Earlier this week Amanda McCellon, a city government worker and community organizer in the greater Oklahoma City metropolitical area, finally decided to pray to the algorithm.
“Dear Algorithm,” began her post on Threads, Meta’s newish Twitter clone. From there, McCellon asked to be connected with 10 topics and like-minded folks on Threads, including members of the LGBTQIA community, people in local politics, dog moms, horror movie fans, lovers of the TV show The Office, and “late in life bloomers.”
McCellon had been on Threads since it launched this summer as an extension of Instagram. She had high hopes for its potential as a community-building tool and a balm for all the chafing that came with late-stage Twitter. But she quickly discovered that it was hard to, well, discover people on Threads. Hence the Dear Algorithm plea.
McCellon is not alone. Over the past couple of months an uncountable number of Dear Algorithm posts have popped up on Threads as people on the platform have tried to find their people. Some posts are delightful, riffing on the fresh chords of a new social network, one that’s still finding its rhythm. They’re also highly specific. In a WIRED analysis of more than two dozen such posts, Threaders search for “UX professionals,” “foodies,” “fellow celiacs,” “strong female leads,” “disability advocacy,” “introvert life,” “dermatologists,” “Disneyland Club 33 members,” “weed,” “silk-screeners,” “Panic! At the Disco,” “the ghost of great men who were seduced by the one ring of power forged by Sauron to gain dominion over the free people of middle earth,” and “a meeting with Tim Burton.”
There’s just one problem: Dear Algorithm posts don’t work.
Meta confirmed to WIRED that sharing a Dear Algorithm post does not affect the posts users ultimately see on Threads. The app’s For You feed is personalized to each user based on a number of signals, Meta spokesperson Seine Kim says, such as accounts and posts that a user has interacted with. These interactions extend to both Threads and Instagram, since Threads is an extension of the Instagram API.
Threads via Lauren Goode
Jonathan Stray, who researches recommendation algorithms at the Berkeley Center for Human-Compatible AI, says this approach is consistent with other social media platforms. “Engagement, in various forms, is definitely the primary signal used by content ranking for essentially all platforms,” Stray says in an email, noting that one of the few exceptions is non-personalized site rankings, such as on Reddit.
So lingering on celebrity accounts, liking delectable food porn, or responding in comments to dog-mom posts on Instagram is very likely going to influence the kind of content that appears in your Threads feed. But simply putting out a text-based plea to the faceless Meta gods (or algorithms) running Threads won’t do the trick. The Dear Algorithm posts have all of the early-days energy, then, of Twitter’s #FollowFridays, which helped users find other cool people, or the sweetly desperate vibes of the copy-paste hoaxes that convinced users Facebook wouldn’t own their uploaded photos if they just posted a copyright warning to their feeds. Of course, #FollowFridays actually worked; copy-pasting poorly written legal jargon demanding your rights from Facebook did not.
The Dear Algorithm trend also suggests spam is starting to infiltrate Threads: In researching these posts on Threads, WIRED observed more than 100 rapid-fire repeats of the same list, shared by multiple accounts within a short time frame, all of them including the phrase “Lover’s of Jesus” in the post. (This is not to suggest Jesus is spam; just the repetitive messaging.) Meta said it is looking into the issue.
Some of the early problems with discoverability on Threads, which claims just under 100 million monthly active users, could be eased with the introduction of hashtags. On Wednesday, Meta chief executive Mark Zuckerberg said that the company was testing tagging on Threads in Australia. Once those are widely rolled out, tagging one of your interests in a Dear Algorithm post might make it easier for other people to find and engage with said post. But Meta hasn’t said when tags will be supported on Threads in other countries, and for now users can add only one tag at a time to a Threads post.
“I’m still hopeful that Threads will bring more people to me who I have things in common with,” says McCellon, the organizer who is trying to galvanize her community around local politics in Oklahoma. “I do find that most of the people who are following me there are already following me on Instagram, but that’s not necessarily helpful if you’re trying to organize or build a new community.”
But, McCellon concedes, if Threads connects her with just one or two people who get a new perspective on local politics, “then the algorithm has done its job.”
