SAIL: Building AI University, Anxiety, Mixed results?, AI Use
September 15, 2025
Welcome to Sensemaking, AI, and Learning (SAIL). I focus on AI and higher education.
Since roughly April 2023, I’ve been involved with building something like an AI university. Or perhaps a human development platform. Or perhaps an AI-enabled learning system. We’re starting our first deployment next week with about eight universities and are planning broader deployment in January 2024. I’m deeply enamored with higher education as a critical counter-balancing institution in society. The cost has escalated, especially in USA, and that raises the need for a new social, cultural, and economic contract with society. Central to this contract is the vision of universities as AI product builders, not only consumers of frontier lab offerings.
I’ve dusted off my blog and started writing about what I’ve learned over the last 2 ½ years. This is the first of many (many) posts on the experience.
I’ve also posted a 25 min video on somewhat of the current state of AI as part one of a four part series. The goal is to get universities to become AI product builders.
Also, we have a pretty fantastic free/open/online conference happening in a few weeks on Empowering Learners for the Age of AI (our 5th event). Registration is open.
AI and Learning
Anthropic released a fascinating report on AI use. It’s long and detailed. Coding remains top use case. Education is continuing rapid growth.
How people use chatgpt. As detailed in figure 9, education/learning is a key use here as well. Not really shocking, but roughly maps the study above.
A forecast for learning and earning. GSV Ventures releases this report, heavily focused on AI trends. More a review and summary than new information.
Navigating the future. One of many attempts to define future skills and navigating a complex world. Self-management skills figure prominently (p.43).
Drop in international students reported in USA.
General AI
Different ecosystems produce different results depending on constraining factors. The USA is cash and tech (GPU) rich and proprietary models reign. China is taking a different approach: open source. More details on their AI+ initiative just released. Focus is on technical innovation, scientific discovery, and manufacturing innovation.
What is context engineering? You can never read too many.
Resistance to AI will become the largest social activist movement in history. The more people know about AI, the less they trust it. And it’s dividing political houses.
The Women in Love with AI Companions. “these large corporations are in effect running a very large-scale experiment on all of humanity.” We shall soon yearn for the softer harms of social media.
Also: “My Boyfriend is AI”. Section 2 provides an overview of existing AI-human relationship literature.
And if that’s not enough emotional/social/relationship anxiety, how about people turning to AI for spiritual direction and confession.
Not sure what this means, but xAI and Google are both reducing their staff. The former involved in data annotation.
Amazon 100% robot staffed in five years? So say some hypesters.
GPT-5 Codex is getting rave reviews. “GPT‑5-Codex is a version of GPT‑5 further optimized for agentic software engineering in Codex.“