- Garbage Day
- Posts
- Some big news...
Some big news...
Read to the end for what three soda pops no lunch feels like
Another Surprise For You All
OK gang, as I’ve been saying, I’ve been working behind the scenes on some big stuff over the last few months and today I can finally share something I’ve been absolutely buzzing about for weeks.
The Garbage Day team is getting a little bigger today. Please help me in welcoming Cates Holderness, who is coming on as managing editor. If you’ve never heard of Cates, well, you actually have. I’d argue there are few people in the world who understand the internet as well as she does. She’s had her fingerprints all over the social web for more than decade now. We first met deep in the content mines at BuzzFeed in 2012, where she taught me everything I know about Being Online (and where I showed her 4chan for the first time lol). In 2015, she was the mastermind behind The Dress. And she’s spent the last six years at Tumblr, where she completely reactivated the site’s community in ways only she could.
Cates is joining Garbage Day at a pivotal moment for us. We’re in spitting distance of 100,000 readers, our podcast Panic World is now doing more than 100,00 downloads a month, and the work we do, as weird as it is for me to say, has never felt so important. You know, because of the horrors and such. She’ll be overhauling Garbage Day as a business, revamping our community features like the subscriber-only Discord and our live events, and, most importantly, helping me make Garbage Day as good as it can be for you, dear reader.
Hiring our head of research Adam Bumas earlier this year and Cates this week would not have been possible without the overwhelming support we’ve received from Garbage Day’s audience. I feel so lucky and privileged that we get to expand like this. There aren’t a lot of big wins in the world of digital media these days and none of this would have been possible without you.
Think About Supporting Garbage Day!
It’s $5 a month or $45 a year and you get Discord access, the coveted weekend issue, and monthly trend reports. What a bargain! Hit the button below to find out more.
Automating The Failures Of The Previous Internet
Molly White, the author of Web3 Is Going Great and Citation Needed, published a piece over the weekend about a custom ChatGPT bot called “Looksmaxxing GPT,” which is basically an incel bot.
“Some of the feedback is explicitly racist,” White wrote. “With the chatbot identifying one set of photos as ‘Low Normie / High Subhuman Borderline (White)’ and offering to explain ‘how and why being White (Caucasoid) helps in the context of PSL rating and SMV classification’.” As of this morning, Looksmaxxing GPT is #6 under “Lifestyle” on the custom GPT store.
I covered the launch of ChatGPT’s custom GPTs for Fast Company back in 2023. At the time, there wasn’t really a discovery tool for them. Users could build chatbots with custom instructions and share them with each other. But it was clear this was the direction OpenAI saw AI tools going. I wrote at the time that it was part of a larger trend of tech platforms selling your own filter bubble back to you. But my thinking about this has become a bit more nuanced over the last two years.
What the AI arms race has actually done is codified and automated all of the failures of the previous internet era. Extremism, misinformation, harassment, non-consensual sexual material, and scams — all the content that tech companies promised to fix, but never could at scale — are now trapped in some AI’s black box of data. And what isn’t already part of a mainstream AI service like ChatGPT can easily be ported into an open source model and offered to users with zero regulation or oversight. And part of me wonders if that was the point all along, at least partially.

(ChatGPT)
In the 2010s, huge tech platforms like Facebook and Google pitched themselves as sleek corporate alternatives to the wild west of the web. They promised that, by using their sites, you could be safe to read, share, and, most importantly, buy anything you wanted online. And for a brief moment, that was true. Things like eating disorder content and ISIS videos were effectively eradicated from users’ feeds. Non-consensual sexual material and child sexual abuse material were rare enough to be newsworthy. But that work is tedious, perpetual, expensive, never perfect, and, most important for be-hoodie’d neoliberal tech companies, inherently political. By the middle of the decade, reports began emerging of traumatized moderators, country-level bans of social networks, and strange, new identitarian movements crafted to bypass filters. And as both Facebook and Google chased impossible scale, desperate to connect the entire planet, it became clear that their initial promise — infinite feeds of safe and vetted user-generated content that advertisers could slot themselves into — was impossible. A brand safe internet is incompatible with the way human beings innately use a computer, it seems.
And so these companies faced a dilemma. Do they tell their investors that moderation at scale is impossible and will require a never-ending budget to burn through to do it even poorly? Do they give up on scale as a goal, turning their back on the techno mercantilism that propelled them to huge heights? Or do they invent a magic tool that just so happens to fix all the problems they couldn’t? The third one, obviously.
Which brings us back to the incel bot. It also fits perfectly into my new understanding what generative-AI bots are actually for, based on my own experience using one for therapy this winter. The marketing has been about productivity or, more laughably, creativity, but the point is emotional dependence. The same dopamine-fueled business model perfected by the social platforms of the last decade. Which is why if you want to know where this is all headed, imagine an incel bot for every bad internet thing you can imagine. A gamergate bot, an eating disorder bot, a QAnon bot, an terrorism bot. Some of these will slip under the radar of big AI companies, some will be more niche and found on the darker corners of the web, and some you won’t ever even know about. Until they’re already a problem, of course.

