Image via 9to5Google
Friday, November 28, 2025
Character.AI just banned all teens from its platform
Character.AI is hitting the brakes on teen access (yikes), pivoting to interactive fiction as the chatbot era for younger users seemingly hits a wall. Meanwhile, OpenAI severed ties with Mixpanel after an analytics breach exposed API user data, and in a wild move to dodge chip sanctions, Alibaba and ByteDance are allegedly training Qwen and Doubao LLMs using Nvidia chips in Southeast Asian data centers. On the capabilities front, it's a bit humbling: these advanced models still can't handle simple geometry tasks that humans solve in minutes. And Google's making a bold move capping free Gemini 3 Pro access due to demand overload. Would you ban teens from AI chatbots?
Image via 9to5Google
Top Stories
Character.AI has banned minors from its chatbots and launched a safer alternative called 'Stories' as mounting evidence of mental health risks and regulatory scrutiny force the AI industry to reckon with AI companion addiction among teens.
The Register
OpenAI cut off Mixpanel after an analytics breach exposed API user data, prompting the company to audit its entire vendor security posture and implement stricter vendor requirements across its ecosystem.
Tom's Hardware
Chinese AI leaders are exploiting regulatory gaps by training advanced LLMs offshore using Nvidia GPUs in Southeast Asia, maintaining competitiveness with Western labs while adhering to the letter—if not spirit—of U.S. export controls.
Personal Blog
Major LLMs fail a seemingly simple task combining counting and spatial reasoning that humans solve easily, exposing gaps in current AI capabilities despite recent advances.
9to5Google
Google is restricting free-tier access to Gemini 3 Pro and AI-powered features due to overwhelming demand, a common pattern as enterprise and consumer adoption strains infrastructure capacity.
Keep Reading
Enjoyed this issue?
Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.