Stephen Downes

While Isaac Mao is talking about innovation at the national level, and specifically China, I think the same logic applies at corporate and institutional levels. Innovation, says Mao, is a process where "collective memory (Ψ) generating new coherent patterns (Φ), which reshape the institutional lattice (Ω), which then feeds back into memory with less entropy than before," and this requires "high Ψ diversity... high Ω permeability (and) fast Φ→Ψ integration" (or, in my own terms: diversity and autonomy, openness, and interaction, respectively). In other words, 'recursive emergence'. A discussion based on this follows, but I think it should be more nuanced. No government (or corporation or institution) is completely authoritarian or non-authoritarian, and they can be more or less authoritarian in different areas. And we are not (yet, at least) globally authoritarian. But the main point here is the definition of innovation, and I think this is not wrong, and I think a lot about how we should structure business, research and society follows from this.

My inclination is to answer this question with a "no" while I think the panel's collective response is to say "do what now?" Because, after all, what exactly is is to decode a memory?"I think," says Michał Januszewski, "that you want to do, if you want to read out a memory, is to have a learned function which maps some inputs to some outputs." So anything that is a 'memory' will be relative to a specific input? Or do we just generalize, and say any future (brain) state that depends on a past state is a memory. But then, is a behaviour (like a bird singing its song) a memory? There's a lot more along these lines; if you're in education, you should at least consider some of these questions. You can read the transcript (16 page PDF) or listen to the panel discussion as a video. 

As this article reports, "Facebook is testing placing a limit on how many links some users can share when they post on the social media platform." I'm sure I don't need to emphasize how anti-Web this is. "Tests like this underline why building a business that's overly dependent on any one platform's goodwill is incredibly risky." Even more to the point, it underlines why we shouldn't depend on centralzied servcies at all.  See also: the Guardian.

I'm a bit surprised to read that Zoom-bombing is still a thing. "'Zoom-bombing' refers to any time uninvited guests join a virtual meeting to disrupt it, typically by sharing loud audio or offensive imagery, or both." That said, I have worked with professors who even as late as this year have preferred to keep their Zoom chat open to whomever wants to drop in. It's definitely not what I recommend, no more than I would recommend allowing just anyone to join the stage at a conference panel discussion. At the same time, I've seen the logic work in reverse, which professors arguing that the discussion shouldn't be advertised so they can keep it 'open' without fear of it being Zoom-bombed. Me, I don't really see a contradiction of values here - 'open' does not entail a responsibility to provide an 'open mic'.

This is quite a long post and it wanders into irrelevant territory quite a bit (for example, a debate on whether liberal or conservative instructors are better teachers) but the message, ultimately, is quite clear: "If universities continue to charge premium tuition for sub-par human instruction that is inferior to a $20/month AI subscription, the market will collapse." So what to do? "The only survival strategy is to offer what AI cannot: the 25% of 'heroic' teachers who provide mentorship, complex critique, and the human accountability that drives actual learning." This, however, requires definition and measurement of teaching quality, and the data for this just isn't there. Hollis Robbins isn't the first to stress teaching quality in higher education, but I still think the concept is more elusive than critics think. Because if it weren't, we'd have some was of measuring it, and if we could, then we'd just have AI optimize for that. As Justin Weinberg says, "calling for a metric that will reduce all the various things a university can do for students to a few numbers does not seem like a promising strategy."

My interest in the world of mergers and acquisitions is virtually zero, but I'll pass along this notice because I imagine some people interested in MOOCs and edtech will want to take note. The announcement states, "they have entered into a definitive merger agreement under which Coursera will combine with Udemy in an all-stock transaction." If you're wondering about the layoffs that inevitably follow such an announcement, they anticipate "annual run-rate cost synergies of $115 million within 24 months of closing", so what's that, a thousand jobs, maybe a quarter of their combined workforce? The announcement also suggests "sustained investment in AI-driven platform innovation, rapid product development, and durable growth initiatives."