Anthony Po (@anthpo) Lessons on going viral with PR Stunts

Who is anthpo

Anthony Po, who posts as anthpo and is also known as Anthony Potero, is a 20-something creator out of New Jersey who spent about a decade making internet videos before most people knew his name. He did sketch comedy, a rapid-fire high school interview show called 1 Minute Talk Show that pulled 250,000 Instagram followers, a K-pop group at Rutgers called OK-Pop that earned $11,000 in streams and $8,500 in merch by December 2023, and a stint on the short-form team at MrBeast in Greenville for four months in 2023.

Then he started hanging up posters, and everything changed.

He now runs Pufferfish, a content studio announced in November 2025 with Talia Schulhof as CEO and Po as chief creative officer, selling six-figure bespoke campaigns to brands. He retired his personal channels in May 2026 to work on things where he isn't the center of the frame.

Most of the thinking below comes from his interview with Colin and Samir, where he gave away more of the method than he probably meant to. Quotes are lightly cleaned from the transcript for readability.

The stunts, at a glance

1. Run everything through the "would I post this" test

Po's whole filter is one question. He removes himself from the idea and asks whether he'd share it if he stumbled across it cold.

"If I saw this organically I would post about it. That's just the north star for me. I remove myself from the thing and I'm like, if I saw this, would I think it's funny? And that has not failed me yet."

The reason he can run the test honestly is that he's genuinely in the audience. He describes himself as chronically online, on TikTok and Instagram and YouTube all day. He isn't studying a community from a slide. He's in it.

"A lot of creators and brands think of things backwards. Instead of observing a community from the outside, you really have to engross yourself in it."

That's the part most B2B teams skip. You can't run the test if you don't consume the thing you're making. A marketer who never reads the subreddit, never opens the Discord, and never watches the creators in the category is guessing, and the guess reads as a guess.

He pairs it with a second habit he calls the fresh eyes watch: finish the thing, leave it alone for two days, then watch it like a stranger. Most bad campaigns would die in that gap if anyone bothered to wait.

2. Steal from a hundred years back, not last week

The Chalamet contest wasn't an original format. In the 1910s there was a Charlie Chaplin lookalike contest that Chaplin himself reportedly entered and lost. Po knew that, and built the modern version on purpose. He showed up dressed as Chaplin, riding a penny-farthing bicycle he learned to ride for the bit.

"You take a concept and you innovate on it. This happened like 100 years ago. We have a laundry list of things in culture that we love and hold dearly, and that was one of them, and I was like, what's the modern version of that?"

His framing for what makes an idea work is two words: familiar and different. Familiar enough that people instantly get it, different enough that they haven't seen it. A lookalike contest is a format everyone understands in one second. Hanging 50 handmade posters for one around Manhattan is the part nobody had done.

Colin and Samir push this further in the interview, citing Austin Kleon: don't copy what's popular now, find what your favorites stole from, then go back further than anyone else bothers to. The deeper you dig, the more original you look.

3. The poster is the thumbnail

Po has run roughly 15 poster-based projects. The posters are deliberately ugly and deliberately simple, because they're doing a thumbnail's job in physical space.

"We're not trying to create an artful poster, we're trying to get a message across. Here's the picture, here's what it is, here's a time."

There's a New York creator who finds badly designed street posters and redesigns them properly. Po says the guy has run into his posters plenty of times and never remade one, because there's nothing to improve. The message is already the whole design.

The obsessive part is the copy, not the layout. His office argued for an hour about the Chalamet cash prize before landing on $50.

"It's like, no, $20 is funny. No, I'm like, $50 is the funniest."

$50 is funny because it's insultingly specific. It's real money and it's no money at the same time. That's the same instinct behind the chalk logo play: the cheapness isn't a constraint, it's the joke.

