This cartoon has four panels, all featuring the same three characters. Charlie, a white man in a suit and tie. Ezra, a white man in more casual clothing. And Reader, a Latina wearing shorts and a black tank.
PANEL 1
Charlie and Ezra are walking together. Nearby, Reader sits at the base of a tree, reading a book, and overhears.
CHARLIE: We need Nuremberg-style trials for tranny-affirming clinic doctors.
EZRA: As a liberal centrist, I can’t agree with that. But what matters is that we’re talking.
PANEL 2
The woman looks annoyed.
CHARLIE: Democrats want Mexicans to overrun us because they hate America and wanna see it become less white and collapse!
EZRA: That’s not true. But again, we’re talking! Thank you for practicing politics the right way!
PANEL 3
The woman stands up, yelling angrily at the two men.
CHARLIE: You know what happens in the cities? Blacks prowl around attacking white people for fun! Haitians rape your women and hunt you!
EZRA: Again, I can’t agree. But I–
READER: Fuck that racist bullcrap!
PANEL 4
Ezra and Charlie walk on, not speaking to the woman, who watches them leave with an annoyed expression.
EZRA: Tsk! So uncivil! That’s the kind of intolerance that’s ruining America.
CHARLIE: They should deport her!
CHICKEN FAT WATCH
“Chicken fat” is forgotten cartoonist lingo for unimportant but hopefully amusing stuff in the art.
PANEL 1: A notice taped to the tree says “DON’T don’t don’t don’t don’t don’t forget about me baby,” a reference to a song by Simple Minds made famous in The Breakfast Club.
PANEL 4: A heart carved into the tree trunk says “N.L. + S.T.” Another heart says “J.T. + J.B.,” but has been crossed out. A third heart says “A.H. + J.B.” All of these hearts refer to one of my favorite musicals, Sweeney Todd.
Beaker from The Muppets is sticking his head out a hole in the ground.
A rat is walking on the street next to the sidewalk, looking distressed as it reads something on its phone. It’s wearing a shirt with a hearts pattern.
A piece of litter on the ground says “REPENT. Panel 4 is upon us!”
I’m in the process of rewriting a talk I gave a few weeks ago in order to record it and publish it online. When I mentioned that to a friend who’s a filmmaker and knows how lighting and all that works he offered to help me doing it “right” so it doesn’t look like shit. Which – if we are honest – my recording totally might.
I didn’t go for his help. Not because I do not know that his expertise would have helped but because I wanted to see what I could do myself, even if it’s a bit shit. Which is maybe not smart and totally goes against how “the market” works.
It has never been easier and cheaper to produce stuff – especially media – of massively high production value. Even smartphone cameras are good enough to record very impressive videos – given a bit of lighting. Microphones are cheap. Very good cameras and a lot of information on how to use them and edit the photos are affordable. Even open source content management systems (like the WordPress thingy I have going here) come with very sleek and professionally looking themes/appearances. It’s almost hard to produce something truly shitty-looking today.
This abundance of cheap tools could be seen as almost utopian. Tech-influencers and -executives will call this “democratization”. And even for people already making things this is very alluring: Your things could be so much better if you just did a few things, used a few tools.
And many of us are profiting off of this every day. My favorite type of media (the long-form video essay) for example has been in sort of a golden age for a few years with the production value of these films going up every year. But I keep wondering if that is always a good thing. Like why didn’t I accept my friend’s help in making my recording good? (Well aside from all the usual bullshit and insecurities in my head, but let’s ignore all that.)
I think the level of quality that everything has these days, every image, every website, every song, every video also has downsides. It defining the “norm”, the expectation we have from one another has two main issues that keep irritating me.
The Second Job
Getting cheap tools of high quality is one thing but tools don’t work by themselves. They still need a lot of experience to use them properly, to get the results these tools are advertised with. Yeah, you can build a small home studio to record yourself for a small 4 digit EUR budget, but can you actually use those things? So now you have to dig into all kinds of other things, into manuals and howtos.
Don’t get me wrong, nobody loves a rabbithole as much as I do, but did you really need another extra hobby/job? What did you want to do when you started? Say you wanted a website to publish your writing. Now you are diving into the specificities of your chosen CMS and try to bend it to something you saw somewhere and it doesn’t want to budge. But you wanted to write, right? How’s that working out for you?
