I've been plagued by The Algorithm, Dead Internet Theory and the like for quite some time now. grading content quality by algorithmically driven success metrics (views, likes, shares, time watched, followers gained/lost, etc) is a doomspiral.

this is especially the case if you're like me, or at least the recent me, and want to write content that... matters. especially in the post-LLM AI Slop universe that the already dead internet is becoming. it is easier than ever to release content, easier than ever to consume content, and harder than ever to produce or consume content that matters.

as I've become more of a content creator media producer I've started following a lot of the cultural throughlines in different spaces. TPOT had its heyday on Twitter, faceless bikinni pics had their horny moment in the limelight, BlueSky and Threads became overnight successes and next-day graveyards. we're in a sort of... in-between time now on a lot of platforms. some platforms, like YouTube, don't change all that much. YouTube content, especially in the engineering space, is pretty banal, pretty uninteresting, pretty static.

but other platforms are very dynamic. Twitter, as much as I hate it, is a fairly useful reflection of overall culture, of where (US) society is, where it's been and where it's headed. a keen eye could have seen the mutterings of what would become Trump Administration Policy gurgle up from the recesses of mask-off natalist and nationalist TPOT.

twitter has also revealed some sad truths about platforms and web 2.0+ in general. the algorithm changes weekly, making it very difficult to get a feel for what, in any given week, will launch one into the algorithm and what will leave you in the dark. it is plainly clear to any critical eye that twitter's algorithm regularly has a thumb on it. when the Trump Administration released an executive order declaring "Antifa" a terrorist organization, it didn't rank in the top 25 trending topics on the site. there were plenty of high-engagement posts on the matter, tons of discussion on both sides, but absolutely zero reach from the Trending tab.

this is an example of "irregular" or "exceptional" algorithmic bias, though it isn't quite exceptional given how frequently it occurs. there are algorithmic biases that have become so entrenched in modern society that they're part of macroculture now. engagement and rage baiting are the norm. threads and posts very often follow incredibly obvious structural algorithms to maximize the potential for engagement and algorithmic boost. what "faceless" bikinni pictures did to instagram, thread templates are doing to twitter. the algorithm for creating these types of post became so easy to master and so widely known that it was a given that they would be implemented into the post-LLM bots that have infested Twitter today.

this is all to say: the dead internet theory is real, and it really isn't just because of bots. we have all been reduced to "NPC-like behavior" whether we like to admit it or not. when we look in the mirror, we all can think of half a dozen times (likely more) that we changed wording, made or deleted a post or responded to someone in a certain way because we knew it would "play well" on social media and in its algorithms. the internet is dead, and some of the zombies strolling its dark corridors have beating hearts instead of glowing wires.

the makerlog alternative

the alternative I am proposing is an experiment. as such, it is limited in scope to The Maker Class, and will involve highly dynamic testing. it also very well may not work or "catch on" in any meaningful way. that said, I think it is an important experiment to try out, and it is relatively simple.

makerlog guidelines

these are guidelines, not rules. as such, they don't have to be followed perfectly, at all times, but completely disregarding the guidelines defeats the purpose. "makerlog content" is defined as follows:

• content should pertain to something being "made" - I am a software engineer, but I do not want this to be limited to software. most of my friends push the boundaries between software development, media production, creative worldbuilding and a myriad of other interesting topics that produce an output. if you create music, writing, videos, software, hardware, metalwork, woodwork, poetry or any melding of the above, it counts.


• content should be "anti-algorithmic" - this is not to say that it should be made badly, or made in such a way that it rejects The Attention Economy entirely. this means you shoudl create your content in such a way that it is made for people, not for algorithms. take this blog for example. my primary platform in terms of reach is (currently) Twitter and YouTube. This blog will perform badly on Twitter (links are still being nuked at the time of writing) and would require a bit of work to transform into a video, that will also likely perform badly. I am writing this blog in Neovim, in pure HTML within a Svelte file. my only hope for "algorithmic reach" is in SEO, which, in my view, is a more pure form of algorithm that is a bit closer to the idea of "delivering relevant content to people who want it." this blog is written for you, the reader, not the twitter algorithm, the youtube algorithm or the tiktok for you page.


• content should be frequent - once upon a time, I said content should be daily, but I realized that my reasoning there was often based in what algorithms like (regularity) and not necessarily what is good for the creator and consumer.


• content should be as Actually Build In Public as possible - one of the more disappointing dead movements I've witnessed recently is the fast birth and faster death of the "Build in Public" movement (in the webdev/SaaS space, at least). obviously, people have been "building in public" for ages, in that they've been building things and talking about them publicly, but "the movement" that was born on Twitter rose to fame, declined to infamy and is now mainly just regurgitated garbage software slop and engagement bait written for algorithms in an attempt to make a quick buck off a big ProductHunt launch. what I am proposing is... not that. I propose returning to the roots: build things publicly, talk about them publicly, put out versions of the thing you're making publicly. if you are making music, don't wait until you're finished with production: take a video of that dope riff and drop it on YouTube for us to watch. if you're making software, release it iteratively and let people break it. if you're doing woodworking, show us what you built while the glue is drying. we have to go back to having joy in building things purely because building things is joyous.

these guidelines are intentionally incomplete and intentionally vague. the purpose is to get across a vibe that you can riff on in your own makerlog. this is designed as a multiplayer adventure in an open world.

as for me, I'm moving my makerlogs from Twitter to my blog. I like the freedom I get from being able to write them as Svelte components, free to create interactive and fun customizations like and other more impactful things in the future, like embedded interactive components.

you can still read all past Makerlogs, including Makerlog 8 on Twitter, head of thread here.