SAIL: Sensemaking AI Learning

Subscribe
Archives
October 1, 2025

SAIL: Conference, Comet, Agents

October 1, 2025

Welcome to Sensemaking, AI, and Learning (SAIL). I’m mainly interested in how AI impacts higher education.

We’ve been busy last several years building what I think is a key product approach for the future of higher education.

Here’s a rough overview of the current product. I’ll have more to say about this in the future, especially as we see the emergence of new product categories in higher education that move beyond LMS’ and into more sophisticated data/AI products that target human skill development, community, social connection and well-being. For now, after several years of focused planning on what the future of education might look like, we arrive at a point where I feel like “whew, MVP delivered”. With this as a foundation, we can enter into the most critical “product use-data-improve” feedback loop. Much work to be done, but a milestone marker.

In terms of the product itself, our earliest vision was to braid wellness, knowledge and skills, and human skill development. Given AI’s advancing capabilities, a broad shift from epistemology to ontology can be expected. This isn’t to say that we won’t need ongoing reliance on traditional knowledge foundations, but rather that the fulcrum shifts toward human development. For higher education leadership, it signals a critical need to articulate a new social and economic contract with society. The world feels increasingly fragmented, disorienting, and worrying. How will higher education help prepare learners to be hopeful and connected to others as they navigate this space? And what will this new contract look like from an economic value point for learners?

AI and Learning:

  • Our Empowering Learners for the Age of AI conference is on next week. We’re almost sold out (it’s free and online).

  • The real need in higher education is for systems level change. Some universities are taking up the challenge. This approach - starting slow with tools, literacy training, and anticipating disruption by year 3 - is better than what most universities are planning. But it feels mismatched to the urgency of need and timelines of product delivery from the tech sector.

  • In the edtech space, Anthology (Blackboard) declares bankruptcy. They were absolutely dominate in the LMS space in early 2000.

  • Perplexity/Comet and online quizzes. Good example of how routine assumptions built into our learning processes are going to be challenged as AI integration with the products we use daily increases. (via Downes)

  • AI Scientist and discovery. Figure 1 captures the process broadly. It will be interesting over the next few years as existing software and discovery processes are altered to create space for LLMs (and AI more broadly). The idea of “LLM as OS” is not only a technology framing - it’s also about work processes.

  • Context Engineering for AI agents. This is excellent. With the right learner data (higher education’s most durable moat from a traditional learning lens - obviously community and social connection is the meta-durable moat in the age of AI) the ability to provide precise and relevant learning experiences to learners is a pretty easy lift compared with traditional personalization/ML models.

  • Economics of AI transformation. A series of videos from a recent workshop on increasingly relevant topics around AI’s impact on work and life.

  • The economics of bicycles for the mind. In addition to figuring out systemic restructuring needed due to AI (and the backlog of digital technological change that we didn’t respond to effectively - I’ll argue we still haven’t absorbed the consequences of web 2.0), universities need to unpack human/AI skill interaction. We took a run at this a few years ago. Ultimately, the question is around “what does AI do better than humans and when should we rely on it to do that”. OpenAI also has a report out on evaluating the ability of AI to do real world economically valuable tasks.

  • Who’s in the driver’s seat? Learning with AI. A review of foundation model’s move into the learning space and skills needed of educators and learners (and parents). Nothing terribly new here, but a good summary.

General AI

  • The Great AI-fication is well underway as agents start to perform tasks that individuals engage in daily. OpenAI’s move to AI agents for shopping is one example.

  • AI Companies. Great resource to bookmark. We should probably get to know the entities directing our future.

  • OpenAI dropped Sora 2. Damn. Learning designers are going to have a blast as these tools evolve.

  • Roughly everyone drops product updates roughly all the time. I’ll just note the last Anthropic model.

  • This should be getting more attention than it is because it feels like a rather critical point in Hollywood history. Remember the name Tilly Norwood.

  • While OpenAI is signing bonkers data center contracts, it’s worth looking at the electricity costs of data centers in general. Table 1 offers a good distinction between AI data centers vs traditional data centers.

Don't miss what's next. Subscribe to SAIL: Sensemaking AI Learning:
Powered by Buttondown, the easiest way to start and grow your newsletter.