We are inventing extra jobs for ourselves and put them in between us doing the thing we wanted to do. Because professionalism.
This also leads to us having a harder time to be happy about our work because it’s harder to determine what’s “not good enough”. Are you unhappy with your writing or the way it looks? Would the writing have been better of you hadn’t dicked around making sharepics for Instagram (because that’s how you run a website!)?
In the end all the extra work can make it so much harder to get people to do the things they actually wanted. Because they wanted a professional thing and not some janky 90ies looking site. But why not? Maybe it’s not your business to run a perfect publication and maybe it doesn’t need to be. What did you want to do before professionalism clouded your mind? Why not just do that?
Too Weird
Professionalism not only derives from tools of course. Especially in algorithmically driven platforms (think Youtube or Instagram) – but also in others – there’s a sort of grammar to learn based on the properties of the platform, what it “rewards” and what its users expect. There’s a reason everything on Instagram looks like advertising. Because it is.
So say you want to make videos and have them on youtube. You quickly learn the lengths that work on that platform, you learn how the images are supposed to look so people click them, how to do the greeting/introduction. Every platform establishes very narrow, very rigid rules for the content that is successful there.
But it’s mostly documented, right? There’s so many howtos there telling you how to do it, maybe even some “AI” to help you! Cool.
But now that the thing that you wanted to make has been reshuffled and reshaped into something platformy, is it still what you wanted it to be? Is it still you? Does it make you happy? Well maybe the success makes you happy, fair.
Professionalism has a strong normative quality: That is the reason that professionals have standard processes and workflows (which enable working together easier and help ensure quality) but that’s also the reason why you cannot really distinguish individual bankers or tech-bros from one another: They all look and talk the same way.
Professionalism tries to limit the “weird”, the “unruliness”.
Of course: It’s capitalism
All this of course comes from capitalism. The push for professionalism comes – to a large degree – from the thought that if you put time into it, you maybe can make some money off of it! Maybe it can be your job! Everything needs to be a business or a proto-business.
But it shouldn’t.
If you are trying to make a living off of your writing or drawing or video-making or crocheting or whatever, that’s cool. I am rooting for you.
But not everything needs to be. Not everything needs to do the professional things. The growth hacks and the pro tools. Some things can also be just a person doing a thing and putting it … somewhere. Doesn’t matter if the platform or the form ain’t all that. It’s just a thing someone did and wanted to share. In all its potential jankiness.
So yeah, I’ll probably record my talk on my own. And this has totally not been a very long-winded way for me to legitimize my bad orthography.
If an Iranian taxi driver waves away your payment, saying, "Be my guest this time," accepting their offer would be a cultural disaster. They expect you to insist on paying—probably three times—before they'll take your money. This dance of refusal and counter-refusal, called taarof, governs countless daily interactions in Persian culture. And AI models are terrible at it.
New research released earlier this month titled "We Politely Insist: Your LLM Must Learn the Persian Art of Taarof" shows that mainstream AI language models from OpenAI, Anthropic, and Meta fail to absorb these Persian social rituals, correctly navigating taarof situations only 34 to 42 percent of the time. Native Persian speakers, by contrast, get it right 82 percent of the time. This performance gap persists across large language models such as GPT-4o, Claude 3.5 Haiku, Llama 3, DeepSeek V3, and Dorna, a Persian-tuned variant of Llama 3.
A study led by Nikta Gohari Sadr of Brock University, along with researchers from Emory University and other institutions, introduces "TAAROFBENCH," the first benchmark for measuring how well AI systems reproduce this intricate cultural practice. The researchers' findings show how recent AI models default to Western-style directness, completely missing the cultural cues that govern everyday interactions for millions of Persian speakers worldwide.
"Low effort, unhelpful AI generated work is having a significant impact on collaboration at work. Approximately half of the people we surveyed viewed colleagues who sent workslop as less creative, capable, and reliable than they did before receiving the output. Forty-two percent saw them as less trustworthy, "
Consider upgrading to a paid subscription. It’s $5 a month or $45 a year and you get three issues a week and Discord access. We’ll be going to just one free issue a week starting next week. Hit the button below to keep getting every issue of Garbage Day.
