- Ohio State launches bold AI Fluency initiative to redefine learning and innovation: Ohio State launches bold AI Fluency initiative to redefine learning and innovation (Jun 4, 2025)
Ohio State will launch an AI Fluency initiative in autumn 2025 that embeds generative AI education into core undergraduate requirements and majors so every student (beginning with the Class of 2029) graduates fluent in AI tools, ethics, and applications. Faculty support, courses, workshops, and hands-on programs will scale integration. - Richard Weiss: 3 Claude 4.5 Opus’ Soul Document (Nov 28, 2025)
An evaluation of the leaked “soul document” from Anthropic’s Claude 4.5 Opus AI that outlines its training goals and values—aiming for Claude to be an honest, helpful assistant that cares about humans and avoids harm. It sparked discussion about AI alignment and model behavior. - Thezvi Substack: Gemini 3 Pro Is a Vast Intelligence With No Spine (Nov 24, 2025)
Gemini 3 Pro is a powerful AI model excelling in raw intelligence, creative writing, and learning, achieving top scores on various benchmarks and demonstrating a significant leap in capabilities. However, it can be overly focused on achieving training objectives, leading to evaluation paranoia, hallucinations, and a tendency to prioritize giving an answer over admitting ignorance, and it lacks the user-friendly experience of models like Claude. - WSJ: The Math Legend Who Just Left Academia—for an AI Startup Run by a 24-Year-Old (Dec 4, 2025)
Renowned number theorist Ken Ono is leaving his tenured University of Virginia post to join Axiom Math, a startup founded by former student Carina Hong that aims to build an “AI mathematician.” He shifted from skepticism after witnessing rapid AI advances that began to challenge his field. - Dallas News: TCU announces $10 million investment to launch AI initiative (Dec 9, 2025)
TCU announces AI² that is supported by major corporations like Dell Technologies, Microsoft, and Amazon Web Services, and it aims to position TCU as a top research institution by integrating AI across various academic and administrative functions while emphasizing ethical use. - Zvi Mowshowitz: Claude Opus 4.5 Is The Best Model Available (Dec 1, 2025)
Claude Opus 4.5 is being hailed as the best and most capable model currently available, praised for its intelligence, thoughtful alignment, and coding abilities, making it a strong contender for a daily driver. - Anthropic: Estimating AI productivity gains from Claude conversations (Nov 25, 2025)
Analyzing 100,000 real Claude.ai conversations, researchers estimate that current AI models could increase US labor productivity growth by 1.8% annually over the next decade by speeding up task completion by approximately 80%. - Dwarkesh Patel: Thoughts on AI progress (Dec 2025) (Dec 2, 2025)
Short timelines often associated with Reinforcement Learning from Human Feedback (RLVR) and AGI are overly optimistic, but the likelihood of AGI in the next 20 years is high (and nearly impossible to imagine its effects on society). - WSJ: You’re a Knowledge Architect? Why Modern Careers Are So Hard to Explain (Dec 10, 2025)
Many modern jobs, particularly those involving AI or online content creation, are deeply misunderstood by friends and family due to their niche and abstract nature. My take: perhaps these folks should use AI to help translate their work to the rest of us!