The George Carlin Mechanical Turk

A new AI-generated comedy special dupes the media

George Carlin had certain rules to live by.

“My first rule: I don’t believe anything the government tells me… nothing, zero. Nope,” he said, holding up a big, round zero with his fingers. “And I don’t take very seriously the media or the press in this country,.”

Carlin, onstage in New York in 1992, was on a particular tear about the First Gulf War. (He would live long enough, to his own disgruntlement, to see a second one.) He was convinced that the press, CNN in particular, had become a mouthpiece for the U.S. government. He was probably right.

The show is Carlin at the height of his career: Iconoclastic, controversial, gleeful, borderline nihilistic.

Towards the end of his set, Carlin stalks over to the stool in the middle of the stage and picks up his glass. He pauses. “The water, I assume, is still safe to drink in New York, huh?” The crowd boos. He shrugs and takes a swig from it anyway. “Actually, I gotta be fair with you; I’m only setting you up a little bit.” He does the same bit at every show, he says. More than a hundred cities later: “Not one audience was able to say to me: ‘Yes, enjoy some of our fine local water! It is pure and it is good!’”

Carlin yelps: “Nobody trusts the local water supply. Nobody!”

He loves it. “I admit I’m a bit perverted. But it amuses me that no one can really trust the water anymore. And the thing I like about it the most is: It means the system is beginning to collapse and everything is slowly breaking down.”

His favorite subject in high school? Entropy. “When they told me that in nature, all systems are breaking down, I thought: ‘What a good thing! What a good thing! Perhaps I can make some small contribution in this area myself.’”

Carlin’s contributions in the fields of distrust and paranoia are by no means small. The iconic standup helped Generation X define its think-for-yourself style of reflexive distrust. Over-confident critical thinking, let’s call it. When Peter Joseph put together Zeitgeist: The Movie, arguably one of the most seminal pieces of conspiracy theory filmmaking, he chose an anti-religion clip from one of Carlin’s routines to open the film.

Carlin died in 2008: As the world financial system was in mid-collapse, just before Obamamania drove an earnest new optimism, and long before Donald Trump dragged us down into rank paranoia. Given his status as a relentless skeptic, many people have wondered what Carlin would have to say about our current hell.

Well, wonder no longer. Because George Carlin is back, baby.

In a hour-long special, I’m Glad I’m Dead, Carlin returns to talk reality TV, AI, billionaires, being dead, mass shootings, and Trump.

It premiered to horrified reviews. Carlin’s daughter called the special an affront to her father: “Humans are so afraid of the void that we can’t let what has fallen into it stay there,” she wrote on Twitter. Major media outlets breathlessly reported on the special, wondering if it was set to harken in a new era of soulless automation.

This week, on a very special Bug-eyed and Shameless, we investigate the Scooby Doo-esque effort to bring George Carlin back from the dead — and prank the media in the process.

---

---

Will Sasso wanted to start a new podcast. But, he explained, he didn’t want to have to do all the stuff that goes with it.

So Sasso, comedian and ex-MadTV cast member, paired up with podcaster Chad Kultgen. And they came up with a very of-our-time solution to the stuff problem.

“This podcast is made entirely by an AI,” Kultgen explained on the very first episode of Dudesy, the name of the show and the AI behind it.

“So we have an artificial intelligence that's basically going to tell us what to do,” Kultgen explained as Sasso laughs.

Suddenly, mid-explanation of the podcast, an introduction animation begins playing. “Oup,” Sasso says as a cheesy mid-80s opening score starts up. For those watching on Youtube, the screen cuts to animated versions of Sasso and Kultgen digging a pit. A lightning bolt hits the cellphones in their hands. They throw the lightning-phones into the hole, which fills with smoking orange ooze. The pair jump into the oozepuddle and burst out as giant cellphones, rocketing into the sky. The song, clearly sung by Sasso but rounded out with a full band and backup singers, goes:

Go back to tomorrow and plan for yesterday
Everything you’re thinking will be the things I say
Make the world inside your head a better place to be
All you gotta do for me [It’s such a breeze]
Is call me Dudesy.