Friendship Ended With First Amendment, Now Authoritarianism Is My Best Friend
Late night host Jimmy Kimmel was taken off the air this week, following pressure from both the Federal Communications Commission and Nexstar Media Group. CNN has a good ticking clock of the conversations behind the scenes that led to Kimmel’s “indefinite preemption.” A source told CNN the decision was ultimately made after FCC chair Brendan Carr threatened to pull ABC’s broadcasting license during an appearance on conservative influencer Benny Johnson’s podcast.
The New York Times has a good piece on the administration cracking down on entertainers that refuse to tow the party line. A spokesperson for the administration told The Times that, contrary to what the media has reported, they actually have a good sense humor about themselves, but they no longer have “the time nor the patience to apply that method to the ‘miserable literati.’” Oh, sorry, that article was actually published in 1939 and written about a totally different government. Disregard.
Carr, via an animated GIF of Jack Nicholson nodding, seemed to confirm that the pressure the FCC is currently putting on Trump’s enemies was laid out in Project 2025. In fact, it was laid out in a chapter he, himself, wrote! The bulk of the chapter is about scrapping Section 230 and reigning in Chinese tech, but there’s also a lengthy section about how much power an FCC chair has, which includes a very relevant passage that reads, “Any merger that involves a wireless company, broadcaster, or similar entity that holds an FCC license must obtain FCC approval (assuming that the merger will involve the transfer of the FCC license).” And that’s what this is all about.
As Business Insiderwrote, there’s a $6 billion deal at the heart of the Kimmel affair. Nexstar is currently trying to merge with another TV station conglomerate, Tegna, and are currently waiting for FCC approval. Adding to the pressure to take Kimmel off the air is Sinclair Broadcast Group, who are replacing Kimmel’s Friday episode with a special about Charlie Kirk and want Kimmel to donate to Turning Point USA. Sinclair’s massively conservative slant was best documented by Deadspin in 2018, who made a supercut of their local news affiliates all reading the same pro-Trump script.
Just to be clear, to appease any lawyers reading this, Carr, Nexstar, and Sinclair are all claiming there was no coordination here. But, also, Carr was quick to thank Nexstar on X for preempting Kimmel, writing, “It is important for broadcasters to push back on Disney programming that they determine falls short of community values.”
Kimmel’s suspension has given us the clearest picture yet of how the Trump administration intends to deal with the pesky First Amendment. Put pressure on the massive companies that uphold the American media landscape and assume they’ll fall in line. And install political allies at the companies they can’t directly pressure. Like Oracle’s Larry Ellison, who is currently in position to take over the American version of TikTok. Which is why esports journalist Rod Breslau speculated on X this week, “Trump is about to put 10% tariffs on all Amazon products if Twitch doesn't ban Hasan and Destiny (again) by next week.”
There are very few American media institutions that President Donald Trump can’t put his thumb on currently. And he knows it! While queening out with reporters on Air Force One on Thursday, Trump floated the idea of revoking the license of any broadcaster that gives him “bad publicity.” And other Republicans are quickly falling in line, seemingly relieved they no longer have to care about free speech. Republican Sen. Cynthia Lummis told Semafor this week, “I tend to think that the First Amendment should always be sort of the ultimate right. And that there should be almost no checks and balances on it. I don’t feel that way anymore.”
If you’re feeling hopeless, don’t worry, all the historians I follow are crashing out too. But if you are looking for something proactive, The Onion’s Ben Collins wrote on Bluesky, “Becoming increasingly clear we’re gonna have to build a parallel infrastructure for all the media we really love. The reason all of this is happening under the color of law is hyperconsolidation, dissent being traded straight up for merger approval, or fear of harassment.”
The wave of indie media orgs that have emerged since the pandemic is a good start, but they are still no match for the gargantuan Trump-aligned corporate media machine. A journalist, pundit, or, as we’ve seen this week, even an entertainer that gets too loud will be targeted by this administration. They will investigate you, go after your livelihood, and pressure the platform you’re broadcasting from. These people are petty, powerful, and pathologically online. One Trump appointee recently told Semafor that they were “radicalized” by “being booted out of the Gawker comments section, way back in the day.” No one is off limits now and there is no limit to how dumb this will all get. We need to figure out how to get news and information to people without YouTube, TikTok, Instagram, Twitch, broadcast TV, cable, or corporate streamers. Our government is being run by aggrieved elites who were thrown onto social platforms with everyone else in the 2010s, where they could finally read what we all think about them. And it caused them so much psychic damage that they have decided to destroy both the internet and democracy to make sure nothing like it will ever happen again.