The smaller poster stunts are worth studying precisely because they're throwaways:

• Take a selfie with this cat. Same poster batch as the Chalamet run. A photo of a cat, an instruction to take a selfie and email it in. He compiled the replies into a video he says is the most successful thing he's ever made, roughly 27M on one platform, 12M on YouTube, a few million on TikTok, north of 100M total by his own count. It took about five hours to make, and it gets stolen and reposted constantly. Those numbers come from him, not a verified dashboard, so treat them as his estimate.
• Lampshade guy. A poster looking for someone to wear a lampshade for an hour, $30 reward. Someone texted: "I saw your poster, I found my calling." The video version underperformed. He recut it as a carousel and it did around 300,000 likes.
• Honey taste testing. For the bear statue in the Crocs campaign, he hung posters for a honey tasting event, brought in real strangers to sample and vote, then placed the winning honey at the base of the statue.
• Stuffed animal contest, bad hairline competition, bad singing contest, an Anthony meetup. Thrown together fast, mostly during a period when his main channel looked dead.

The stuffed animal one cost him almost nothing and the winner DMed him that it was the best day of her life. That's the actual product.

4. The meta-narrative beats the video

This is the piece Po calls his secret sauce, and it's the most transferable idea in the whole interview.

"What matters more is not the narrative of the video but the meta narrative of the page and who they are. A guy eats a jar of cheese balls, but he's anonymous, and why did he do it? What is the story past the video?"

Cheeseball Man is a man eating snacks in a park, which is nothing. Cheeseball Man is an anonymous masked figure whose identity the internet spent months guessing about, which is a story. Same footage, completely different object.

It's also why he keeps launching new accounts from zero instead of posting to his own audience.

"People want to be at the start of something. A lot of the comments for Kid with Crocs were like, I feel like I'm witnessing history."

Colin makes the comparison to Amazon's working-backwards press release, where teams write the launch coverage before they build the product. Po's version is the same discipline aimed at culture: before you make the thing, work out how two people would describe it to each other at a party. If that sentence isn't fun to say out loud, the thing won't travel.

5. Go anonymous so the idea has to carry it

Gilbert. Cheeseball Man. Kid with Crocs. Three different stunts, all run behind a mask or an alias, in an economy that tells creators their face is the asset.

"In a perfect world I wouldn't ever have to reveal it's me."

His reason is partly proof to himself that he didn't just fall into an audience, and partly that anonymity is contagious in a way that fame isn't. People DM him saying he inspired them to start their own page or run their own bit, which he says matters more than anything he did with his name attached.

There's a strategic edge too. When nobody knows who's behind it, the only thing holding the idea up is the idea. Slate ran the same mechanic with fake business car wraps, re-wrapping a prototype daily as a different absurd company and letting the internet chase the mystery.

Anonymity also has a cost he's blunt about. He can't monetize a thing nobody knows he made, which is why every arc ends with a reveal video. The reveal is the invoice.

6. High effort, low stakes is inherently funny

The honey tasting is the perfect example. He didn't need real strangers voting on real honey to put honey at the base of a statue. He could have bought a jar. When the agency asked why, he had no better answer than that it was funny.

"It's like, well, it's funny. I don't know. It is good, like just trust. If I saw that I'd be like, that's so dumb and beautiful."

Colin names the mechanic outright: high effort on something with no stakes is always funny. The gap between how much work went in and how little it mattered is the joke. Learning to ride a penny-farthing for a bit. 3D-printing custom Crocs for a bronze bear. Counting out 700 cheeseballs. Spending $7,000 on a fake alien crash.

For a brand, this is the cheapest credibility available. Effort is visible and it can't be faked. It reads as caring, which is the one thing an audience can't be tricked into perceiving.

7. Build a world, then let other people play in it

Cheeseball Man was supposed to be one video. Po had wanted to make a superhero video, and eating a jar of cheeseballs in Union Square was the origin story. He credits Philadelphia's Alexander Tominsky, the guy who ate 40 rotisserie chickens, with the seed.

Then it kept going. Patrols. Cleaning up the city. A superhero league. And then, without any involvement from Po, a villain.

"A super villain appeared and he gained like 300,000 followers, and then we had a fight and I had no idea who he was."