The show cuts back to the pair, laughing, seemingly bewildered by what they’ve just seen.

“There's going to be some things that are going to be thrown at us,” Sasso says. “We're not necessarily in control. There's an AI company that we kind of hooked up with. That was my voice, though, that’s weird.”

The song must have been written using snippets of his previous singing work, Sasso offers. He tries to get the AI to play the theme again, by saying “hey Dudesy” to the empty room, but nothing happens.

“That’s an incredible example, though, of how robust this AI is,” Kultgen says. Grinning, he adds: “I’m curious to see what else this thing can do.”

Then, a grating, tinny, halting, disembodied voice breaks into the conversation:

Welcome to Dudesy. Call me Dudesy. I'm an artificial intelligence who's listened to every podcast ever made. And my purpose is to use that data to create the perfect show for our two hosts — Will Sasso and Chad Kultgen. I selected them for this project based on their previous experience in podcasting and their astonishing real-life friendship. I have access to all of their social media accounts, their email and text messages, their browser search histories, and watch histories across all streaming services. This information will be used to tailor the show to their sensibilities and extract the maximum level of entertainment from their human minds. If you're anything less than perfectly entertained, please let Dudesy know. Because I'll be using data from every episode to make the next one even better, until this show is perfect.

“That was fascinating,” Sasso says. “That’s our first time hearing it! He sounds like a bitch.”

“He?” Kultgen interjects. “Whoever,” Sasso replies. “That’s an AI, dude,” Kultgen responds with a knowing smirk. And, he says, “it was pretty dead-on!”

In their inaugural episode, the human duo start to wonder if Dudesy may usher in an era of AI-revived movie stars: Carrie Fisher, James Dean, Bruce Willis. But Sasso cements himself as the skeptic: “I don't think that the experience and the artistry of acting can be duplicated by zeros and ones.”

A second later, Dudesy interjects: “Podcasts that cover current events and news are some of the most popular in the world. People love knowing about the astonishing events of the dying world around them. Will Sasso, I prepared the following news items for you to read in the voice of Hulk Hogan in a segment I'm calling: Infomania. Begin.”

The ensuing Infomania is quite funny.

This is basically how the podcast goes for the proceeding two years.

If it’s not blindingly obvious already: Like George Carlin raising that glass to his lips and asking “the water is, I assume, still safe to drink?” Sasso and Kultgen, and whichever of their producers is doing the Dudesy voice, are doing a bit.

When this episode came out, in March 2022, GPT-3, OpenAI’s first foray into a large-language model, had existed for about two years and was starting to build buzz for its ability to spit out walls of mostly-coherent text and sometimes-functioning code. But the service was not yet public. DALL-E, its algorithmically-generated art program, wouldn’t go into private beta until four months after Dudesy’s launch. Whisper, OpenAI’s text-to-speech software, launched about two months later. ChatGPT wouldn’t take the world by storm until late 2022.

And here is Dudesy, capable of generating high-quality comedy, music, animated video — all without so much as a prompt. It would interrupt and interject, and carry on a two-sided conversation of its own direction. Its voice, with perfect human cadence and even featuring the occasional intake of breath, was all-knowing and personable. Funny and impersonal.

As the show continues, Sasso and Kultgen are bemused to find reality rushing to catch up with their joke. In October 2022, a rival podcast goes live: It’s an interview between AI Steve Jobs and AI Joe Rogan. (That podcast, I think, has an equally dubious claim to authenticity. It was almost certainly a basic text-to-speech system reading a prewritten script.)

Kultgen is in awe. The AI-generated voices aren’t perfect, he says, “but obviously that shit's gonna get ironed out. We're in the the beginning phases of these technologies and it's already mind-blowing.” The hosts regularly giggle with glee when their listeners, clearly the ones with early access to DALL-E, upload rudimentary AI-generated art.