Garbage Day is doing a three-night residency at Baby’s All Right in Brooklyn this July. Tickets are officially on sale. We’re going to try and save democracy in America. Or, at the very least, figure out how we broke it. Each night has a different theme and different guests. You can grab tickets for each night by clicking the links below.
Let’s Check In On How Google’s Gemini Is Going
Who Punched Elon Musk?

(Photo by Tom Brenner For The Washington Post via Getty Images)
Elon Musk appeared in the West Wing over the weekend with a visible black eye. He was asked by a reporter about it, to which he quipped, "I wasn’t anywhere near France," a reference to a video that emerged last week of French President Emmanuel Macron being shoved by his wife Bridgette while getting off a plane. Musk eventually explained that he got the shiner after “horsing around” with his son X. “I said, ‘Go ahead, punch me in the face.’ And he did,” he said. All very normal.
The black eye has quickly turned into a bit of a conspiracy theory among liberals and leftists. Seeing as how it showed up only days after it was reported that Musk is leaving the White House and taking Stephen Miller’s wife Katie with him. She’ll be working across Musk’s various companies, but many users are convinced this is some weird IVF breeding thing in the works.
“This is a black eye delivered by someone who can’t throw a punch and Stephen Miller’s wife just left him for Elon Musk,” X user @TheWapplehouse wrote. “Also, Stephen Miller is left handed. Probably unrelated.”
In other Musk drama news, Steve Bannon is gunning for him. As I wrote on Friday, about completely different example of fascist infighting, the only thing right-wingers hate more than themselves, and each other, is weakness. And Musk is looking real weak right now.
Asmongold Content Warning
—by Adam Bumas
Zack “Asmongold” Hoyt, North America’s top Twitch streamer and proud warrior in the fight against deodorant, had a busy weekend. On Sunday, he appeared alongside other streamers in a new video by very online rapper bbno$ (pronounced “baby no money”). Within a few hours, though, the video had been updated with a cartoon lizard pasted over Hoyt’s face, and a note reading, “At the time of recording, I was unaware of certain individuals’ ethical beliefs. As a result, I’ve chosen to censor those whose values I do not support.”
So what are Hoyt’s values, besides the deodorant thing? You can get a sense from Hoyt’s other announcement the day before. He said that he was going to start streaming to multiple platforms, but focusing primarily on X (with personal support from less honorable gamer Elon Musk) and on the offshore casino-owned platform Kick.
These plans have been in the works for months, after Hoyt was temporarily banned from Twitch for some Deep Thoughts about Palestinians. Given all that, bbno$’s decision makes more sense — though presumably, he was also standing close enough to Hoyt while making the video to pick up on the smell.
The Bluesky Slump
There are conflicting reports about the health of Bluesky at the moment. It managed to capture a decent amount of prominent Democrats post-election, even if they all hate using it. And while daily like counts aren’t spiking like they used to last year, it has landed, for now, at a pretty healthy plateau. But, according to a recent Pew study, Bluesky adoption hasn’t actually replaced X usage. The amount of what Pew calls “news influencers,” doubled after the election, but those “news influencers” are still just as active on X.
(This is where I take a second to tell the readers who complain that I’m still highlighting X content that it is still, unfortunately, relevant.)
It’s also possible that Bluesky just feels like its slumping because of the users on there. The initial cohort of ornery, but entertaining Weird Twitter refugees has been totally overwhelmed by MSNBC boomers who are extremely aggressive. They also don’t appear to understand how the site works. There have been a lot of concerns recently that Bluesky’s “algorithm” is suppressing users, which isn’t possible, because there isn’t an algorithm that can do that. As TechDirt’s Mike Masnick reminded folks last week, Bluesky lets anyone make their own algorithmic feeds.
Which is all to say, social networks live and die on vibes and if Bluesky's vibe feels like it’s petering out it doesn’t really matter if it’s true or not.
A Good TikTok
Some Stray Links
P.S. here’s what three soda pops no lunch feels like.
***Any typos in this email are Cates’ fault***
Reply