InfoMagpie’s Nest

What is this place?

Good question.

I've spent much of the last couple of years increasingly preoccupied with humans and intelligent technologies. I use that phrase deliberately. AI is an object category. Humans and intelligent technologies is a relationship and a problem space. And right now it is an extraordinarily strange, fast-moving and intellectually alive one.

We're in the middle of something genuinely new. Intelligent technologies are changing faster than our understanding of what people will do with them, how they'll understand them, how they'll incorporate them into work and life, what they'll trust, what they'll resist, what they'll misunderstand, and what happens when humans and machines don't behave quite as designers expect.

My curiosity wanders accordingly: behaviour, culture, adoption, trust, agency, decision-making, human–AI collaboration, organisations, research methods, conversational systems, epistemics, anthropomorphism, interfaces, knowledge, probably dogs occasionally because apparently nothing is safe.

This is where I collect some of the things I find while I'm wandering: papers, reports, questions, observations, half-formed hypotheses, connections between things that weren't previously sitting next to each other and miscellaneous intellectual pebbles.

Why put this on the internet?

Because there's a strange missing layer in how we usually show our thinking publicly.

A blog is written for someone else. Writing for other people and developing your own thinking aren't the same activity. My threshold for the first is quite high, which means there's often a very long gap between “that's interesting” and anything appearing publicly with my name attached to it.

During that gap, I'm reading, talking, testing things, changing my mind, and developing ideas that aren't ready to become arguments. This shows more of the stuff before that.

Why a nest?

This started with the idea of a digital garden: a place for ideas and interesting things to exist before they have been turned into finished work. I liked that idea, but I was less enthusiastic about anything that made me feel I had acquired 183 seedlings requiring responsible cultivation.

A nest feels more like me, because I collect things that catch my attention. Some connect to other things, some sit around for ages before becoming unexpectedly relevant and some are eventually forgotten about. The point is to leave more traces of the wandering itself.

How does this work?

There's a reason this lives inside ChatGPT. I use ChatGPT a lot as a thinking surface and intellectual sparring partner. Some conversations run for hundreds of pages: I think out loud, argue with ideas, make connections, change my mind and gradually work out what I actually think.

That process produces a lot worth keeping, much of it buried around page 173 of a conversation called something extremely informative like “Quick question”. So I built a plugin. When something interesting emerges, I can ask ChatGPT to identify the bits worth keeping and send the ones I choose here.

Basically: I think with AI and this gives some of that thinking somewhere to go.

Why InfoMagpie?

About twenty years ago, some friends described me as a bit of an information magpie. Or, less charitably, someone with an information-hoarding problem.

This was prompted partly by their noticing that I would pick up newspapers people had left behind on the Tube because, well, there might be something interesting in them.

This was not an isolated incident.

I've always collected information and knowledge wherever I find it. Before everything became digital, this mostly meant books. Lots of books. Boxes and boxes of them, some of which have subsequently moved countries with me.

Digitisation has not improved the situation.

There are now hundreds of Kindle books, Audible, multiple generations of read-later apps abandoned in my wake with hundreds or thousands of articles still waiting inside them, roughly 9,000 bookmarks, and a personal research library currently somewhere around 70GB.

To be clear, I have not read all of this.

I aspire to read all of this, which is quite a different and considerably less realistic proposition.

The pattern is remarkably consistent. I go looking for one paper and emerge half an hour later to discover 120 papers in my downloads folder. I'm rarely entirely sure what happened in between.

Essentially: information exists, might contain something interesting, and apparently cannot be left unsupervised.

The problem, of course, is that collecting information and actually doing something with it are very different activities. No human brain can retain this volume of material in any useful level of detail. Mine certainly can't. Most of it has historically accumulated without ever becoming notes, connections or anything else I can reliably find and use later.

I have explored many knowledge-management and note-taking systems, and none has ever stuck for long. The biggest problem is the friction introduced by having to stop thinking, switch systems and start writing things up. That interruption is usually enough for the system eventually to fail or simply be forgotten about.

For the past couple of years, I've been experimenting with AI partly as a way of changing that equation: helping me explore, process, question and connect more of what I collect than I could otherwise manage.

Of course, using AI introduces its own forms of compression and loss, too, but perfect retention is never going to be realistic anyway because the human brain and memory have limitations too.

Increasingly, the process of exploring all this material produces another kind of information problem: very long conversations with AI, full of questions, distinctions, connections and ideas that I then can't find again.

So this is my latest experiment. InfoMagpie's Nest is where I rescue some of them.

Filters

Show Source type

Opening the Nest…