1. On Human curation and algorithms

    Human curation will always win.

    Why?

    Because humans are wired for the story, and curation is a specific type of storytelling.

    Because humans have ‘taste’ and taste is a great way to align interests. Find someone whose taste you admire and chances are that what they will lead you to new things as they share their influences, inspirations, and work (and so help you improve your taste).

    Because humans are the ultimate validators of what is useful and helpful to humans: there is no more powerful recommendation than someone you know and trust saying “this worked for me”, or “this is good” because they have tried or applied it in the real world.

    We have mischaracterised the early, more successful versions of the algorithms that powered our feeds and filtered our searches: we fetishised them as the work of machines. But the most successful algorithms were not machine curation but rather machines detecting and amplifying human signals at a scale that most humans could no longer operate at. As scale increases, the skill and dedication required to process information and keep it accessible increases. Case-in-point: Librarians (especially research librarians). Eventually something like Google Pagerank arrives, but the signals for the algorithm are still essentially proxies for detecting and amplifying human value and curation.

    But now the scale has changed again.

    When all content is viewed as a fungible commodity …

    When human edification, education, and entertainment are flattened into a single metric: engagement …

    When generative large language models are capable of auto-completing entire articles instead of merely words or sentences, and output homogenous reductions and remixes of what already exists without any new insight, context, or validation against reality …

    When there is no human value, no human creation in the production of vast amounts of ‘content’ …

    When the humans who do create content make it primarily as an offering to the algorithm in hopes that it will bless them with audience/engagement/cash …

    When those who control and (hope to) profit from the infrastructure of attention no longer value or respect the humans they claim to serve

    … then there is no human signal left to find and amplify, and there is no ‘taste’ to discover or develop, and the algorithms eat themselves, and all the while those who profit from those algorithms scream at us, “Are you not entertained engaged‽”

    The algorithms were always at their best as utilities to help you find your humans.

    Humans will always be

    • the best recommenders
    • the best discovery services
    • the best curators
    • the best at developing ‘taste’

    There is still a human-curated web out there under all the noise. It is more human and humane in both pace and scale. Opt-in!

    Go find your humans!

    Add their RSS feeds to your feed reader. Subscribe to their newsletters. Follow them in the Fediverse. Curate your inputs. Share wholeheartedly.

  2. Invisible tools

    A good tool is an invisible tool. By invisible, we mean that the tool does not intrude on your consciousness; you focus on the task, not the tool…Good tools enhance invisibility.
    …Unfortunately, our common metaphors for computer interaction lead us away from the invisible tool, and towards making the tool the center of attention.
    …magic is about psychology and salesmanship, and I believe a dangerous model for good design and productive technology. The proof is in the details—magic ignores them. Furthermore magic continues to glorify itself

    — Mark Weiser, The World Is Not A Desktop, ACM Interactions, January 1994

    See also: Not invented here

  3. Tech fashion

    …because tech is a pop culture, it will apply techniques and approaches because they are fashionable even when the outcomes are worse.

    — Baldur Bjarnason https://www.baldurbjarnason.com/2023/book-production-ai-spa/ (commenting on the computational limits of photography and the ‘fashion’ for software processing in phone cameras)

    I have been thinking a lot lately about applying pace layers to technology, particularly in the current climate of hype, where everything is declared a breakthrough (probably because slow, incremental progress toward stable, usable platforms is too ‘boring’)

    Bjarnsason provides another example of software moving at the speed of fashion when its users need it to move slower, be more reliable, to learn from fashion (and either adopt or reject as appropriate) rather than to be fashion.

  4. Stop and simply be

    When was young, I had to choose between the life of being and the life of doing. And I leapt at the latter like a trout to a fly. But each deed you do, each act, binds you to itself and to its consequences, and makes you act again and yet again. Then very seldom do you come upon a space, a time like this, between act and act, when you may stop and simply be. Or wonder who, after all, you are.

    — Ursula K. Le Guin, The Farthest Shore
  5. Writing for humans, by humans

    In truth, one reason I enjoy writing these newsletter so much is because they land in your email inbox, sorted by date and time, with my name as the sender. This takes enormous pressure off writing a damned title that is SEO friendly and clicky…Sure, writers must hustle relentlessly to promote their newsletter to potential subscribers, but they do not have to play the exhausting game of ‘can I make you click’ for each post…

    — Anne Trubeck, Notes from a Small Press - Amazon Is Changing How We Write

    A ‘captive’ audience of willing—paying, even!—subscribers means there is less pressure to play the SEO game. This doesn’t mean titles, subject lines, and headings don’t have to be well written, just that they don’t have to be keyword-stuffed—written for humans, not The Algorithm.

More entries can be found in the archive.