and Apple’s Private Cloud Compute turns it into an unalterable reference version. Later, that reference image can be compared with edited copies to show whether the photo was changed. Apple also said the new phones will support SynthID, a standard that helps identify images created or altered with AI.

🤔 Hunter’s take: As AI images get better, the camera is becoming less about taking pretty pictures and more about proving what actually happened. That could matter everywhere from newsrooms to courtrooms to group chats and social media.

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?