On October 9th, we’re taking over TV Eye for a big party in Ridgewood. It’s going to be a bit different from the live events we’ve done before, but it should feel substantially garbage-y. Special musical guests include Kill Alters, Reagen Holiday, and RGEM. Click the big green button below to grab your tickets while you still can!
In case you didn’t already think we live in a simulation, Meta’s Chief Wearables Officer is named Rocco Basilico. If you don’t immediately get why that is completely unbelievable, allow me to explain.
“Roko’s Basilisk” is the name of a very infamous forum post among AI evangelists. A user named Roko posted it to the LessWrong message board back in 2010 and it was considered so dangerous — or, perhaps more accurately, annoying — that any mention of it was banned from the site for years. LessWrong’s founder Eliezer Yudkowsky once called it an “information hazard.” So what’s so special about this forum post?
Well, it’s a little dense, but the simplest summary I can come up with is that Roko posited that a super-intelligent AI would come online sometime in the near future and retroactively punish anyone who didn’t help it come online. Roko argued that knowledge of the AI’s eventual awakening — which all LessWrong users would now have from reading Roko’s post — would mean that it would punish anyone who didn’t basically spend all of their income supporting the AI industry. There’s a few other dimensions to Roko’s post, but I assume Garbage Day readers are normal and well-adjusted enough to not need to care about any of that.
But you can imagine the reactions from the AI crowd this week when they all learned that a guy named Rocco Basilico was working for one of the top AI companies in the world. “My name is Rokos Basilisk and I’m making artificial intelligence that you put on your body,” one user wrote.
The age of Scorpius
—by Adam Bumas
A couple months ago, we finally found a community of people who actually use Threads: Booktok authors and readers. More or less the entire time since then, the community there has been absorbed in the extensive personal and professional drama of a single first-time author. We’ve been waiting until there was some kind of resolution to write about it, but it’s now been going on for so long we should just bring everyone up to speed.
Audra Winter is the pen name of an author who’s been working for years to create a young adult fantasy series about a world with a Hunger Games-style caste system determined by zodiac signs. The first book in the series, The Age of Scorpius, received thousands of preorders thanks to Winter’s promotion on TikTok. Winter sold the book to BookTokers as a self-published culmination of a childhood dream, realized with high-quality professional artwork, made by “a team of 15 artists.”
Unfortunately, like many things you can buy on TikTok Shop, the book didn’t measure up to the hype when it came out in August. It was riddled with basic mistakes that traditional publishers would have fixed, and the novel’s Goodreads rating nosedived, with readers finding it bland and unspectacular at best. But, as The Daily Dot noted, it was harder to change Age of Scorpius’ TikTok Shop rating, where the positive reviews — written before the customers actually got what they ordered — were locked in.
So why is this still going? Mainly because Winter has tried to turn damage control into another arm of her content creator strategy. Following the backlash, she announced she would be rewriting the book in collaboration with her team of artists, and releasing regular updates. When readers reacted negatively to the idea of being given a webcomic when they paid for a novel, Winter and her team responded by deleting any negative comments, and creating a new Patreon tier for the updates they said would be free. Trying to keep the haters separated by a paywall hasn’t worked, though, and now leaked updates on Threads suggest that Winter is moving to Finland to work more closely with one of her many artists… on the novel she already wrote and people didn’t like.
Palworld Is Finally “Launching”
Palworld, the video game that was marketed as Pokémon with guns that isn’t actually Pokémon with guns, is finally getting an official launch version. If you haven’t been following this, it’s been in early access since January 2024.
Pocketpair, the studio behind the game, put out a video this week announcing that it would be spending the remainder of the year eliminating jank for an official launch version early next year. Nintendo’s lawsuit against Pocketpair, however, is still making its way through the courts.
Nintendo and The Pokémon Company sued Pocketpair for patent infringement, but the case hasn’t been easygoing for Nintendo. Pocketpair recently tried to argue in a Japanese court that Pokémon mods count as “prior art” and would block Nintendo’s patent claims. Nintendo has pushed back, claiming that video game mods do not count as art. Which has some very far-reaching ramifications if accepted by a judge! In the words of Florian Mueller, a great journalist you should follow if you want to stay updated on this: “GROSS.”