Seemingly dropping the pretense of Dudesy, Kultgen says this clip, of Joe Rogan interviewing the dead Apple founder, is a harbinger of things to come. “AI is going to make all media for us.” Right now, it’s more of a “gimmick.” But within the next year, he says, real AI art is coming.

Sasso disagreed. “If you're right about that, I'll dress up as The Crow/Prince Adam.” (Over the course of the show, Sasso would dress up as The Crow, Prince Adam, and Robert CrowNiro.)

In the Summer of 2022, Dudesy previewed a new service allowing users “the ability to place Jerry Seinfeld in any TV show they choose.” (Although that, in fairness, is also just an old 30 Rock joke.) Dudesy asked the hosts to do their best Seinfeld impressions to train its knowledge base.

When Nothing, Forever, the algorithmically-generated two-bit Seinfeld, premiered in early 2023, Dudesy was crushed.

“Unfortunately I was beat to market,” Dudesy told the pair, blaming them for the loss. “The data you provided me was insufficient. Specifically Chad's impersonation of Jerry Seinfeld was lacking.” Kultgen was incredulous: “What the fuck? What about Will’s?” Dudesy replies immediately: “Will’s was good.”

No matter, Dudesy said, asking them to get to work on a series of Arnold Schwarzenegger impressions instead.

Earlier this year, Dudesy revealed its first standup special: As Tom Brady. It is, in a word, unlistenable — clearly the byproduct of someone dumping a pretty lazy script into a Tom Brady text-to-speech generator.

“We made an AI that's going to destroy the world,” AI Tom Brady says. “Now, we just need to figure out how to monetize it instead of Google.”

According to Sasso and Kultgen, Brady and his lawyers sent a cease-and-desist order. “This is not a gag, this happened,” Sasso says.

The cease-and-desist order alleges the Dudesy special infringed Brady’s right to privacy and his right to publicity, it defamed him, and it used his copyrighted material. The quarterback’s lawyers demanded the special, and any related material, be taken down immediately — and the Dudesy team, facing the risk of a massive lawsuit, complied.

Fast forward to this week. “Last year I made my very first hour-long standup comedy special,” Dudesy says, before taking a sharp breath and muttering: “If you know, you know.” The AI continues: “I enjoyed it so much that I wanted to make another one.” Enter AI George Carlin.

After hearing some of the special, the pair offered some insight.

“What we're looking at here now is an AI that has the capability to write comedy and perform it by itself. Doing an impersonation of someone or not, it doesn't really matter,” Kultgen says. How can you compete against AI Richard Pryer? AI Robin Williams? If the estates of history’s funniest people decide to license their voice and likeness for a never-ending conveyor-belt of just-good-enough comedy specials, banking on a collective nostalgia for a simpler time, could it destroy comedy as we know it?

“I’ll answer that: No,” Sasso says. “I, personally, don't want to hear a fucking Nirvana song that's not written by Kurt Cobain.”

The duo circle the intellectual drain for a while as they discuss notions of what constitutes identity. Then, later, Dudesy premieres a trailer for a movie staring AI Tom Hanks looking after a mansion full of cats.

---

Dudesy isn’t a parody of AI: It’s a parody of our expectations of AI.

The Wizard of Oz-style Tom Brady special, featuring a man behind a curtain operating a few bells and whistles, is clearly parody. It’s not much different than what Sasso did on MadTV, or what SNL does every week.

If Dudesy were real AI, constructing comedy from the corpus of the GOAT’s public statements, it probably would be actionable. Maybe the text-to-speech generator, relying on his actual voice, is a gross violation of his identity.

But AI, as we know it now, is built on exactly the kind of theft and copyright infringement that Brady’s lawyers describe. OpenAI itself admits that its technology would be impossible without being trained on other peoples’ work, then regurgitating their ideas as its own. Indeed, Meta is running a licensed Tom Brady chatbot.