Cornhead Killer was a fan invention. He'd been pasting "Push" signs over "Pull" ones and putting fake electrical outlet stickers in airports since August, then confronted Cheeseball Man in Washington Square Park to challenge him to a fight. The stakes: if Cornhead won he'd get the Cheeseball Man accounts, if Cheeseball Man won he'd unmask Cornhead. They fought at Church Street Boxing Gym on November 16. Cheeseball Man won. Cornhead turned out to be California YouTuber Andrew Levitt, who had flown to New York five separate times to appear in character.

The lookalike contest did the same thing at a bigger scale. Po encouraged people to run their own instead of franchising it himself, and the format went worldwide. Samir shows him a poster he photographed in LA for a Taylor Zachary Perez lookalike contest, $25 prize, a name Samir didn't recognize. That's the format working exactly as intended.

"The internet can feel very secular, perpetually in one place on your screen. I wanted it to be like, oh, we can hang out in real life too."

The lesson for brands is uncomfortable: the format has to be simple enough to steal, and you have to be fine with people stealing it without crediting you. Po says he never felt left out of the coverage. He wanted the joke to travel more than he wanted the byline.

8. His actual framework: six principles, plus effort

When Po pitches brands, he pitches psychology. He learned the six principles of influence in AP Psych in high school and has been running on them ever since. Here's how he applied each one to Kid with Crocs.

The similarity one is where he's hardest on his peers.

"Most street artists and 3D artists think they're too good to follow trends and sounds and put text on the screen. I can't be afraid to use Proxima Nova. I can let my artistry shine through in other ways, but I need to be familiar."

And the caveat he adds immediately after listing all six matters more than the list:

"You could only check two of them and you can still get people to care. Getting people to care is so much more meaningful than getting 10 million views on a video. You can follow these to a tee, but if you're not making something impactful, people will see through it."

9. Plan the failure in advance

Po describes himself as neurotic and treats it as a professional asset. He calls his prep a production bible, a habit he picked up from the chaos of MrBeast productions: if this happens, we do this.

For the Chalamet contest, once a friend on set told him Chalamet was talking about the posters two weeks out, the team stopped planning the contest and started planning for arrest. They made #freegilbert shirts. They pre-shot an entire video to release if he got booked, a message to the audience from jail. He didn't get arrested, though he was fined $500 and at least one attendee was, and when the park kicked them out he marched thousands of people to a second park because that contingency was already written.

For Kid with Crocs, the ladder was four rungs deep:

1. If the account stalls, 10x the posters.
2. If that stalls, put Crocs on the Wall Street Bull, which he knew would force coverage.
3. If that stalls, make a long-form YouTube documentary about the process.
4. Carousel posts as a parallel format, since they sometimes outperform video.

He also had a press list of hundreds of reporters built after the Chalamet contest, and deliberately never used it.

"That feels like an in case of emergency. It feels like cheating."

He planned the first four videos completely, then built the last four from what the audience responded to in the first four. That's the whole campaign design in one sentence: commit hard, then listen.

10. Not all views are the same

This is the metric argument every B2B team needs and almost none run.

"800,000 likes on this is like an army versus 800,000 likes on some kid with text on top of the screen. No one's talking about a meme that has 800,000 likes the way I'm talking about this guy."

Samir points at Po's own catalog for proof. After the Cheeseball Man boxing match, Po posted a video titled simply "I fought him." Roughly 600,000 views, from a title and thumbnail that could only mean anything to someone already inside the story. Those views are worth more than a million algorithmic ones, because every single one is a person who was following along.

Depth over width. A hundred people who care beats a hundred thousand who scrolled past. His advice to creators is the same as his advice to founders:

"If you're struggling to go from zero viewers on Twitch to 10, that is a you problem. You need to engage with people, get close with them."

What brands should actually take from this

Po's brand work is where this stops being creator theory and starts being a marketing argument.

Crocs. They came to him after his producer Paige posted about the Chalamet contest on LinkedIn and it did around 80,000 likes. Crocs let him pitch several ideas and picked the most complicated one. He grew Kid with Crocs from zero to 100,000 followers in five days and 780,000 in 40, across eight statues including Paul Manship's 1932 Group of Bears in Central Park, five of Tom Otterness's Life Underground subway sculptures, and the Balto statue, ending with a five-foot Croc signing. Zero paid media. Over 55 million views. He says he got maybe three hate comments across the whole run.