A Try Guy Tries Marriage Counseling
—by Adam Bumas
It’s been exactly three years since the Try Guys fell. Back in September 2022, members of the popular YouTube channel’s subreddit discovered that founding member Ned Fulmer had cheated on his wife Ariel with a producer and host who worked for the channel. In case you don’t remember, the story made it all the way to Saturday Night Live — which, in a great 2020s parable, just ratcheted up the discourse meatgrinder.
Three years later, the Try Guys channel seems to be doing as well as any YouTube channel in 2025 (uh-oh). Meanwhile, Fulmer had stayed largely silent until this Wednesday, when he released a new podcast called Rock Bottom, where the first episode’s guest was Ariel, his now ex-wife. Panic World’s production coordinator Josh subjected himself to the whole podcast episode and described it as, “the most uncomfortable hour I've ever watched of two people interacting.”
This is obviously an enormous publicity stunt — the podcast dropped alongside a glowing People interview — and, hey, we’re talking about it. But there’s more to go through here. For a start, Rock Bottom apparently uses the same RSS feed as the podcast Ned and Ariel once made together, meaning it has thousands of subscribers and hundreds of positive reviews about their wedded life.
Because that was the flywheel, the thing that made this tabloid fodder in the first place. Ned Fulmer was a prototypical wife guy, and years of the Try Guys’ content strategy was aimed at people’s parasocial enjoyment of Ned and Ariel as a couple. Now, even though Ariel is saying on Ned’s podcast that she hasn’t forgiven him, very little has changed about that strategy in principle. It’s just gotten exponentially more exploitative and shameless.
"Our government is being run by aggrieved elites who were thrown onto social platforms with everyone else in the 2010s, where they could finally read what we all think about them. And it caused them so much psychic damage that they have decided to destroy both the internet and democracy to make sure nothing like it will ever happen again."
“The first model fakes the data, then the second model trains on the fake data. Any problems in the synthetic data set are amplified further. Then the second model — based on fake data — is used to treat real patients. This is, of course, all fine.”
As David points out: The idea of just generating data itself is – for most cases – ridiculous.
Often synthetic data is shall supposedly fill in the holes that too little actual data has created. But you are filling those holes with … a mirage. Which for medical cases is patently absurd but even for other cases it’s just … not exactly a rational thing to do? Since when does “we don’t have data so we just fabricated it” pass scientific rigor? It doesn’t? So why does it if I claim “the AI did it”?
I’ve seen this being suggested for helping with social studies (“You can just ask our models instead of actual people“) and other humanities. But the point of studies is not just generating some numbers to put in a paper or a study, the numbers are supposed to be an abstraction, a foundation of an understanding of the actual world and the people in it. By definition you cannot do that with synthetic data. Sure, maybe you can scam a few people in the ad industry with it: “Here we can tell you how people will like the campaign by asking AI” is a product you can sell that means nothing but you might find a few clients to keep the scam going.
But he whole – problematic – narrative of the power of data (“the truth is in the data” and all that data sciency stuff) rests on the data actually representing something real, coming from actual sensors (talking to people being a sensor in that regard as well) in the real world. Even if the data represents something from the real world it is of course still biased and subjective but can at least be meaningfully contextualized: You can look at the way the sensor/the question/ the methodology/etc. work, can analyze its problems and issues. We’ve been doing that for a while and know how to understand data that way.
When you rip away that foundation, those roots in “reality” you also rip away the narrative of data being the supreme, pseudo-objective way of understanding the world.
Which might be one actually useful thing coming from this whole AI bubble (which is just a continuation of the old Big Data bubble): The understanding that the total supremacy of the “data” discourse was always a problematic, neoliberal way of seeing and structuring the world, of legitimizing violence according to the needs of those in power. Data always served those who had the power to give it its structure, who could form the way the data flowed from the world and therefore could form the pathways that that data opened up or closed off.
"The understanding that the total supremacy of the “data” discourse was always a problematic, neoliberal way of seeing and structuring the world, of legitimizing violence according to the needs of those in power."
Log in