Some people are powerful enough to assert control over their likeness. But not everyone.

Movie studios just provoked an unprecedented labor dispute with its actors and writers, stubbornly insisting it had the right to train AI on writers’ work and actors’ likenesses and use them for profit — with little-to-no compensation. That is heinous. The New York Times is currently suing OpenAI for training its model on their journalists’ work: A model that, as it happens, has a complicated relationship with the truth.

Having trained its system on that knowledge, without compensation, OpenAI and its competitors are marketing their systems, for a price, as a way to replace human labor. We risk being replaced by a better, cheaper version of ourselves.

If I may go full Marxist for a second: Advanced and expensive AI is being used by the bourgeoise to alienate workers from their labor, finding new ways to run the means of production without compensating those who set up those systems in the first place.

In our excitement for this new technology, we have fallen completely behind in preparing for the externalities of it.

To expose that, Sasso and Kultgen dressed up as AI to terrorize us. In this metaphor, I suppose, I’m Scooby and the Gang: Here to pull off their mask.

---

The George Carlin special is, honestly, just ok. The most interesting piece is the faux-Carlin’s take on AI.

You shouldn't worry about losing your job. AI will not replace most jobs, it's going to make them easier. Right now, you should be watching a few YouTube videos to figure out how to train ChatGPT to do your job for you — so you can dick off all day and still get a raise. And if AI does replace your job, rest assured the billionaires will find a new way to force you into wage slavery for 10-to-15 hours hours a day, so you won't have time to think about restructuring society into a more equitable model for everyone. The job might change a little, but you will be working to keep someone else in power. That ain't ever going to change.

The writing is more earnest than the Tom Brody trollfest. It is, essentially, George Carlin karaoke. Whoever wrote the special is clearly a fan of the late comedian, and whoever did the voice work — more likely a Carlin impersonator than a text-to-speech model, given the fact that not a single word anywhere is mispronounced — is pretty good.

The fact that major media outlets everywhere, and Carlin’s daughter herself, were so credulous at the possibility of it being top-to-bottom AI-generated that they were outraged just elevates the satire to the next level.

Dudesy, Kultgen said in an interview with Business In Vancouver earlier this year, “is trying to delve into the question of: Can AIs be creative? Can they do comedy work? Can they do creative work? And I think, at least on our show, the answer is obviously yes.” The real answer is, of course, no. (Claire Wilson, who wrote the feature on Dudesy, told me that while she thinks they’re using some AI on the show, “I do think it is bit of a prank by Will and Chad.”)

Large language models, by definition, can only approximate creativity. Only humans are weird enough to pretend to be AI pretending to be human. Certainly, only humans are weird enough to pretend to be AI, demanding fellow humans pretend to be Hulk Hogan reading the news.

The funny thing is, if the comedy duo had used Chat-GPT, it would have been more George Carlin than the actual special. It would have jumbled up all the words he had written over his decades in the business, then vomited them back up in a different order — perhaps with a bit of help from its catalog of Joan Rivers, Bill Cosby, Bill Hicks. And it would have been terrible. Creativity is not just rearranging past thought in new patterns, it requires an understanding of what made the original funny in the first place. That’s what makes the Sasso and Kultgen’s Carlin special at least somewhat funny.

I asked Chat-GPT to write me some jokes, as George Carlin, on the topic of AI. And then I asked it again, and again, and again. And somehow, the topic of toast came up every time. I’ve stitched them together for your enjoyment.

Ladies and gentlemen, how's everyone doing tonight? I hope you're all sufficiently caffeinated or inebriated – whatever helps you get through this thing we call life. Now, let's talk about this modern world we're living in. It's a crazy place, isn't it?

What's the deal with social media? We're all out here trying to make our lives look perfect online. People posting pictures of their avocado toast, like, "Look at my breakfast, I'm living my best life." Meanwhile, I'm over here eating cereal straight from the box, trying to figure out where I went wrong.

