-
i think one exciting output of GPT and the like is that they give everyone a intuitive sense of what "unsurprising" or "low information" writing is like. if a machine can accurately predict what you're about to say next, it's probably not conveying much information.Permalink Look on archive.org
♻️ 13 Retweets ❤️ 67 Favorites Mood +7 🙂
-
which doesn't mean that the ideal for writing is some sort of superdense knot of meaning, which the reader must work at for hours to unpick (a shame, because my writing tends that way)Permalink Look on archive.org
-
sometimes you're writing an email and Gmail see you writing "I hope" and offers to autocomplete "this email finds you well", and, y'know, you do want to express an anodyne polite sentiment, that's fine, that writing is doing it's job.Permalink Look on archive.org
-
and if you spend much time online you're probably already familiar with low-information writing - the kind of things you find in small text at the bottom of a shop page, designed to be read by a Google spider & not by human eyes.Look on archive.org
-
but it's a new and strange thing to write within a box which gives you real time feedback on just how anodyne and predictable your phrases arePermalink Look on archive.org
-
it's not just me @rachelcoldicutt/1511017003431124994Permalink Look on archive.org