In my last post I argued that the AI debate runs simultaneously at five levels (task, person, organization, economy, civilization) and that no one has agreed on which level should have standing. The civilization level, where the hardest questions live and where costs tend to concentrate, keeps getting treated as decorative. Nobody is lying. Everyone…
The debate about AI’s effects has been running for three years and producing mostly heat. Not because the researchers are wrong. Because they are not studying the same thing. Here is the clearest version of what I mean. A major 2025 meta-analysis of 69 studies found that ChatGPT produces significant positive effects on academic performance,…
When new data enters the pipeline, a full rerun used to mean hours. Pull everything, re-join everything, wait. That was the cost of keeping enrollment data current. The architecture I described last week does not work that way. New data drops into the stage where it belongs and everything downstream picks up from there. A…
I ran the same two prompts on two different AI systems this morning. First, a haiku about artificial intelligence. Then a ten-stanza poem in iambic pentameter. A parlor trick, the kind of thing you do when you’re testing something and want a clean finish line. Qwen 3.6, running locally on my machine, took 80.9 seconds…
Every admissions cycle, colleges make thousands of offers. A fraction of those students say yes and show up in September. The rest go somewhere else. Enrollment managers have always suspected they knew where. Suspicion and data are different things. For most IR offices, the standard answer to “where did our admits go?” was a spreadsheet….
At most small institutions, the problem is not that IR can’t produce data. The problem is that nobody has time to think about what it means – including the IR director producing it. I run a three-person IR office at a Lasallian Catholic university in the Bronx. We produce enrollment reports, retention analyses, IPEDS submissions,…
I thought the next layer would be more Obsidian configuration. It wasn’t. More graph view, more backlinks, more manual connection-making. That is what I used to do in Roam. I would spend real time threading notes together, following paths between ideas, maintaining a map of how everything connected. I did not do any of that….
I did not add backlinks. That was supposed to be the next step. The May 9 post said so explicitly. I had Obsidian pointed at the folder, the graph view was open, and the Zettelkasten people would tell you this is where the system comes alive. You start linking notes to each other. The connections…
I added Obsidian last. That was the right order. Most people building a personal knowledge system do it backwards. They install Obsidian first, configure a vault, explore the plugins, build a folder structure for notes they have not written yet. Then they discover that the system needs content to be useful, and content is harder…
The system runs on one folder. Insid that folder are plain text files, organized into directories. No proprietary database. No SaaS subscription that can change its pricing or deprecate a feature. The intelligence is not stored in the files. The files are context. The intelligence is in the AI that reads them before it…