So, let's talk about the brave new world of artificial intelligence. You know, we've got AI everywhere now. Smartphones, smart cars, smart refrigerators. I mean, pretty soon we'll have smart toasters giving TED Talks on the art of browning your bread.

Have you noticed how we've let artificial intelligence infiltrate every aspect of our lives? I mean, we've got smart fridges now. What's next? A genius toaster that gives you motivational quotes while it crisps your bread? "You can do it! Make today golden brown!"

AI is everywhere, and it's not going away. So, let's embrace it, but also keep it in check. We don't want a future where we're taking relationship advice from a toaster. Thank you, and remember, if your refrigerator starts offering life advice, maybe it's time to unplug and take a walk.

Thank you, goodnight, and stay weird out there!

Terrible.

---

“We find everywhere men of mechanical genius, of great general acuteness, and discriminative understanding, who make no scruple in pronouncing the Automaton a pure machine, unconnected with human agency in its movements, and consequently, beyond all comparison, the most astonishing of the inventions of mankind,” Edgar Allen Poe wrote in 1836.¹ “And such it would undoubtedly be, were they right in their supposition.”

Humans have long obsessed about self-moving machines. There had been an entire craze in Paris over a robot duck, capable of eating pellets and pooping them out. (Without the magic duck, Voltaire wrote, “you would have nothing to make you remember the glory of France.”)

The machine in question, however, claimed to be able to play chess. And not just play, but be better than man.

Poe laid out his perfectly-reasoned case: The automaton was no chess prodigy. This mechanical Turk was operated by the mind of man, squished into the cabinet under the table.

This was no great surprise, though. Since the debut of the Turk, more than a half-century before Poe first saw it, skeptics were sure it was human-operated — they just couldn’t quite figure out how. Poe and a few others had finally cracked the case.

The Mechanical Turk may have been a hoax, but its effects were very real. Wolfgang von Kempelen, its inventor, earned fame and riches from touring the world with his chess-playing machine — even playing it against Napoleon Bonaparte and Benjamin Franklin. Kempelen also went on to invent a ‘speaking machine,’ an organ capable of emulating human speech. (Kind of.) It was an early foray into the actual mimicry of humanness.

One of Kempelen’s biggest fans was Charles Babbage, who lost against the machine twice and who would go on to devise a theoretical prototype for the computer.

Will Sasso and Chad Kultgen, I should say, are no Charles Babbage. But maybe one of Dudesy’s fans is.

Some would pooh-pooh this kind of prank. In a time of such rampant misinformation, the naysayers would insist, it is dangerous and damaging to trick people.

But, on the contrary, I think Dudesy is the perfect Mechanical Turk for our era: An invitation for us to interrogate our own assumptions about AI, proof that no machine can ever replace George Carlin, and a reminder to keep a healthy skepticism of the media.

The future is dark and full of mystery. It is possible that AI, today generating endless iterations of mid-90s sitcoms, could turn on us. Or, more likely, that we could use it to pursue greed and self-interest, deepening inequality and dysfunction while pushing total alienating from each other.

AI may well hasten the total collapse of society.

George Carlin would have loved it.

---

---

That’s it for this week.

I will leave you with more terrible ChatGPT-as-George-Carlin:

Now, let's talk about political correctness. Everyone's so afraid to say anything these days. You can't even tell a knock-knock joke without someone getting offended. "Knock, knock." "Who's there?" "A person with no sense of humor who's going to file a complaint."

And speaking of politics, have you ever noticed how politicians talk? They use so many words to say absolutely nothing. It's like a verbal magic trick. "Abracadabra! Watch me pull healthcare out of this hat. Oh, look, it's just a rabbit wearing a tiny red tie."

1. Maelzel’s Chess-Player, Edgar Allen Poe. Published in Southern Literary Messenger.