Stephen Downes

As reported here, "WikiWhiz is a free daily guessing game built on top of English Wikipedia and other sister projects. If you've played Wordle, the rhythm will feel familiar: one puzzle a day, same for everyone, and a result you can share with friends. The difference is the answer. Instead of a five-letter word, it is a Wikipedia article, and the clues come from the knowledge our communities have spent more than two decades building." Good fun on a slow news Thursday.

If AI agents visit your website you might be interested in this: "WebMCP (not to be confused with MCP) gives web developers a standard way to publish structured tools to access the capabilities of their website enabling agents to call those functions with structured inputs." This article describes them, shows how to build them, and explains how to put them into action.

A pretty good joke made the rounds this week suggesting that a large number of online pundits disappeared following Ukrainian strike on Russian data centres like Yandex. It's funny because it feels true, but it's also probably based more on myth than fact. The data seem compelling (see this chart, in German) but "the unknown or possibly non-existent data base as well as the inaccurate scaling and values "from the future" show that this is not a meaningful, fact-based graphic created according to statistical methods."

Videos and notes from the FediForum conference I attended last week. Most recommended is Anil Dash's keynote address in which he laid down some truth on app developers hoping to jump on the VC gravy train (sadly there's no transcript). There's also a recording of the presentation on ATmosphere Spaces and a number of live demos, including the IndieBeat radio stream, Acorn Bluesky communities, the FediDeck Tweetdeck alternative for Mastodon and Bluesky, and the AllPeeP community app.

Some realistic thinking about the role of AI in philosophy: "With a little human guidance, you can crank out the kind of papers that make up 95% of what we publish in a few hours. Soon (already?), human guidance may not even be necessary. Do I find this depressing? Just the opposite. Sure, I built my career on this skill, honed it for decades, and now a computer can do it in a hundredth of the time. In some ways, it even does better work than I ever could. But this was always a goofy way of making a living. And, in truth, it's always been one of the worst things about the job." It really depends, I think, on what you valued about the work. If it was for being recognized as deep and erudite, then AI isn't for you. But if what matters to you are the questions themselves, AI might be what you need. Imagine, for example, AI solves the problem of consciousness. It's a loss only if your job was based on perpetuating the question. "What do we lose if LLMs kill the journal system? Nothing but some heavy ass chains, and a few weak heuristics for assessing CVs quickly." Via Justin Weinberg, who offers a feeble response defending the continued counting of beans.

I can accept some version of the evolutionary description of human cognition (illustrated). As a map it probably suffers from some overgeneralization and misattribution, but I wouldn't think it's in any way fundamentally wrong. Human cognition is complex; nobody denies that. Where I disagree with Benjamin Riley is in the characterization of computational thinking, described as 'cognition' in another diagram, and composed of attention, reasoning, decision making, working memory and long-term memory. It's the usual use of the 'mind as computer' metaphor so many cognitive scientists love. But modern AI-based computation doesn't work like that at all. It's complex, so complex that explaining how it works is a major challenge. Is there any reason an AI could not function as described in the illustration? Probably not, though we would ask, why would it? It's the human brain that struggles against evolutionary limitations, not AI.