NYC bans AI in school

[Sponsor FryAI]

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

What’s cookin’? New York City is banning student use of generative AI from kindergarten through eighth grade for the 2026-27 school year, affecting about 600,000 students. Younger students will not be allowed to use AI tools or companion chatbots while the city studies how the technology affects learning, attention, and child development. High school students, however, will still get AI literacy lessons twice a year, covering bias, ethics, and critical thinking. Some high school classrooms may also test AI in closely monitored pilot programs. Teachers can still use approved AI tools for lesson planning and administrative work.

🤔 Hunter’s take: When used too early or often, AI can quietly replace the struggle that builds thinking. The interesting move here is not “AI bad.” It is “AI later, with guardrails.” Nobody is going to get this perfect, but we need schools and universities to try stuff like this so we can learn and adapt to the methods that work best.

What’s cookin’? Google just released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, two new AI models built for harder, longer tasks. Gemini 3.8 Flash is designed to help with coding, research, business analysis, legal work, and multi-step reasoning, while keeping the same speed and low cost as the previous Flash model. The big change is that it “works harder” on difficult problems by taking more reasoning steps, using tools repeatedly, and checking its work more carefully. The Cyber version is built specifically for trusted defenders, helping find software vulnerabilities and create patches before attackers can exploit them.

🤔 Hunter’s take: This follows the same trend we saw from Fable 5.1 earlier this week: AI models are moving from “answer machines” to “do-the-work machines.” The real shift is persistence. These systems are getting better at staying inside a task, using tools, checking their own work, and pushing toward a finished result. That matters because real work usually doesn’t consist of a simple answer, but detailed, multi-step processes.

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?