On reading old internet writing

The internet writing from roughly 2005 to 2015 is, on average, better than the internet writing from 2015 to now. The studio has been spending more time in the older archive lately and the difference is more pronounced than I expected.

A few hypotheses about why.

The earlier writing was not optimised for engagement. There was no algorithm to game. The writers were not getting paid by the click. The voice was, on average, more relaxed, more idiosyncratic, and more willing to be wrong out loud. The current internet rewards a particular kind of confident-sounding generality. The earlier internet did not.

The earlier writing was on personal sites. Substack did not exist. Medium did not exist. Twitter did, but its design pushed long-form writing back to personal sites where the writer could format the post as they wanted and link to the relevant pieces in the depth they wanted. The result was a network of substantial blog posts, linking to each other, that read as a conversation between writers who were not performing for an audience.

The earlier writing assumed less. Posts were written for a smaller, denser network of readers who knew each other’s work. The writer could refer to a previous piece, by name, expecting the reader to have read it. Most current internet writing has to start from scratch each time, because the audience is presumed to be drive-by traffic from a search engine. The result is more setup, less substance.

A few archives worth spending time in.

Cool Tools by Kevin Kelly. Twenty years of one person noting what works. A reading discipline I respect.

Ribbonfarm by Venkatesh Rao. Long, idiosyncratic, occasionally wrong, almost always interesting. The kind of writing that gets less common every year.

Daring Fireball by John Gruber. The archive going back to 2002. The post style has barely changed in over twenty years and the consistency is, in retrospect, remarkable.

Idle Words by Maciej Cegłowski. Pinboard’s founder. The talks are particularly worth reading.

A List Apart’s archive, particularly the issues from the late 2000s. Many of the articles are technically out of date and most of the writing on craft has held up.

The Pinboard archive of “popular bookmarks” from 2010-2015 is a useful index of what the design and software internet was paying attention to in that period. Different from the current attention map.

This is not nostalgia. Most of the older internet was as bad as most of the current internet. But the best of the older internet was published on personal sites by writers who were not optimising for anything other than being read by people who would also read the response, and the best of that is better than what the field is publishing now.

If you are trying to write for the web in 2024, the old archive is, in my experience, more useful reading than the current discourse. Read backwards. Most of the field’s mistakes are easier to spot when they have already played out.