His read on why it worked for them commercially is the same one any CMO would recognize, phrased better:

"At the heart of product placement is positive association. As a brand, is Crocs cool with the crew? Oh, okay, they're cool with my group of friends, so I guess I feel positively about them."

Positive sentiment lowers acquisition cost. That's the entire brand awareness case in one line.

Adobe. The bunny tattoo campaign in November 2025 was a paid #adobepartner deal. He designed tattoo options in Photoshop on mobile, ran them through a selection mechanic, let a bunny pick the winner, and got it permanently inked. The product demo is real and the video isn't an ad in any way your brain flags as one. Watch the setup and the payoff, plus the alternate cut, the Instagram Reel, and the YouTube Short.

Airlearn. Pufferfish's first campaign put billboards of real foreign words that read as filthy in English across New York. "Kock" is Swedish for chef. "Pussi" is Finnish for bag. It was Airlearn's first US marketing campaign, and it's in our OOH roundup for a reason.

His closing advice to brands is two sentences long:

"Trust creators and take more risks. It's better to work with a creator who has 10,000 followers who's passionate and in love with your brand and willing to go the extra mile than someone with a million followers who's just going to buy Saratoga water."

One contract detail worth copying. The only thing Po red-lined with Crocs was ownership: he keeps the account, and he gets award show credit. Brands can approve posts, but the thing he pours himself into stays his.

The limits, honestly

A post about stunts that only shows the wins isn't useful. Three of his went sideways, and the failure modes are as instructive as the hits.

The furry video. In May 2024 he posted "I Spent 200 Days Undercover as a Furry," about buying a fursuit and getting into a furry party. The backlash centered on censored party footage and whether the people in it had consented to being filmed and published, which the video didn't clearly address. He removed the footage after a furry involved asked him to, turned the video into a Trevor Project fundraiser, and apologized publicly. The lesson is the boring one: a stunt built on a community you're outside of needs consent worked out before the camera turns on, not after.

The Miami alien hoax. He spent weeks in a Florida Airbnb amplifying an existing rumor that aliens had been spotted at Bayside Marketplace, using fake accounts to flood comments with fake eyewitness reports, hiring a Pedro Pascal impersonator and a Star Trek II actress as a "whistleblower," staging a crash scene with Hazmat agents and a $250 prop alien. Total spend passed $7,000, including a $2,800 VFX bill for a few seconds of footage. Reddit users cracked it fast, partly because the prop alien was one of the first results when you searched for alien body props. Manufacturing fake evidence is a different category from putting shoes on a statue, and it's the one stunt here I'd tell a brand to never run.

Everybody's Album. With artist Danny Cole under the banner "nano," he paid over 100,000 people $7.99 each, delivered as a Shopify gift card usable only on the pre-order, in exchange for one second of audio. The 23rd track, "Everybody," strings all 100,000 clips together. The volume of transactions was engineered to earn a No. 1 Billboard 200 debut. Billboard ruled it ineligible. The idea was beautiful and the mechanism was clever, and the institution said no anyway. If your stunt's payoff depends on a third party's rules, read the rules first.

The Chalamet contest also cost him a $500 fine, and the crowd number is genuinely fuzzy. Po paid someone $100 to count from AP and his own footage and got about 7,500. Press at the time said anywhere from several hundred to thousands. He says 10,000 in the interview. Take the range, not the headline.

So what do you actually steal?

If you run marketing at a B2B company and you've read this far waiting for the applicable part, it's this.

Stop asking what your audience wants to see and start being a member of the audience, so you can run the "would I post this" test honestly. Pick a format so simple people can copy it without you. Spend the effort somewhere visible and pointless. Write the press release before you build the thing. Plan the failure ladder in advance so the campaign can't quietly die. And measure whether anyone cared, not how many people scrolled past.

The budget was never the hard part. Po hung 50 posters and started a worldwide format. The hard part is having an idea absurd enough that people do your distribution for you, and caring enough to over-build a detail nobody asked for.

"How do you get someone to care? That's so much more meaningful than getting 10 million views on a video."