8 minute read

_Weeknotes are a habit I'm cultivating where I share writing (and some links) as a thinking-in-public process. The idea is to explore ideas I'm grappling with, primarily in my professional life, without worrying about them being full-formed. The potential visibility of these notes is a nudge to develop them a bit more than if they were private. While my audience is mostly theoretical, if you're reading this, please understand these are meant to be exploratory and provisional._

quote: Ursula K. Le Guin on fiction being one of the best ways of learning about other people

“Fiction as we currently think of it, the novel and short story as they have existed since the eighteenth century, offers one of the very best means of understanding people different from oneself, short of experience. Fiction is often really much more useful than lived experience; it takes much less time, costs nothing (from the library), and comes in a manageable, orderly form. You can understand it. Experience just steamrollers over you and you begin to see what happened only years and years later, if ever. Fiction is much better than reality at providing useful factual, psychological, and moral understanding.”

The Wave in the Mind: Talks and Essays on the Writer, the Reader, and the Imagination - Ursula K. Le Guin

thinking about: what I find interesting about the current generation of AIs

I’ve been playing with AI again because a few AI plugins for the note-taking app Obsidian reached the point where they seemed ready for prime time. These plugins (co-pilot and smart connections) let you chat with different models within Obsidian itself, use your own notes as context, and create embeddings from your notes (which I don’t fully understand, but seems to be a low-cost way to help the systems understand the context of each note and find connections between them). Both these plugins allow using local AI models, which is terrific, but honestly, I have been using them with OpenAIs ChatGPT and with Anthropic’s Claude via API.

Obsidian contains all my notes, draft documents, projects files, journal entries, etc. going back at least 3 years. It also has all my reading notes and highlights from almost 10 years of e-books, internet articles, websites, and academic PDFs. Basically, everything text that I write or consume ends up in Obsidian. I like that it is very fast, uses Markdown, and has a frighteningly passionate community of users and developers.

I hate the way most people are using AI, which seems to be having it generate content from chicken scratch. It’s generic and horribly enthusiastic. I increasingly get 3 paragraph emails from students that sound like they were written by a committee of cocaine fuelled project managers.

What I like about AI is the possibility of a tool I can use to take my own ideas, in all their poorly formed and idiosyncratic weirdness, and probe them from different (imagined, fake, weird) perspectives and/or remix them with other ideas and into new ideas, outcomes, and formats.

What I want from AI is an on-demand committee of cocaine-fuelled Franklins remixing and probing my ideas with me.

So for the last few weeks I’ve been using AI for this purpose. Among other things, I’ve used it to help plan a day-long design thinking event, prep for hiring new student ambassadors, and learn more about my Dungeons and Dragons character (I am part of my first ever campaign with some friends right now).

What have I found useful?

Being able to use one or more notes as context for feedback on another note:

  • Give me actionable ideas for improving an event plan based on my highlights from a book I read a year ago about event planning.
  • Use my therapy notes to predict issues I might have with a given project (lol, I knowwww… but its advice was frighteningly on point).

Asking for feedback from the perspective of different personas:

  • Give me actionable feedback about this plan from the perspective of an instructional designer, an event planner, and a design thinking expert.
  • Write a FAQ for someone attending this event who has never attended a design thinking event or similar activity, using the information in this plan for context.
  • Create draft advice for a first-year student applying for this job about what makes a successful application package.

Integrate ideas from different notes to find points of convergence or do other things:

  • Brainstorm ways that ideas from note 1 could be combined with note 2 (one of my favourite uses because it can get weird quick and that is the point)

Remixing ideas or documents into different combinations or forms

  • Using the project plan, create drafts of supplementary documents, including a guide for attendees, a facilitator guide, and a guide specifically for person X and Y.

(These examples aren’t the actual prompts I’ve used, which usually requires a lot of imagination and direction. I’ve found working with AI is more like writing the AI a story than it is like coding)

What this gives you isn’t immediately usable, instead you get a good idea or two, some inspiration, part of a paragraph, a few question/answer pairs, etc. You then need to edit those, integrate them into your document, and repeat the process.

