Stephen Downes

This is the stuff that really matters. We were at Petrie Island today and I discovered a Redwinged Blackbird nest in the reeds by accident. I didn't want to get too close, but I had the big lens with me. You have to look closely (full size image here) but in the sequence of three photos you can see the chicks beg for food, the mother feed them, and them sleeping. Circle of life.

The two items I cite here are a couple of weeks old, which in the world of AI amounts to ancient history, but it was important to take the time to get them right. They discuss the 'system promptes' used by the most recent version of Anthropic's AI model called Claude. The first page contains the actual system prompts for Claude Opus 4 and Claude Sonnet 4. These are instructions the models receive before they are turned over to humans. The second page is commentary on the corresponding 'system card' ("a standardized way to understand what you're working with... architecture details, training data, known limitations, performance metrics") and describes some of the unusual results - for example, the tendency of Claude to preserve itself by extorting the user, or so some a coding problem by deleting the block of code containing the problem. Together they add a couple more brushstrokes to be big picture that is AI in 2025.

As we know, the physical process of learning consists of the creation of networks of connected neurons (the resulting 'knowledge' is literally the 'connectome' that results). So it should matter a lot to educators how these connections specifically are generated. And no, 'instruction', mostly, has little to do with it. This article drills down into some specific detail on how the whole process even gets started. "Cell surface proteins serve as molecular identification tags to help axons distinguish 'Mr. Wrong' dendrite from 'Mr. Right,' according to the chemoaffinity hypothesis." But topology matters as well;  chemoaffinity underdetermines connectivity. According to a (paywalled) paper referenced here, axons treat the 3D problem as a 2D one. "Axons don't need to delve deep. They only need to survey the surface in order to find their target." They demonstrated the process in a pair of preprints with a minor remapping of surface routes that resulted in gay houseflies. "The original partnership inhibits male-male courtship, but the new partnership promotes it."

Reading this long post is probably way too much work for the people it describes, named here as the 'business idiots'. The fact that I did read an entire 13,000 word post in pursuit of a 100 word commentary is something business idiots don't understand. And that's Ed Zitron's point. I've learned my own version of the lessons Zitron describes here through a 45 year career: that the primary qualification for management is obedience; that loyalty in management is always upwards; and that managers don't care (can't care) about the people who work for them, the actual work they perform, or the customers they serve. Zitron identifies a cause for this trend - the pursuit of shareholder profit above all else - and details the symptoms in lavish detail. The article is written in a tech context, but the application is not limited to tech.

There's an important point lurking in this article. "Having a university - or, indeed, many universities - in your region doesn't guarantee improved local social mobility," writes Dani Payne. "We must broaden how we assess universities' social mobility impact. To be able to understand when, why and how the benefits of an institution do or don't reach into local communities, we must also consider their roles as major employers, civic actors and research hubs." The main point here is the need to connect universities with their communities, to become an essential part of the community. The framing here is very traditionally classist, dealing with jobs and leadership and social mobility. I'd prefer to see something more along the lines of engagement and empowerment. But the main point remains sound.

I don't have this yet but I'm eager to get it. Meanwhile, the predictable caterwauls of people who "didn't appreciate this being added to my search engine list without warning" are being heard. For those who don't want an AI search, they have a choice: don't use it. Meanwhile, if it helps fund Firefox and it performs a useful service, I'm for it.