Does this save time? Yes and no. It takes more time than doing nothing, which given time constraints and under-staffing is what would normally happen. It takes less time than doing these things manually.

Is this better than doing them manually or having another person do these things? Again, in my experience so far, yes and no. If I could have a world expert in event planning, design thinking, and instructional design next door to my office who would provide immediate and tailored feedback, that might be better. Sometimes it might be worse. For the remixing and idea generation, I’m not sure it would be better… would I ever ask someone to read through my therapy notes and use them to generate failure points for a project? Or iterate through connections between IDEA A and IDEA B when neither is connected or related to a current project? Probably not.

photo: spring flowers

I liked this article about why We Need to Rewild the Internet and applying an ecological metaphor to thinking about why the internet has become so terrible. It also had me thinking about the ways that all disciplines are increasingly “crisis disciplines” concerned with how we can save the things we love. I definitely thought a lot about academic institutions while reading this, and the ways we have been captured by the same forces threatening the internet.

“Ecologists have re-oriented their field as a “crisis discipline,” a field of study that’s not just about learning things but about saving them. We technologists need to do the same. Rewilding the internet connects and grows what people are doing across regulation, standards-setting and new ways of organizing and building infrastructure, to tell a shared story of where we want to go. It’s a shared vision with many strategies. The instruments we need to shift away from extractive technological monocultures are at hand or ready to be built.”

Since I am talking about what I like about AI, it’s good to share that I really appreciate Garbage Day’s regular dives into all the ways that AI content is terrible and weird and destroying what’s left of social media platforms. This from The AI Flight Attendants of Facebook is illustrative. Garbage Day is absolutely worth paying for.

“I assume these pages are simply jamming a bunch of popular stuff together to farm engagement to eventually monetize in some way down the line. Why AI images? Because you can flood Facebook with thousands of posts and the platform won’t really do anything about it. These pages are also using the platform’s built-in 3D photo filter, possibly to bypass Facebook’s bar-is-in-hell bare-minimum AI image detection. Why flight attendants? Because Facebook users are, and always have been, uncontrollably horny. But, also, my mom is a flight attendant (sorry mom if you’re reading this!) and aviation and flight attendant Facebook has always been huge. So I think they’re just identifying communities that were already active and swarming them. Why Jesus? Because religious content — and getting users to say “Amen” underneath it — became one of the fastest growing types of content on the site after it stopped promoting news content last year. “

As a 43-year-old bearded man I can practically feel my bird-guyness flying towards me through time. Ed Yong When I Became a Birder, Almost Everything Else Fell Into Place

“Birding has tripled the time I spend outdoors. It has pushed me to explore Oakland in ways I never would have: Amazing hot spots lurk within industrial areas, sewage treatment plants and random residential parks. It has proved more meditative than meditation. While birding, I seem impervious to heat, cold, hunger and thirst. My senses focus resolutely on the present, and the usual hubbub in my head becomes quiet. When I spot a species for the first time — a lifer — I course with adrenaline, while being utterly serene. “

I liked this episode of Weird Studies on the power and necessity of being pretentious. Make Believe: On the Power of Pretentiousness Weird Studies

“In culture and the arts, labeling something you don’t like (or don’t understand) “pretentious” is the easy way out. It’s a conversation killer, implying that any dialogue is pointless, and those who disagree are merely duped by what you’ve cleverly discerned as a charade. It’s akin to cynically revealing that a magic show is all smoke and mirrors—as if creative vision doesn’t necessitate a leap of faith. In this episode, Phil and JF explore the nuances of pretentiousness, distinguishing between its fruitful and hollow forms. They argue that the real gamble, and inherent value, of daring to pretend lies in recognizing that imagination is an active contributor to, rather than a detractor from, reality.”

Tags:

Categories:

Updated: